What are the principles of stroke management?

The goals of a comprehensive acute stroke treatment management are (1) to minimize the amount of brain damage by reversing penumbra and by preventing further bleeding in acute hemorrhages , (2) to prevent early neurological and medical complications, and (3) to promote early recovery.

What are the general management of stroke?

Basic life support, control of bleeding, seizures, blood pressure (BP), and intracranial pressure , are critical. Medications used in the treatment of acute stroke include the following: Anticonvulsants - Prevent Seizure recurrence. Antihypertensive Agents - Reduce BP and other risk factors of Heart Disease.

What is the best medical management stroke?

Emergency IV medicine. Quick treatment improves your chances of survival and may reduce complications. An IV injection of recombinant tissue plasminogen activator (TPA) is the gold standard treatment for ischemic stroke. The two types of TPA are alteplase (Activase) and tenecteplase (TNKase).

What is the definition of a stroke in PDF?

stroke as ' rapidly developed clinical signs of focal . (or global) disturbance of cerebral function, lasting . more than 24 hours or leading to death, with no . apparent cause other than of vascular origin '.

What are the ABCs of stroke management?

These care processes were then grouped into 3 domains and labeled the ABCs of Stroke Prevention: (1) anticoagulation management, (2) blood pressure management, and (3) communication .

What are the best interventions for stroke patients?

Physical therapy uses exercises to help you relearn movement and coordination skills you may have lost because of the stroke. Occupational therapy focuses on improving daily activities, such as eating, drinking, dressing, bathing, reading, and writing.

What is the correct management of a stroke?

The main treatment for an ischemic stroke is a medicine called tissue plasminogen activator (tPA) . It breaks up the blood clots that block blood flow to your brain. A healthcare provider will inject tPA into a vein in your arm. This type of medicine must be given within 3 hours after stroke symptoms start.

What is the most important treatment priority for a stroke patient?

The first goal is to find and control the bleeding . If you take blood thinners, the doctor will take you off them. The next step depends on what caused your stroke.

What heals the brain after a stroke?

Brain exercises can help people regain thinking, reasoning, and memory skills after a stroke. Other brain-strengthening activities include eating a heart-healthy diet, getting regular exercise, and limiting alcohol consumption. Stroke recovery begins before you leave the hospital.

How to calm a stroke patient?

Emotional well-being When caring for a stroke survivor: Be patient: Allow the person to do or say things in their own time and in their own way . Give them time and space to practice speech, movement, or other skills. Be respectful: Treat the person as a partner rather than a victim.

What is stroke in one word?

: the act of striking . especially : a blow with a weapon or implement. 2. : a single unbroken movement. especially : one of a series of repeated or to-and-fro movements.

What is stroke in simple words?

A stroke can occur when blood flow to the brain is blocked or there is sudden bleeding in the brain . There are two types of strokes. A stroke that occurs because blood flow to the brain is blocked is called an ischemic stroke. The brain cannot get oxygen and nutrients from the blood.

What are the general principles of stroke rehabilitation?

planning rehabilitation, including goal setting. screening, assessment and interventions for cognitive impairment and psychological changes, and difficulties with communication and movement. providing information and support. self-care and long-term health and social support.

What are the principles or procedures for the first aid management of a stroke?

Conscious casualty Place in a position of comfort and support the head and shoulders with pillows . Loosen tight clothing. Ensure airway is clear and open and wipe any secretions away from the mouth. Do not give anything to eat or drink, as swallowing may be impaired.

Stroke: Practical Management - PDF Free Download (2024)

9781405127660_1_pre.qxd 10/13/07 11:15 AM Page i

Stroke Practical Management THIRD EDITION

9781405127660_1_pre.qxd 10/13/07 11:15 AM Page ii

Dedication To our many colleagues, young and old, with whom we have shared the care of so many stroke patients, and with whom we have discussed so many interesting ideas.

Acknowledgement The authors of this book are particularly grateful to Joanna Warldlaw, who has drawn much of the line artwork throughout the three editions.

9781405127660_1_pre.qxd 10/13/07 11:15 AM Page iii

Stroke

Practical Management THIRD EDITION

C. Warlow J. van Gijn M. Dennis J. Wardlaw J. Bamford G. Hankey P. Sanderco*ck G. Rinkel P. Langhorne C. Sudlow P. Rothwell

9781405127660_1_pre.qxd 10/13/07 11:15 AM Page iv

© 2007 C. Warlow, J. van Gijn, M. Dennis, J. Wardlaw, J. Bamford, G. Hankey, P. Sanderco*ck, G. Rinkel, P. Langhorne, C. Sudlow, P. Rothwell Published by Blackwell Publishing Blackwell Publishing, Inc., 350 Main Street, Malden, Massachusetts 02148-5020, USA Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK Blackwell Publishing Asia Pty Ltd, 550 Swanston Street, Carlton, Victoria 3053, Australia The right of the Author to be identified as the Author of this Work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher. First edition published 1996 Second edition published 2001 Third edition published 2008 1

2008

Library of Congress Cataloging-in-Publication Data Stroke : practical management / C. Warlow . . . [et al.]. – 3rd ed. p. ; cm. Includes bibliographical references and index. ISBN 978-1-4051-2766-0 (hardcover : alk. paper) 1. Cerebrovascular disease. 2. Cerebrovascular disease – Treatment. I. Warlow, Charles, 1934[DNLM: 1. Cerebrovascular Accident – therapy. 2. Intracranial Hemorrhages – therapy. 3. Ischemic Attack, Transient – therapy. WL 355 S9208 2007] RC388.5.S847 2007 616.8’1 – dc22 2007022955 ISBN: 978-1-4051-2766-0 A catalogue record for this title is available from the British Library Set in 9/12pt Stone Serif by Graphicraft Limited, Hong Kong Printed and bound in Singapore by Fabulous Printers Pte Ltd Commissioning Editor: Martin Sugden Editorial Assistant: Jennifer Seward Development Editor: Lauren Brindley Production Controller: Debbie Wyer For further information on Blackwell Publishing, visit our website: http://www.blackwellpublishing.com The publisher’s policy is to use permanent paper from mills that operate a sustainable forestry policy, and which has been manufactured from pulp processed using acid-free and elementary chlorine-free practices. Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The Publisher is not associated with any product or vendor mentioned in this book. The contents of this work are intended to further general scientific research, understanding, and discussion only and are not intended and should not be relied upon as recommending or promoting a specific method, diagnosis, or treatment by physicians for any particular patient. The publisher and the author make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of fitness for a particular purpose. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of medicines, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. Readers should consult with a specialist where appropriate. The fact that an organization or Website is referred to in this work as a citation and/or a potential source of further information does not mean that the author or the publisher endorses the information the organization or Website may provide or recommendations it may make. Further, readers should be aware that Internet Websites listed in this work may have changed or disappeared between when this work was written and when it is read. No warranty may be created or extended by any promotional statements for this work. Neither the publisher nor the author shall be liable for any damages arising herefrom.

9781405127660_1_pre.qxd 10/13/07 11:15 AM Page v

Contents

Contributors, vi Acknowledgements, vii Abbreviations, viii 1 Introduction, 1 2 Development of knowledge about cerebrovascular disease, 7 3 Is it a vascular event and where is the lesion?, 35 4 Which arterial territory is involved?, 131 5 What pathological type of stroke is it, cerebral ischaemic or haemorrhage?, 181 6 What caused this transient or persisting ischaemic event?, 259 7 Unusual causes of ischaemic stroke and transient ischaemic attack, 353 8 What caused this intracerebral haemorrhage?, 411 9 What caused this subarachnoid haemorrhage?, 457 10 A practical approach to the management of stroke and transient ischaemic attack patients, 503 11 What are this person’s problems? A problem-based approach to the general management of stroke, 537 12 Specific treatments for acute ischaemic stroke, 635 13 Specific treatment of intracerebral haemorrhage, 703 14 Specific treatment of aneurismal subarachnoid haemorrhage, 719 15 Specific interventions to prevent intracranial haemorrhage, 767 16 Preventing recurrent stroke and other serious vascular events, 789 17 The organization of stroke services, 903 18 Reducing the impact of stroke and improving the public health, 953 Index, 980 Colour plates are found facing p.550

9781405127660_1_pre.qxd 10/13/07 11:15 AM Page vi

Contributors

Charles Warlow University of Edinburgh, Western General Hospital, Edinburgh, UK

Jan van Gijn Utrecht University, Utrecht, the Netherlands

Martin Dennis University of Edinburgh, Western General Hospital, Edinburgh, UK

Joanna Wardlaw University of Edinburgh, Western General Hospital, Edinburgh, UK

John Bamford St James’ University Hospital, Leeds, West Yorkshire, UK

Graeme Hankey Royal Perth Hospital, Stroke Unit, Perth WA, Australia

Peter Sanderco*ck University of Edinburgh, Western General Hospital, Edinburgh, UK

Gabriel Rinkel Utrecht University, Utrecht, the Netherlands

Peter Langhorne Academic Section of Geriatric Medicine, Royal Infirmary, Glasgow, UK

Cathie Sudlow University of Edinburgh, Western General Hospital, Edinburgh, UK

Peter Rothwell Department of Clinical Neurology, Radcliffe Infirmary, Oxford, UK

9781405127660_1_pre.qxd 10/13/07 11:15 AM Page vii

Acknowledgements

We have had invaluable help and advice from many people in the preparation of this third edition. So thank you all, including:

Also, thank you to our teachers and colleagues from whom we have learned so many worthwhile things over the years:

Sheena Borthwick Judi Clarke Carl Counsell Ann Deary Alice Emmott Hazel Fraser Paut Greebe Gord Gubitz Ingrid Kane Sarah Keir Alistair Lammie Lynn Legg Richard Lindley Mike McDowell Michael Mackie Ian Marshall Nick Morgan Ross Naylor Sarah Pendlebury David Perry Rustam Al-Shahi Salman Cameron Sellers Mark Smith Ian Starkey Stuart Taylor Brenda Thomas Theo vanVroonhoven Nic Weir

Henry Barnett Lou Caplan David Chadwick Iain Chalmers Rory Collins Hans van Crevel Richard Doll Geoff Donnan Stuart Douglas Shah Ebrahim Rob Edis Barbara Farrell C. Miller Fisher Chris Foote John Fry Mike Gent Sonny Gubbay Michael Harrison Jim Heron Steff Lewis Bryan Matthews Richard Peto Alex Pollock Geoffrey Rose David Sackett Robin Sellar David Shepherd Jim Slattery Rien Vermeulen Ted Stewart-Wynne Derick Wade Eelco Wijdicks

vii

9781405127660_1_pre.qxd 10/13/07 11:15 AM Page viii

Abbreviations

We don’t care much for abbreviations. They are not literate (Oliver Twist was not abbreviated to OT each time Dickens mentioned his name!), they don’t look good on the printed page, and they make things more difficult to read and understand, particularly for non-experts. But they do save space and so we have to use them a bit. However, we will avoid them as far as we can in tables, figures and the practice points. We will try to define any abbreviations the first time they are used in each chapter, or even in each section if they are not very familiar. But, if we fail to be comprehensible, then here is a rather long list to refer to. ACA ACE AChA ACoA ACST ADC ADH ADL ADP ADPKD AF AFx AH AICA AIDS AMI ANCA ANF APS APT APTT ARAS ARD ASA ASD ATIII

Anterior cerebral artery Angiotensin converting enzyme Anterior choroidal artery Anterior communicating artery Asymptomatic Carotid Surgery Trial Apparent diffusion coefficient Antidiuretic hormone Activities of daily living Adenosine diphosphate Autosomal dominant polycystic kidney disease Atrial fibrillation Amaurosis fugax Ataxic hemiparesis Anterior inferior cerebellar artery Acquired immune deficiency syndrome Acute myocardial infarction Antineutrophil cytoplasmic antibody Antinuclear factor Antiphospholipid syndrome Antiplatelet Trialists’ Collaboration Activated partial thromboplastin time Ascending reticular activating system Absolute risk difference Atrial septal aneurysm Atrial septal defect Antithrombin III

ATP ATT AVF AVM BA BIH BMI BP C CAA CADASIL

Adenosine triphosphate Antithrombotic Trialists’ Collaboration Arteriovenous fistula Arteriovenous malformation Basilar artery Benign intracranial hypertension Body mass index Blood pressure Celsius Cerebral amyloid angiopathy Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy CAST Chinese Acute Stroke Trial CAVATAS Carotid and Vertebral Artery Transluminal Angioplasty Study CBF Cerebral blood flow CBFV Cerebral blood flow velocity CBV Cerebral blood volume CCA Common carotid artery CEA Carotid endarterectomy CHD Coronary heart disease CI Confidence interval CK Creatine kinase CMRO2 Cerebral metabolic rate of oxygen CMRglu Cerebral metabolic rate of glucose CNS Central nervous system CPP Cerebral perfusion pressure CPSP Central post-stroke pain CSF Cerebrospinal fluid CT Computed tomography CTA Computed tomographic angiography CVR Cerebrovascular resistance DBP Diastolic blood pressure DCHS Dysarthria clumsy-hand syndrome DIC Disseminated intravascular coagulation DNA Deoxyribose nucleic acid DSA Digital subtraction angiography DSM Diagnostic and statistical manual of mental disorders

viii

9781405127660_1_pre.qxd 10/13/07 11:15 AM Page ix

Abbreviations

DVT DWI EACA EADL EAFT ECA ECASS ECG EC-IC ECST EEG EMG ESR FDA FIM FLAIR FMD fMRI FMZ GCS GEF GKI HACP Hg HITS HIV HMPAO HTI HU IAA IAA ICA ICH ICIDH ICP ICVT IADSA INR IST IVDSA IVM kPa L LACI LACS LGN LP LSA M MAC

Deep venous thrombosis (in the legs or pelvis) Diffusion weighted (MR) imaging Epsilon-aminocaproic acid Extended activities of daily living European Atrial Fibrillation Trial External carotid artery European Cooperative Acute Stroke Study Electrocardiogram Extracranial-intracranial European Carotid Surgery Trial Electroencephalogram Electromyography Erythrocyte sedimentation rate Food and Drug Administration Functional independence measure Fluid attenuated inversion recovery Fibromuscular dysplasia Functional magnetic resonance imaging Flumazenil Glasgow Coma Scale Glucose extraction fraction Glucose, potassium and insulin hom*olateral ataxia and crural paresis Mercury High intensity transient signals Human immunodeficiency virus Hexamethylpropyleneamine oxime Haemorrhagic transformation of an infarct Hounsfield units Internal auditory artery Intra arterial angiography Internal carotid artery Intracerebral haemorrhage International classification of impairments, disabilities and handicaps Intracranial pressure Intracranial venous thrombosis Intra-arterial digital subtraction angiography International normalized ratio International Stroke Trial Intravenous digital subtraction angiography Intracranial vascular malformation Kilopascals Litre Lacunar infarction Lacunar syndrome Lateral geniculate nucleus Lumbar puncture Lenticulostriate artery Molar Mitral annulus calcification

MAOI MAST-I MCA MCTT MES MI MLF MLP MMSE MR MRA MRC MRI MRS MRV MTT NAA NASCET NELH NG NIHSS NINDS NNT NO OCSP OEF OHS OR PACI PaCO2 PaO2 PACS PCA PChA PCoA PCV PE PEG PET PFO PICA PMS PNH POCI POCS PD PSE PSS PT PTA PVD

Monoamine oxidase inhibitor Multicentre Acute Stroke Trial – Italy Middle cerebral artery Mean cerebral transit time Microembolic signals Myocardial infarction Medial longitudinal fasciculus Mitral leaflet prolapse Mini mental state examination Magnetic resonance Magnetic resonance angiography Medical Research Council Magnetic resonance imaging Magnetic resonance spectroscopy Magnetic resonance venogram Mean transit time N-acetylaspartate North American Symptomatic Carotid Endarterectomy Trial National Electronic Library for Health Nasogastric National Institute of Health Stroke Score National Institute of Neurological Disorders and Stroke Number-needed-to-treat Nitric oxide Oxfordshire Community Stroke Project Oxygen extraction fraction Oxford Handicap Scale Odds ratio Partial anterior circulation infarction Arterial partial pressure of carbon dioxide Arterial partial pressure of oxygen Partial anterior circulation syndrome Posterior cerebral artery Posterior choroidal artery Posterior communicating artery Packed cell volume Pulmonary embolism Percutaneous endoscopic gastrostomy Positron emission tomography Patent foramen ovale Posterior inferior cerebellar artery Pure motor stroke Paroxysmal nocturnal haemoglobinuria Posterior circulation infarction Posterior circulation syndrome Proton density Present state examination Pure sensory stroke Prothrombin time Percutaneous transluminal angioplasty Peripheral vascular disease

9781405127660_1_pre.qxd 10/13/07 11:15 AM Page x

Abbreviations

PWI QALYs RAH RCT RIND RNA ROR RR RRR rt-PA SADS SAH SBP SCA SD SEPIVAC SF36 SIADH SK

Perfusion weighted (MR) imaging Quality adjusted life years Recurrent artery of Heubner Randomized controlled trial Reversible ischaemic neurological deficit Ribonucleic acid Relative odds reduction Relative risk Relative risk reduction Recombinant tissue plasminogen activator Schedule for affective disorders and schizophrenia Subarachnoid haemorrhage Systolic blood pressure Superior cerebellar artery Standard deviation Studio epidemiologico sulla incidenza delle vasculopathie acute cerebrali Short form 36 Syndrome of inappropriate secretion of antidiuretic hormone Streptokinase

SLE SMS SPAF SPECT SVD TACI TACS TCD TEA TENS TGA TIA TMB TOAST TTP TTP US VA VB WHO WFNS

Systemic lupus erythematosus Sensorimotor stroke Stroke prevention in atrial fibrillation (trial) Single photon emission computed tomography Small vessel disease Total anterior circulation infarction Total anterior circulation syndrome Transcranial Doppler Tranexamic acid Transcutaneous electrical nerve stimulation Transient global amnesia Transient ischaemic attack Transient monocular blindness Trial of ORG 10172 in Acute Stroke Therapy Thrombotic thrombocytopenic purpura Time to peak Ultrasound Vertebral artery Vertebrobasilar World Health Organization World Federation of Neurological Surgeons

9781405127660_4_001.qxd 10/13/07 11:15 AM Page 1

Introduction 1.1 Introduction to the first edition

1.2 Introduction to the second edition 1.3 Introduction to the third edition

1.1 Introduction to the first edition

1.1.1 Aims and scope of the book We, the authors of this book, regard ourselves as practising – and practical – doctors who look after stroke patients in very routine day-to-day practice. The book is for people like us: neurologists, geriatricians, stroke physicians, radiologists and general internal physicians. But it is not just for doctors. It is also for nurses, therapists, managers and anyone else who wants practical guidance about all and any of the problems to do with stroke – from aetiology to organization of services, from prevention to occupational therapy, and from any facet of cure to any facet of care. In other words, it is for anyone who has to deal with stroke in clinical practice. It is not a book for armchair theoreticians, who usually have no sense of proportion as well as difficulty in seeing the wood from the trees. Or, maybe, it is particularly for them so that they can be led back into the real world. The book takes what is known as a problem-orientated approach. The problems posed by stroke patients are discussed in the sort of order that they are likely to present themselves. Is it a stroke? What sort of stroke is it? What caused it? What can be done about it? How can the patient and carer be supported in the short term and long term? How can any recurrence be prevented? How can stroke services be better organized? Unlike traditional textbooks, which linger on dusty shelves, there are no ‘-ology’ chapters. Aetiology, epidemiology, pathology

Stroke: practical management, 3rd edition. C. Warlow, J. van Gijn, M. Dennis, J. Wardlaw, J. Bamford, G. Hankey, P. Sanderco*ck, G. Rinkel, P. Langhorne, C. Sudlow and P. Rothwell. Published 2008 Blackwell Publishing. ISBN 978-1-4051-2766-0.

and the rest represent just the tools to solve the problems – so they are used when they are needed, and not discussed in isolation. For example, to prevent strokes one needs to know how frequent they are (epidemiology), what types of stroke there are (pathology), what causes them (aetiology) and what evidence there is to support therapeutic intervention (randomized controlled trials). Clinicians mostly operate on a need-to-know basis, and so when a problem arises they need the information to solve it at that moment, from inside their head, from a colleague – and we hope from a book like this.

1.1.2 General principles To solve a problem one obviously needs relevant information. Clinicians, and others, should not be making decisions based on whim, dogma or the last case, although most do, at least some of the time – ourselves included. It is better to search out the reliable information based on some reasonable criterion for what is meant by reliable, get it into a sensible order, review it and make a summary that can be used at the bedside. If one does not have the time to do this – and who does for every problem? – then one has to search out someone else’s systematic review. Or find the answer in this book. Good clinicians have always done all this intuitively, although recently the process has been blessed with the title of ‘evidencebased medicine’, and now even ‘evidence-based patientfocused medicine’! In this book we have used the evidence-based approach, at least where it is possible to do so. Therefore, where a systematic review of a risk factor or a treatment is available we have cited it, and not just emphasized single studies done by us or our friends and with results to suit our prejudices. But so often there is no good evidence or even any evidence at all available, and certainly no systematic reviews. What to do then? Certainly not what most doctors are trained to do: ‘Never be wrong, and if you are, never admit it!’ If we do 1

9781405127660_4_001.qxd 10/13/07 11:15 AM Page 2

Chapter 1 Introduction

not know something, we will say so. But, like other clinicians, we may have to make decisions even when we do not know what to do, and when nobody else does either. One cannot always adopt the policy of ‘if you don’t know what to do, don’t do it’. Throughout the book we will try to indicate where there is no evidence, or feeble evidence, and describe what we do and will continue to do until better evidence becomes available; after all, it is these murky areas of practice that need to be flagged up as requiring further research. Moreover, in clinical practice, all of us ask respected colleagues for advice, not because they may know something that we do not but because we want to know what they would do in a difficult situation.

1.1.3 Methods We were all taught to look at the ‘methods’ section of a scientific paper before anything else. If the methods are no good, then there is no point in wasting time and reading further. In passing, we do regard it as most peculiar that some medical journals still print the methods section in smaller letters than the rest of the paper. Therefore, before anyone reads further, perhaps we should describe the methods we have adopted. It is now impossible for any single person to write a comprehensive book about stroke that has the feel of having been written by someone with hands-on experience of the whole subject. The range of problems is far too wide. Therefore, the sort of stroke book that we as practitioners want – and we hope others do too – has to be written by a group of people. Rather than putting together a huge multiauthor book, we thought it would be better and more informative, for ourselves as well as readers, to write a book together that would take a particular approach (evidence-based, if you will) and end up with a coherent message. After all, we have all worked together over many years, our views on stroke are more convergent than divergent, and so it should not be too terribly difficult to write a book together. Like many things in medicine, and in life, this book started over a few drinks to provide the initial momentum to get going, on the occasion of a stroke conference in Geneva in 1993. At that time, we decided that the book was to be comprehensive (but not to the extent of citing every known reference), that all areas of stroke must be covered, and who was going to start writing which section. A few months later, the first drafts were then commented on in writing and in detail by all the authors before we got back together for a general discussion – again over a few drinks, but on this occasion at the Stockholm stroke conference in 1994. Momentum restored, we went home to improve what we had written,

and the second draft was sent round to everyone for comments in an attempt to improve the clarity, remove duplication, fill in gaps and expunge as much remaining neurodogma, neurofantasy and neuroastrology as possible. Our final discussion was held at the Bordeaux stroke meeting in 1995, and the drinks that time were more in relief and celebration that the end was in sight. Home we all went to update the manuscript and make final improvements before handing over the whole lot to the publisher in January 1996. This process may well have taken longer than a conventional multiauthor book in which all the sections are written in isolation. But it was surely more fun, and hopefully the result will provide a uniform and coherent view of the subject. It is, we hope, a ‘how to do it’ book, or at least a ‘how we do it’ book.

1.1.4 Using the book This is not a stroke encyclopaedia. Many very much more comprehensive books and monographs are available now, or soon will be. Nor is this really a book to be read from cover to cover. Rather, it is a book that we would like to be used on stroke units and in clinics to help illuminate stroke management at various different stages, both at the level of the individual patient and for patients in general. So we would like it to be kept handy and referred to when a problem crops up: how should swallowing difficulties be identified and managed? Should an angiogram be done? Is raised plasma fibrinogen a cause of stroke? How many beds should a stroke unit have? And so on. If a question is not addressed at all, then we would like to know about it so that it can be dealt with in the next edition, if there is to be one, which will clearly depend on sales, the publisher, and enough congenial European stroke conferences to keep us going. It should be fairly easy to find one’s way around the book from the chapter headings and the contents list at the beginning of each chapter. If that fails, then the index will do instead. We have used a lot of crossreferencing to guide the reader from any starting point and so avoid constant reference to the index. As mentioned earlier, we have tried to be as selective as possible with the referencing. On the one hand, we want to allow readers access to the relevant literature, but on the other hand we do not want the text to be overwhelmed by references – particularly by references to unsound work. To be selective, we have tried to cite recent evidence-based systematic reviews and classic papers describing important work. Other references can probably mostly be found by those who want to dig deeper in the reference lists of the references we have cited.

9781405127660_4_001.qxd 10/13/07 11:15 AM Page 3

1.2 Introduction to the second edition

Finally, we have liberally scattered what some would call practice points and other maxims throughout the book. These we are all prepared to sign up to, at least in early 1996. Of course, as more evidence becomes available, some of these practice points will become out of date.

1.1.5 Why a stroke book now? Stroke has been somewhat of a Cinderella area of medicine, at least with respect to the other two of the three most common fatal disorders in the developed world – coronary heart disease and cancer. But times are gradually changing, particularly in the last decade when stroke has been moving up the political agenda, when research has been expanding perhaps in the slipstream of coronary heart disease research, when treatments to prevent, if not treat, stroke have become available and when the pharmaceutical industry has taken more notice. It seems that there is so much information about stroke that many practitioners are beginning to be overwhelmed. Therefore, now is a good time to try to capture all this information, digest it and then write down a practical approach to stroke management based on the best available evidence and research. This is our excuse for putting together what we know and what we do not know, what we do and why we do it.

1.2 Introduction to the second edition

Whether we enjoyed our annual ‘stroke book’ dinners at the European stroke conferences too much to abandon them, or whether we thought there really was a lot of updating to do, we found ourselves working on this second edition four short years after the first. It has certainly helped to have been so much encouraged by the many people who seemed to like the book, and find it useful. We have kept to the same format, authors, and principles outlined above in the introduction to the first edition. The first step was for all of us to read the whole book again and collect together any new comments and criticisms for each of the other authors. We then rewrote our respective sections and circulated them to all the other authors for their further comments (and they were not shy in giving them). We prepared our final words in early 2000. A huge technical advance since writing the first edition has been the widespread availability of e-mail and the use of the Internet. Even more than before, we have

genuinely been able to write material together; one author does a first draft, sends it as an attachment across the world in seconds, the other author appends ideas and e-mails the whole attachment back to the first author, copying to other authors for comments perhaps, and so on until it is perfect. Of course, we still do not all agree about absolutely everything all of the time. After all, we want readers to have a feel for the rough and ragged growing edge of stroke research, where there is bound to be disagreement. If we all knew what to do for stroke patients there would be no need for randomized controlled trials to help us do better – an unrealistic scenario if ever there was one. So where there is uncertainty, and where we disagree, we have tried to make that plain. But, on the whole, we are all still prepared to sign up to the practice points. In this second edition, we have been able to correct the surprising number of minor typographical errors and hope not to have introduced any more, get all the X-rays the right way up, improve on some of the figures, remove some duplication, reorder a few sections, put in some more subheadings to guide the readers, make the section on acute ischaemic stroke more directive, improve the index, and generally tidy the whole thing up. It should now be easier to keep track of intracranial venous thrombosis and, in response to criticism, we have extended the section on leukoaraiosis, even though it is not strictly either a cause or a consequence of stroke. We have also introduced citations to what we have called ‘floating references’ – in other words, published work that is constantly being changed and updated as new information becomes available. An obvious example is the Cochrane Library, which is updated every 3 months and available on CD-ROM and through the Internet. There are no page numbers, and the year of publication is always the present one. We have therefore cited such ‘floating references’ as being in the present year, 2000. But we know that this book will not be read much until the year 2001 and subsequent years, when readers will have to look at the contemporary Cochrane Library, not the one published in 2000. The same applies to the new British Medical Journal series called ‘Clinical Evidence’ which is being updated every 6 months, and to any websites that may be updated at varying intervals and are still very much worth directing readers towards. Rather to our surprise, there is a lot of new information to get across on stroke. Compared with 4 years ago, the concept of organized stroke services staffed by experts in stroke care has taken root and has allowed the increasingly rapid assessment of patients with ‘brain attacks’. It is no longer good enough to sit around waiting 24 h or more to see if a patient is going to have a transient ischaemic attack or a stroke, and then another 24 h for a ·

9781405127660_4_001.qxd 10/13/07 11:15 AM Page 4

Chapter 1 Introduction

computed tomography brain scan to exclude intracerebral haemorrhage. These days we have to assess and scan stroke patients as soon as they arrive in hospital, perhaps give thrombolysis to a few, and enter many more into clinical trials, start aspirin in ischaemic stroke, and get the multidisciplinary team involved – and all of this well within 24 h of symptom onset. Through the Cochrane Library, which was in its infancy when the first edition was published, there is now easy, regularly updated electronic access to systematic reviews of most of the acute interventions and secondary prevention strategies for stroke, although the evidence base for rehabilitation techniques is lagging behind. Catheter angiography is giving way to non-invasive imaging. Magnetic resonance techniques are racing ahead of the evidence as to how they should be used in routine clinical practice. For better or worse, coiling cerebral aneurysms is replacing clipping. The pharmaceutical industry is still tenaciously hanging on to the hope of ‘neuroprotection’ in acute ischaemic stroke, despite numerous disappointments. Hyperhom*ocysteinaemia and infections are the presently fashionable risk factors for ischaemic stroke, and they may or may not stand the test of time. So, in this second edition, we have tried to capture all these advances – and retreats – and set them in the context of an up-todate understanding of the pathophysiology of stroke and the best available evidence of how to manage it. Of course, it is an impossible task, because something new is always just around the corner. But then ‘breakthroughs’ in medicine take time to mature – maybe years until the evidence becomes unassailable and is gradually accepted by front-line clinicians. And then we can all sit back doing what we believe to be ‘the right thing’ for a few more years until the next ‘breakthrough’ changes our view of the world yet again. We hope that the ideas and recommendations in this book will be sufficient 99% of the time – at least for the next 4 years, when we will have to see about a third edition. ·

1.3 Introduction to the third edition

Six years have gone quickly by since the second edition, much has happened in stroke research and practice in the meantime, and two of the authors are on the edge of retirement – so it is time for this third edition of what we fondly refer to as ‘the book’. Maybe because the original authors were feeling tired, or increasingly unable to cover in depth all we wanted to, or perhaps because we

wanted to ensure our succession, we have recruited four new and younger authors, all of whom have worked closely with us over many years, and whose help we acknowledged in the earlier editions – Gabriel Rinkel, Peter Langhorne, Cathie Sudlow and Peter Rothwell. But, even with their help, the rewriting has had to compete with all the far less interesting things which we have to do these days to satisfy managers, regulatory authorities and others keen to track and measure our every move. And maybe there is less imperative to write books like this which are out of date in at least some ways even before they are published. But then searching the Internet for ‘stroke’ does not come up with a coherent account of the whole subject of managing stroke patients using the best available evidence, which is what this book is all about. So, with the help and encouragement of Blackwell Publishing, here is the third edition of ‘the book’ at last. We have written the book as before with most of the authors commenting on most of the chapters before all the chapters were finally written in the form you can read them in now. Again, you will have to guess who wrote what because we can all lay claim to most of the book in some sense or another. There has been a slight change in the arrangement of the chapters, but loyal readers of the earlier editions will not find this too upsetting – they will still find what they want in more or less its familiar place, and as ever we hope the index has been improved. The practice points we all sign up to and our day-to-day practice should reflect them. The uncertainties we all share – they will be gradually resolved as more research is done, and more uncertainties will then be revealed. The biggest change in this edition is succumbing to the space saving offered by a numbered reference system, and a change in the colour scheme from a pastel green to various shades of purple. As with the second edition, much has changed and there has been more updating than we originally anticipated – what we know about stroke has moved on. Neuroprotection is even less likely to be an effective treatment for ischaemic stroke than it was in the 1990s, we still argue about thrombolysis, clopidogrel cannot very often be recommended, carotid stenting has still to prove its worth, routine evacuation of intracerebral haemorrhage is definitely not a good idea, and hormone replacement therapy far from protecting against vascular disease actually seems to increase the risk. But on the positive side, much has improved in brain and vessel imaging, it is now clear how much blood pressure lowering has to offer in secondary stroke prevention, and cholesterol lowering too. Carotid surgery can now be targeted on the few who really need it, not recommended for the greater number who may or may not need it.

9781405127660_4_001.qxd 10/13/07 11:15 AM Page 5

1.3 Introduction to the third edition

Coiling has more or less replaced clipping of intracranial aneurysms, an astonishing change in practice brought about by a large trial energetically led by an interventional neuroradiologist and neurosurgeon. And it is not just acute stroke that needs urgent attention nowadays, transient ischaemic attacks must be assessed and managed very quickly to minimize the early high risk of stroke. Stroke services continue to improve all over the world, stroke has moved up the political agenda as we have managed to wrench it out of the rubric of ‘cardiovascular’ disease which always emphasized the cardiac rather than the cerebral, and more and more people are involved in stroke research, which is now a much more

crowded and competitive field than it was when some of us started in the 1970s. Will there be a fourth edition? We don’t know; this will be in the hands of the remaining authors as Charles Warlow and Jan van Gijn dwindle into retirement of a sort, or at least a life that will not require the relentless battle to keep up with all the stroke literature, critique it, absorb anything that is worthwhile, and then put it into the context of active clinical practice. No one can write well about stroke unless they can connect research with their own active clinical practice – we are not, we hope, armchair theoreticians; we try to practise what we preach.

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 7

Development of knowledge about cerebrovascular disease 2.1

Ideas change slowly

2.2

The anatomy of the brain and its blood supply

2.3

What happens in ‘apoplexy’?

2.4

Cerebral infarction (ischaemic stroke)

2.5

Thrombosis and embolism

2.6

Transient ischaemic attacks

2.7

Intracerebral haemorrhage

2.8

Subarachnoid haemorrhage

2.9

Treatment and its pitfalls

10 14

2.10 Epilogue 28

‘Our knowledge of disorders of the cerebral circulation and its manifestations is deficient in all aspects’ was the opening sentence of the chapter on cerebrovascular diseases in Oppenheim’s textbook of neurology at the beginning of the 20th century.1 More than 90 years later this still holds true, despite the considerable advances that have been made. In fact, the main reason for Oppenheim’s lament, the limitations of pathological anatomy, is to some extent still valid. True, our methods of observation nowadays are no longer confined to the dead, as they were then. They have been greatly expanded, first by angiography, then by brain imaging and measurement of cerebral blood flow and metabolism, and most recently by non-invasive methods of vascular imaging such as ultrasound and magnetic resonance angiography. Yet, our observations are still mostly anatomical, and after the event. It is only in rare instances that are we able to reconstruct the dynamics of a stroke. At least in haemorrhagic stroke, brain computed tomography (CT) or magnetic resonance imaging (MRI) in the acute phase gives an approximate indication of where a blood vessel has ruptured (though not why exactly there and then) and how far the extravasated blood has invaded the brain parenchyma or the subarachnoid space. With ischaemic stroke, the growth of our understanding has been slower. The ubiquity of

the term ‘cerebral thrombosis’ up to the 1970s exemplifies how deficient our understanding was even at that time.2 Embolic occlusion, now known to result more often from arterial lesions than from the heart, can be detected in an early phase by non-invasive angiographic techniques or inferred by means of perfusion imaging, but so often the source of the clot is still elusive. We have also learned to distinguish many causes of cerebral infarction other than atherothrombosis, such as arterial dissection, mitochondrial cytopathies and moyamoya syndrome, but the precise pathogenesis of these conditions is still poorly understood. So it is with humility, rather than in triumph, that we look back on the past. In each era the problems of stroke have been approached by the best minds, with the best tools available. Of course many ideas in the past were wrong, and so presumably are many of our own. Even though we are firm believers in evidence-based medicine, some – perhaps many or even most – of our own notions will not survive the test of time. Our knowledge may have vastly increased in the recent past but it is still a mere island in an ocean of ignorance.

2.1 Ideas change slowly Stroke: practical management, 3rd edition. C. Warlow, J. van Gijn, M. Dennis, J. Wardlaw, J. Bamford, G. Hankey, P. Sanderco*ck, G. Rinkel, P. Langhorne, C. Sudlow and P. Rothwell. Published 2008 Blackwell Publishing. ISBN 978-1-4051-2766-0.

The history of medicine, like that of kings and queens in world history, is usually described by a string of dates 7

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 8

Chapter 2 Development of knowledge about cerebrovascular disease

and names, by which we leapfrog from one discovery to another. The interval between such identifiable advances is measured in centuries when we describe the art of medicine at the beginning of civilization, but in mere years where our present times are chronicled. This leads to the impression that we are witnessing a dazzling explosion of knowledge. Some qualification of this view is needed, however. First of all, any generation of mankind takes a myopic view of history in that the importance of recent developments is overestimated. The Swedish Academy of Sciences therefore often waits for years, sometimes even decades, before awarding Nobel prizes, until scientific discoveries have withstood the test of time. When exceptions were made for the prize in medicine, the early accolades were not always borne out: WagnerJauregg’s malaria treatment for neurosyphilis (1927) is no longer regarded as a landmark, while Moniz’s prize (1949) for prefrontal leucotomy no longer seems justified; at least he also introduced contrast angiography of the brain, although this procedure may again not survive beyond the end of this century. We can only hope that the introduction of X-ray CT by Hounsfield (Nobel prize for medicine in 1979) will be judged equally momentous by future generations as by ourselves. Another important caveat if one looks back on progress in medicine is that most discoveries gain ground only slowly. Even if new insights were quickly accepted by peer scientists, which was often not the case, it could still be decades before these had trickled down to the rank and file of medical practitioners. The mention of a certain date for a discovery may create the false impression that this change in medical thinking occurred almost overnight, like the introduction of the single European currency. In most instances, this was far from the truth. An apt example is the extremely slow rate at which the concept of lacunar infarction became accepted by the medical community, despite its potentially profound implications in terms of pathophysiology, treatment and prognosis. The first pathological descriptions date from around 1840,3,4 but it took the clinicopathological correlations of C. Miller Fisher (Fig. 2.7) in the 1960s before the neurological community and its textbooks started to take any notice.5–7 And it was not until new techniques for brain imaging in the 1980s provided instantaneous clinicoanatomical correlations that no practising neurologist could avoid knowing about lacunar infarcts – some 150 years after the first description! It is best to become reconciled to the idea that a slow rate of diffusion of new knowledge is unavoidable. The problem is one of all times. Franciscus Biumi, one of the early pathologists, lamented in 1765: ‘Sed difficile est adultis novas opiniones inserere, evellere insitas’ (But it is difficult to insert new opinions in adults and to remove rooted ones).8 How slowly new ideas were accepted and acted upon, against

the background of contemporary knowledge, can often be inferred from textbooks, particularly if written by fulltime clinicians rather than by research-minded neurologists. Therefore we shall occasionally quote old textbooks to illustrate the development of thinking about stroke. A reverse problem is that a new discovery or even a new fashion may be interpreted beyond its proper limits and linger on as a distorted idea for decades. Take the discovery of vitamin B1 deficiency as the cause of a tropical polyneuropathy almost a century ago; the notion that a neurological condition, considered untreatable almost by definition, could be cured by a simple nutritional supplement made such an impact on the medical community that even in some industrialized countries vitamin B1 is still widely used as a panacea for almost any neurological symptom. So broadly speaking there are two kinds of medical history, that of the cutting edge of research and that of the medical profession as a whole. The landmarks are easy to identify only with the hindsight of present knowledge. In reality, new ideas often only gradually dawned on consecutive scientists, instead of the popular notion of a blinding flash of inspiration occurring in a single individual. For this reason, accounts of the history of stroke are not always identical.9,10 Also many important primary sources are not easy to interpret – not only because they were written in Latin, but also because ‘new observations’ have sometimes been identified only by later historians, in retrospect, while the authors at the time attached no importance to them.11

2.2 The anatomy of the brain and its blood supply

From at least the time of Hippocrates (460–370 BC), the brain was credited with intelligence and thought, and also with movements of the opposite side of the body, through observation of unilateral convulsions after head wounds on the contralateral side.12 Yet, stroke, or ‘apoplexy’ (Greek for ‘being struck down’), was defined as a sudden but mostly general, rather than focal, disorder of the brain. The pathogenesis was explained according to the humoral theory, which assumed a delicate balance between the four humours: blood, phlegm, black bile and yellow bile. Anatomy played almost no part in these explanations. Apoplexy was often attributed to accumulation of black bile in the arteries of the brain, obstructing the passage of animated spirits from the ventricles.13 Galenus of Pergamon (131– 201), a prolific writer and animal experimenter whose

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 9

2.2 The anatomy of the brain and its blood supply

views dominated medicine up to the 17th century,14 distinguished ‘karos’ from ‘apoplexy’, in that respiration was unaffected in the former condition.15 Leading Islamic physicians like Avicenna (980–1037) tried to reconcile Galenic tenets with the Aristotelian view of the heart as the seat of the mind.16 In Western Europe, mostly deprived of Greek learning until the fall of Constantinople in 1453 prompted the Renaissance,17 these Arabic texts were translated into Latin before those of Galen and Hippocrates.18 All these theories had no anatomical counterpart; dissection of the human body was precluded by its divine connotations. Any illustrations of the human brain that are known before the 16th century are crude and schematic representations of Galenic theories, rather than attempts at copying

Fig. 2.1 Plate depicting the blood vessels, from Vesalius’s Tabulae Anatomicae Sex, of 1538.21 This shows the carotid arteries ending up in a network (b) at the base of the brain; the structures marked (a) represent the choroid plexus in the lateral ventricles. The network of blood vessels (rete mirabile) is found in oxen; Galen had assumed it was found also in the human brain, a belief perpetuated throughout the Dark and Middle Ages, up to the early Renaissance. Leonardo da Vinci had also drawn a (human?) brain with a ‘rete mirabile’ at its base.22 Vesalius retracted the existence of a network in his atlas of 1543.

the forms of nature. As a consequence, many nonneurological disease conditions with sudden onset must have been misclassified as ‘apoplexy’. In 1543 Andries van Wesele (1514–1564), the great Renaissance anatomist who Latinized his name to Andreas Vesalius, produced the first accurate drawings of the brain in his famous book De humani corporis fabrica libri septem, with the help of the draughtsman Johan Stephaan van Calcar and the printer Oporinus in Basle.19 It was the same year in which Copernicus published De revolutionibus, proclaiming the sun and not the earth as the centre of the universe.20 Vesalius largely ignored the blood vessels of the brain, although he retracted an earlier drawing (Fig. 2.1) depicting a ‘rete mirabile’, a network of blood vessels at the base of the brain that

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 10

Chapter 2 Development of knowledge about cerebrovascular disease

Galen had found in pigs and oxen and that had been extrapolated to the human brain ever since.21,22 Before him, Berengario da Carpi had also denied the existence of the rete.23 Vesalius was vehemently attacked by traditionally minded contemporaries as an iconoclast of Galenic dogmas. Nevertheless, initially, he did not go as far as outright opposition to the central Galenic tenet that blood could pass through the septum between the right and left ventricle of the heart, allowing the mixture of blood and air and the elimination of ‘soot’. Instead, he praised the creator for having made the openings so small that nobody could detect them, another striking example of how the power of theory may mislead even the most inquisitive minds. Only later, in the 1555 edition of his De humani corporis fabrica, did he firmly state that the interventricular septum was tightly closed. The decisive blow to the humoral theory came in 1628, through the description of the circulation by William Harvey (1578–1657);24 it need no longer surprise us that it took many decades before these views were widely accepted. Harvey’s work formed the foundation for the recognition of the role of blood vessels in the pathogenesis of stroke. Thomas Willis (1641–1675) is remembered not so much for having coined the term ‘neurology’, or for his iatrochemical theories, a modernized version of humoral medicine, or for his part in the successful resuscitation of Ann Green after judicial hanging,25 as he is for his work on the anatomy of the brain, first published in 1664,26 especially for his description of the vascular interconnections at the base of the brain (Fig. 2.2).27 Before him, Fallopius, Casserio, Vesling and Wepfer had all observed at least part of the circle,28–31 in the case of Casserio and Vesling even with an illustration.32 But undisputedly, it was Willis who grasped the functional implications of these anastomoses in a passage illustrating his proficiency in performing necropsies as well as postmortem experiments (from a posthumous translation):33 We have elsewhere shewed, that the Cephalick Arteries, viz. the Carotides, and the Vertebrals, do so communicate with one another, and all of them in different places, are so ingraffed one in another mutually, that if it happen, that many of them should be stopped or pressed together at once, yet the blood being admitted to the Head, by the passage of one Artery only, either the Carotid or the Vertebral, it would presently pass thorow all those parts exterior and interior: which indeed we have sufficiently proved by an experiment, for that Ink being squirted in the trunk of one Vessel, quickly filled all the sanguiferous passages, and every where stained the Brain it self. I once opened the dead Carcase of one wasted away, in which the right

Fig. 2.2 Illustration of the base of the brain from Willis’s Cerebri Anatome (1664),26 showing the interconnections between the right and left carotid systems, and also between these two and the posterior circulation (drawing by Christopher Wren).

Arteries, both the Carotid and the Vertebral, within the Skull, were become bony and impervious, and did shut forth the blood from that side, notwithstanding the sick person was not troubled with the astonishing Disease. It seems that the idea of infusing coloured liquids into blood vessels, practised from 1659 onwards and later perfected by Frederik Ruysch (1638–1731) and in the next century by John Hunter (1728–1793),34,35 had come from Christopher Wren (1632–1723).25 Wren also made the etchings for Willis’s book (he is now mainly remembered as the architect of St Paul’s Cathedral and many other churches built after the great fire of London in 1666).

2.3 What happens in ‘apoplexy’?

Willis’s ‘astonishing Disease’, apoplexy, had of old intuitively been attributed to some ill-defined obstruction, whether from want of ‘animal spirits’ via the nerves in

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 11

2.3 What happens in ‘apoplexy’?

the tradition of Greek medicine, or, after Harvey’s time, by deprivation of blood flow. Yet, it should be remembered that the notion of an intrinsic ‘nervous energy’ only slowly lost ground. Even the great 18th-century physician Boerhaave, though clearly recognizing the role of blood vessels and the heart in the development of apoplexy, invoked obstruction of the cerebrospinal fluid.36 In Table 2.1 we have provided a schematic representation of the development of ideas about apoplexy through the ages, together with its relationship to arterial lesions. That Willis had found ‘bony’ and ‘impervious’ arteries in patients who actually had not died from a stroke was probably the reason that he was not outspoken on the pathogenesis of apoplexy. His contemporaries, Wepfer (1620–1695) in Schaffhausen, and Bayle (1622–1709) in Toulouse, only tentatively associated apoplexy with ‘corpora fibrosa’,31 or with calcification of cerebral arteries.37 Wepfer (Fig. 2.3) not only recognized arterial lesions, but he also prompted one of the great advances in the knowledge about stroke by distinguishing between, on the one hand, arterial obstruction preventing the influx

Fig. 2.3 Johann Jakob Wepfer (1620–1695).

of blood and, on the other, extravasation of blood into the substance of the brain or the ventricular cavities. His interpretation was, however, that blockage of arteries as well as extravasation of blood impeded the transmission of ‘spiritus animalis’ to the brain.11 Accordingly, he regarded apoplexy as a process of global stunning of the brain, while the focal nature of the disease largely escaped him. The four cases of haemorrhage Wepfer described were massive, at the base of the brain or deep in the parenchyma. In cases with obvious hemiplegia, incidentally a term dating back to the Byzantine physician Paulus Aegineta (625–690),38 Wepfer suspected dysfunction of the ipsilateral rather than the contralateral side. He also observed patients who had recovered from apoplectic attacks, and noted that those most liable to apoplexy were ‘the obese, those whose face and hands are livid, and those whose pulse is constantly unequal’. That the paralysis was on the opposite side of the apoplectic lesion was clearly predicted by Domenico Mistichelli (1675–1715) from Pisa on the basis of his observation of the decussation of the pyramids (Fig. 2.4).39 A landmark in the recognition of the anatomical substrate of stroke – and of many other diseases – was the work of Morgagni (1682–1771), professor of medicine and subsequently of pathological anatomy in Padua. In 1761 Morgagni published an impressive series of clinicopathological observations collected over a lifetime (he

Fig. 2.4 Illustration from Mistichelli’s book on apoplexy (1709) in which he shows the decussation of the pyramids and also the outward rotation of the leg on the paralysed side.39

Non-haemorrhagic

Rostan (Paris) (1790–1866)57

‘Ramollissem*nt’ (1820): – softening more frequent than haemorrhage – condition not inflammatory?

Rostan

Baillie (London) (1761–1823)42

Morgagni (Padua) (1682–1771)40 ‘Sanguineous apoplexy’ (1761)

Boerhaave

Willis (Oxford) (1621–1675)33

Boerhaave (Leiden) ‘Stoppage of the spirits’ (1668–1738)36 ‘Serous apoplexy’, extravasation of serum? (1761)

Paralysis is unilateral, and crossed with respect to lesion (1709)

Ossification of cerebral arteries (1820)

Hardening of arteries associated with haemorrhage? (1795)

Narrowing due to cartilaginous change (1735)

‘Bony, impervious arteries’ (1684)

1815 Battle of Waterloo; Schubert writes Erlkönig

1776 US Declaration of Independence

1729 Bach writes St Matthew Passion

1707 Union between England and Scotland

1682 Peter I ascends Russian throne

Mistichelli (Pisa) (1675–1715)39

Calcifications (1677)

Bayle (Toulouse) (1622–1709)37

1642 Rembrandt paints Night Watch

‘Corpora fibrosa’

Wepfer (1658)

Extravasation of blood in brain tissue (1658)

Historical events

Wepfer (Schaffhausen) (1620–1695)31

Medical scientist Observations on arterial lesions

0 Birth of Jesus Christ

Sudden loss of consciousness, as a result of brain disease

Haemorrhagic

Ideas about ‘apoplexy’

Galenus Sudden loss of consciousness, as a result of brain disease (Pergamum and Rome) (131–201)15

Hippocrates (Kos) (460–370 BC)13

Medical scientist

Table 2.1 Development of ideas about ‘apoplexy’ and its relationship with arterial lesions.

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 12

Chapter 2 Development of knowledge about cerebrovascular disease

Cerebral softening caused by obliteration of arteries? (one of possible causes; 1838)

Carswell (London) (1793–1857)62

1837 Queen Victoria ascends the throne of the British Empire

Due to ossification of arteries?

1869 Opening of the Suez Canal 1871 Stanley meets Livingstone at Ujiji 1877 Bell invents telephone, Edison the phonograph 1895 Röntgen discovers X-rays in Würzburg 1907 Ehrlich introduces arsphenamine as treatment for syphilis

Arteriosclerosis leads to thrombosis; thrombi may be torn off and lodge distally (‘embolism’) (1856) End-arteries most vulnerable; paradoxical embolism Thrombosis at the carotid bifurcation may cause secondary embolization to brain (1905)

Cerebral softening caused by capillary Virchow congestion, secondary to ‘irritation’ (1842)

Cohnheim

Chiari (Prague) (1851–1916)70

‘Yellow softening’ of the brain is secondary to arterial obliteration; any inflammation is secondary (1856) ‘Infarction’ (stuffing) is haemorrhagic by definition, as opposed to ischaemic necrosis (1872)

Virchow (Berlin) (1821–1902)66

Cohnheim (Berlin) (1839–1884)68

1859 Darwin publishes The Origin of Species 1863 Manet paints Le Déjeuner sur l’herbe

1848 Year of revolutions; Louis Napoleon elected president of France

1829 Stephenson builds the railway engine called ‘The Rocket’

‘Arteriosclerosis’ (1829)

Cruveilhier (Paris) (1791–1874)63

‘Encephalomalacia’ (1844): – white, or serous (congestion) – red (inflammatory) – yellow (frequent; unexplained)

Abercrombie

Cerebral softening analogous to gangrene of limb? (1836)

Abercrombie (Edinburgh) (1780–1844)60

Rokitansky (Vienna) (1804–1878)219

Lobstein (Strasbourg) (1777–1835)67

Cerebral softening is definitely inflammatory in nature (1824)

Lallemand (Montpelier) (1790–1853)58

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 13

2.3 What happens in ‘apoplexy’? 13

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 14

Chapter 2 Development of knowledge about cerebrovascular disease

was 79 at the time of publication), in which he firmly put an end to the era of systemic (humoral) theories of disease and replaced them by an organ-based approach, though he did not include even a single illustration; characteristically, the title of the book was ‘De sedibus et causis morborum . . .’ (about the sites and causes of disease).40 Morgagni not only confirmed the notion of crossed paralysis but also firmly divided apoplexy into ‘sanguineous apoplexy’ and ‘serous apoplexy’ (and a third form which was neither serous nor sanguineous). A decade later, Portal (1742–1832) rightly emphasized that it was impossible to distinguish between these two forms during life.41 However, it would be a mistake to assume that ‘serous’ (non-haemorrhagic) apoplexy was recognized at that time as being the result of impaired blood flow, let alone of mechanical obstruction of blood vessels. Some even linked the arterial hardening with brain haemorrhages and not with the serous apoplexies.42 Although we quoted 17th-century scientists such as Bayle and Wepfer in that they associated some non-haemorrhagic cases of apoplexy with obstruction of blood flow, in the 18th century medical opinion swayed towards ‘vascular congestion’, a kind of pre-haemorrhagic state. That explanation was propounded not only by Morgagni40 but also by many of his contemporaries and followers.41,43,44 John Cheyne (1777–1836) pointed out that autopsy in patients who had survived a ‘stroke of apoplexy’ for a considerable time might show a cavity filled with rusty serum that stained the adjacent brain tissue, but he may have been describing a residual lesion after cerebral haemorrhage rather than infarction.45 The anatomical, organ-based approach exemplified by Morgagni reflected the Italian practice, in which the separation between physicians and surgeons was much less strict than in northern Europe with its more theoretical framework of medicine. The protagonists of the Northern school of thinking were Herman Boerhaave (1668–1738) in Leiden and later William Cullen (1710– 1790) in Edinburgh, the most influential clinical teachers of their time. They established a nosological classification that was based much more on holistic theory, in terms of a disturbed system, than on actual observations at the level of the organ, at least with 20th-century hindsight.46 Probably our own time will be branded as the era of exaggerated reductionism! In the intellectual tradition of the Dutch-Scottish school, purely clinical classifications of apoplexy were proposed in the early 19th century by Serres (with and without paralysis),47 by Abercrombie (primary apoplexy, with deprivation of sense and motion, and sometimes with convulsions, a second type beginning with headache, and a third type with loss of power on one side of the body and of speech, often with recovery),48 and by Hope and Bennett (transient apoplexy, primary apoplexy with death or slow

recovery, ingravescent apoplexy with partial recovery and relapse, and paraplexic apoplexy with paralysis).49 There are several reasons why the brain lesion in what we now call cerebral infarction was not identified until the middle of the 19th century. First, it was impossible to recognize ischaemic softening in patients who had usually died not long after their stroke. Fixation methods were not available until the end of the 18th century; Vicq d’Azyr, Marie Antoinette’s physician, was the first to use alcohol as a tissue fixative,50 while formaldehyde fixation was not employed until a century later.51 Second, it is probable that many patients diagnosed as having died from apoplexy in fact had suffered from other conditions. If in our time the diagnosis is wrong in 20–30% of patients referred with a presumed stroke,52–54 the diagnostic accuracy was presumably no better in centuries past.

2.4 Cerebral infarction (ischaemic stroke)

After Morgagni’s seminal book the organ-based approach to medicine quickly spread from Italy to other countries. In France, the first proponents were surgeons. After the French revolution the strict distinction between medicine and surgery disappeared, driven by the reorganization of hospital care (no longer managed by the church but by the state) and by the need to train a large number of new doctors, for military as well as civilian duties (‘peu lire, beaucoup voir, beaucoup faire’).55,56 It was Leon Rostan (1790–1866; Fig. 2.5), a physician at the Salpêtrière in Paris, who clearly recognized softening of the brain as a separate lesion, distinct from haemorrhage, although the pathogenesis still escaped him. He published his findings in an unillustrated monograph, the first edition of which appeared in 1820.57 The lesions were most commonly found in the corpus striatum, thalamus or centrum semiovale, but they also occurred in the cerebral cortex, brainstem and cerebellum. Old cases showed a yellowish-green discoloration, whereas if the patients had died soon after the event the colour of the lesion was chestnut or reddish. The softening might be so extreme as to lead to the formation of a cyst. In other patients it was difficult to detect any change in firmness or in colour. Rostan distinguished softening of the brain from ‘apoplexy’, a term he no longer used for stroke in general, but which he regarded as being synonymous with haemorrhagic stroke. He supposed that softening of the brain was more common than brain haemorrhage, although some haemorrhages were secondary to softening. The clinical manifestations were thought

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 15

2.4 Cerebral infarction (ischaemic stroke)

seemed for a long time the most logical ‘paradigm’ to fall back on to explain liquefaction of brain tissue.59 The first inkling of a relationship between arterial disease and ‘ramollissem*nt’, as many English writers continued to call brain softening in deference to Rostan, was voiced by Abercrombie, in a later edition of his textbook.60 He drew an analogy with gangrene, caused by ‘failure of circulation’, this in turn being secondary to ‘ossification of arteries’. The role of arterial obstruction as a primary cause of softening of the brain was confirmed by others,61,62 but the theory of inflammation continued to be defended by a few adherents.63,64 Some were aware that apoplexy could be caused by ‘cerebral anaemia’ (as opposed to congestion), not only through loss of blood but also by a reduced vascular pressure, particularly in the case of heart disease.44 Other missing links in the understanding of cerebral infarction were clarified by Rokitansky (1804–1878) in Vienna and by Virchow (1821–1902) in Berlin. Rokitansky divided cerebral softening (which he termed encephalomalacia) into three varieties: red (haemorrhagic) softening, inflammatory in nature; white softening (synonymous with ‘serous apoplexy’) caused by congestion and oedema; and, the most common variety, yellow softening, of which the pathogenesis was unknown.219 Virchow (Fig. 2.6) revolutionized medical thinking about vascular disease by firmly putting the

Fig. 2.5 Léon Rostan (1790–1866).

to occur in two stages: first ‘fugitive’ disturbances in the use of a limb, in speech or in visual or auditory perception, sooner or later followed by hemiplegia and coma, in a slowly progressive fashion. Although Rostan recognized ‘ossification’ of the cerebral arteries, he did not associate these lesions with cerebral softening via obstruction of the arterial system. At any rate he doubted the prevailing opinion that the primary lesion was some kind of inflammatory response. After all, there was redness and swelling (rubor, tumor), if not warmth and pain (calor, dolor), to complete the cardinal signs of inflammation delineated by Celsus in the first century AD. Rostan’s contemporary Lallemand (1790– 1853) was much more outspoken and had little doubt that inflammation was at the root of cerebral softening.58 Readers trained in the 21st century may find this difficult to understand but they should be aware that ‘inflammation’ was a rather common explanation for disease from the middle of the 18th century until some hundred years later.46 Just as in our time some poorly understood medical conditions are often interpreted in terms of autoimmune disease, perhaps erroneously, inflammation

Fig. 2.6 Rudolph Virchow (1821–1902) teaching at a postmortem in the Charité Hospital in Berlin.

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 16

Chapter 2 Development of knowledge about cerebrovascular disease

emphasis on changes in the vessel wall rather than in blood; Schiller called it the victory of ‘solidism’ over ‘humoralism’.9 Virchow also firmly established that thrombosis of arteries was caused not by inflammation but by fatty metamorphosis of the vessel wall, even if he had to found his own journal before his papers could be published.65,66 To describe the changes in the arterial wall Virchow revived the term ‘arteriosclerosis’, first used by Lobstein.67 Virchow’s disciple Julius Cohnheim coined the culinary term ‘infarction’ (from the Latin verb infarcire, ‘to stuff into’), but strictly reserved it for haemorrhagic necrosis (‘stuffing’, by seeping of blood into ischaemic tissue, through damaged walls of capillaries) as opposed to ischaemic necrosis.68

2.5 Thrombosis and embolism

Virchow observed thrombosis as the result of atherosclerosis, and also embolism, in patients with gangrene of the lower limbs caused by clots from the heart. The term ‘embolism’ was newly coined by him, at least in medical parlance. He extrapolated these events to the cause of cerebral softening: Here there is either no essential change in the vessel wall and its surroundings, or this is ostensibly secondary. I feel perfectly justified in claiming that these clots never originated in the local circulation but that they are torn off at a distance and carried along in the blood stream as far as they can go.65 The relationship between vegetations on the heart valves and stroke had in fact been suggested a century earlier by Boerhaave’s pupil Gerard van Swieten, personal physician to the Austrian empress Maria Theresa and founder of the Viennese school of medicine: It has been established by many observations that these polyps occasionally attach themselves as excrescences to the columnae carneae of the heart, and perhaps then separate from it and are propelled, along with the blood, into the pulmonary artery or the aorta, and its branches . . . were they thrown into the carotid or vertebral arteries, could disturb – or if they completely blocked all approach of arterial blood to the brain – utterly abolish the functions of the brain.69 For more than a century after Virchow’s accurate pathological descriptions of arterial occlusions, the term ‘cerebral embolism’ was almost synonymous with embolism from the heart (parenthetically, it still is, in many contemporary textbooks and papers – another

illustration of how slowly ideas change). Sources of embolism in the extracranial arteries were hardly considered until the 1960s, at least in teaching. By the same token, the term ‘cerebral thrombosis’ remained firmly entrenched in clinical thinking as being more or less synonymous with cerebral infarction without associated heart disease, the implication being that in these cases the site of the atheromatous occlusion was in the intracranial vessels. For example, this is what the sixth edition of Brain’s Diseases of the Nervous System says on the subject in 1968: Progressive occlusion of cerebral blood vessels impairs the circulation in the regions they supply. The effects of this depend upon the size and situation of the vessel, and the rate of onset of the occlusion particularly in relation to the collateral circulation. Actual obstruction of an artery by atheroma, with or without subsequent thrombosis, causes softening of the region of the brain supplied by the vessel.2 That the notion of ‘local thrombosis’ persisted for such a long time must have been because of its appealing simplicity, not by lack of observations to the contrary. As long ago as 1905, Chiari drew attention to the frequency of atherosclerosis in the region of the carotid bifurcation and suggested that embolization of atheromatous material might be a cause of cerebral softening,70 and not much later Hunt described the relationship between carotid occlusion and stroke.71 The much later general acceptance of extracranial atherosclerosis as an important cause of cerebral ischaemia was prompted by two further developments. The first was the attention generated by Fisher’s studies, in which he re-emphasized the role of atherosclerosis at the carotid bifurcation, at least in white patients.72 He clinically correlated these lesions not only with contralateral hemiplegia but also with attacks of monocular blindness in the ipsilateral eye.73 The second development was imaging. Cerebral angiography by direct puncture of the carotid artery had been introduced by Moniz in 1927,74,75 but imaging of the carotid bifurcation in patients with stroke became common only after the advent of catheter angiography,76 and later of ultrasound techniques. Now it is modern CT or MR angiography that often shows abnormalities of the internal carotid artery near its origin, at least in patients with transient or permanent deficits in the territory of the main trunk of the middle cerebral artery or one of its branches. The earlier patients are investigated, the greater the chance of detecting the site where the embolus has become impacted in the cerebral arterial tree. The therapeutic implications of identifying lesions in the carotid artery in symptomatic patients became clear through the two large randomized controlled trials of carotid endarterectomy in the 1980s and 1990s,

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 17

2.6 Transient ischaemic attacks

which showed overall benefit from the operation for severe degrees of stenosis.77 In 25–30% of patients with temporary or permanent occlusion of large intracranial vessels no source of embolism can be found in the neck or in the heart.78,79 Pathological observations suggesting that the aorta may harbour atherosclerotic lesions80 have been confirmed in a large autopsy series and by transoesophageal echocardiography during life.81,82 Of course, there is more to ischaemic stroke than embolism from large vessels, but the history of small vessel disease and non-atheromatous causes of ischaemia is rather more recent. Before concluding the sections on cerebral infarction, thrombosis and embolism, we should like briefly to draw attention to the term ‘cerebrovascular accident’, which enjoyed some undeserved popularity in the middle half of the last century. The problem was that sometimes the term was used as a synonym for cerebral infarction, at other times to denote stroke in general. In this day and age the term is a highly specific sign of woolly thinking. We can do no better than quote Schiller:9 That rather blurry and pompous piece of nomenclature must have issued from the well-meant tendency to soften the blow to patients and their relatives, also from a desire to replace ‘stroke’, a pithy term that may sound unscientific and lacking gentility. ‘Cerebrovascular accident (CVA)’ can be traced to the early 1930s – between 1932, to be exact, when it was still absent from the 15th edition of Dorland’s Medical Dictionary, and the following edition of 1936 where it first appeared. The occasional medical student or junior doctor who still takes refuge in the term ‘CVA’ in an attempt to cover up ignorance about the precise type of cerebrovascular event in a given patient (while avoiding sharing the term ‘stroke’ with the laity) should either find out or come clean about not knowing.

2.6 Transient ischaemic attacks

It is difficult to trace the first descriptions of what we now call transient ischaemic attacks (TIAs) of the brain or eye, because symptoms representing focal deficits were not clearly distinguished from non-specific symptoms of a more global nature such as fainting or headache.83 Wepfer recorded that he had seen patients who recovered from hemiplegia in one day or less.31 An 18th-century account has been retrieved in the patient’s own words, not muddled by medical interpretation, which makes it as lucid as it would have been today.

The subject is Jean Paul Grandjean de Fouchy, writing in 1783, at the age of 76 years:84 Toward the end of dinner, I felt a little increase of pain above the left eye and in that very instant I became unable to pronounce the words that I wanted. I heard what was said, and I thought of what I ought to reply, but I spoke other words than those which would express my thoughts, or if I began them I did not complete them, and I substituted other words for them. I had nevertheless all movements as freely as usual . . . I saw all objects clearly, I heard distinctly what was being said; and the organs of thought were, it seemed to me, in a natural state. This sort of paroxysm lasted almost a minute. Once it had become established, in the middle of the 19th century, that cerebral softening was not caused by an inflammatory process but by occlusion of cerebral arteries, temporary episodes of ischaemia were recognized increasingly often in the next few decades.1,85–89 In the course of time, three main theories have been invoked to explain the pathophysiology of TIAs, at least in relation to atherosclerosis: the vasospasm theory, the haemodynamic theory and the thromboembolic theory.83

2.6.1 The vasospasm theory Arterial spasm as a cause of gangrene of the extremities was described by Raynaud (1834–1881) in his doctoral thesis of 1862.90 Others extrapolated his theory of vasospasm to the cerebral circulation.91,92 Russel, writing in 1909 about a 50-year-old farmer who had suffered three attacks of tingling and numbness in the right arm and the right side of the face, dismissed thrombosis (‘Thrombus, once formed, does not break up and disappear in some mysterious way’) and instead invoked a phenomenon of ‘local syncope’, analogous to Raynaud’s disease or some cases of migraine: ‘There must be some vessel constriction, local in site, varying in degree and in extent, coming and going, intermittent’.92 Even the great Osler mounted the bandwagon of the vasospastic theory to explain transient attacks of aphasia and paralysis: ‘We have plenty of evidence that arteries may pass into a state of spasm with obliteration of the lumen and loss of function in the parts supplied’.89 Vasospasm remained the most popular theory to explain TIAs in the first half of the 20th century and provided the rationale for so-called cerebral vasodilators. Up to the 1980s these useless drugs were still widely prescribed in some European countries, not only for TIAs but for ‘senility’ in general; in France they were the third most commonly prescribed medication in 1982.93 In the front line of medicine, however, the vasospastic theory went into decline soon after World War II, firstly

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 18

Chapter 2 Development of knowledge about cerebrovascular disease

because the cerebral arteries are among the least reactive in the body,94,95 and secondly because more plausible theories emerged (see below). Only under strictly defined conditions can vasospasm be a causal factor in the pathogenesis of cerebral ischaemia, that is, after subarachnoid haemorrhage or in association with migraine, and even in these conditions its role is contentious. Nevertheless, vasospasm has resurfaced as a possible cause of episodes of transient monocular blindness that are frequent and stereotyped and have no altitudinal distribution,96 or even of transient motor or sensory deficits not related to migraine.97 Such events must be extremely rare.

2.6.2 The haemodynamic theory The notion of ‘low flow’ without acute vessel obstruction as a cause of cerebral ischaemia should perhaps be attributed to Ramsay Hunt, who drew an analogy between the symptoms of carotid stenosis or occlusion and the symptoms of intermittent claudication in patients with severe peripheral arterial disease.71 But, it was especially after 1951, when Denny-Brown suggested that TIAs might be caused by ‘episodic insufficiency in the circle of Willis’,95 that interest in the haemodynamic aspects of TIAs was fully aroused. Indeed, it was mainly the surgical community for which the concept of ‘cerebral intermittent claudication’ continued to have great appeal, despite the incongruity of the relatively constant blood flow to the brain and the large fluctuations in flow that occur in the legs, depending on their level of activity. Clinical studies failed to support the notion of haemodynamic failure. After artificial lowering of the blood pressure by means of hexamethonium and postural tilting, in 35 patients who had either experienced TIAs or who had known carotid artery disease, only one of the patients developed symptoms of focal cerebral ischaemia before a syncopal attack which signified global rather than focal ischaemia of the brain.98 Similarly, cerebral ischaemia with naturally occurring attacks of hypotension, such as cardiac arrhythmias, is almost always syncopal and not focal in nature,99 and cardiac arrhythmias do not occur more often in patients with TIAs than in controls.100 Once the first successful carotid reconstruction had been reported,101 the intuitive belief in the haemodynamic theory led to an ever-increasing number of carotid endarterectomies being performed (indeed, often called ‘carotid disobstruction’) in patients with and even without TIAs, despite the absence of any formal proof of efficacy. These developments caused understandable concern in the neurological community.102,103 Fortunately the controversy prompted well-designed clinical trials, which have served to define to a large extent the role of this operation.77

That the haemodynamic theory does not apply to most patients with TIAs is not to say that the exceptional patient cannot suffer from ‘misery perfusion’.104 In the presence of multiple occlusions or stenoses of the extracranial arteries, the haemodynamic reserve may be so poor that minor changes in systolic blood pressure cannot be compensated for. Such triggering events include a change from a sitting to a standing position, turning the head, heating of the face or looking into bright light.105–107 Perhaps for this small group of patients extracranial-intracranial bypass surgery has something to offer after all, despite the negative results of the randomized trial in a large but relatively unselected group of patients with occlusion of the internal carotid or middle cerebral artery.108

2.6.3 The thromboembolic theory In the 1950s C. Miller Fisher (Fig. 2.7) not only gave new impetus to some older observations about the relationship between stroke and atheromatous lesions of

Fig. 2.7 C. Miller Fisher (1913–).

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 19

2.6 Transient ischaemic attacks

F C B A D

H G B A D

C B A D

(F) 9:37 E

•

• •

•

H G

C B A D

(E) 9:35

M F

J F

(G) 9:55 E

(D) 9:30

B A D

H G

(B) 9:19

B A D

(A) 9:15

H G

B A

the carotid bifurcation, but he also provided evidence that the pathogenesis was more complex than could be explained by fixed arterial narrowing. First, he saw a patient in whom hemiplegia had been preceded by attacks of transient monocular blindness in the contralateral eye, that is, ‘the wrong eye’.73 Second, through assiduous ophthalmoscopic observations, he saw white bodies passing slowly through the retinal arteries during an attack of transient monocular blindness (Fig. 2.8), the whitish appearance and friability of the moving material suggesting that these were emboli, largely made up of platelets.109 These findings were confirmed by Ross Russell,110 whilst others saw atheromatous emboli in the retinal vessels, which did not move but had become impacted.111,112 After these direct observations of the ocular fundus, additional – but more indirect – arguments corroborated the notion of artery-to-artery embolism as an important cause of TIAs. • In many patients with attacks involving the cortical territory of the middle cerebral artery there is an associated lesion of the internal carotid artery, but in only very few of them is the stenosis severe enough, with a residual lumen of 1–2 mm, for blood flow to be impaired below critical levels, even assuming there is no collateral circulation.113 In addition, the stenosis is constant but the episodes of ischaemia transient,

G Cattle trucking in veins

C Fig. 2.8 Diagrams of observations in a patient with an attack of transient monocular blindness in the left eye (except the upper temporal quadrant); the attack had started at 8.55 am, 20 minutes before the beginning of the observations. The column of blood in the retinal arteries was in some places interrupted by white segments, initially at the stems of the superior and inferior retinal arteries (A); also the column of blood in at least six venous branches of the superior half of the retina was broken into transverse bands (so-called cattle trucking). The white segments in the retinal arteries slowly passed through the superior temporal artery (B–H). At (C) the vision in the upper half of the visual field had returned. At (D) a fine trickle of erythrocytes moved slowly along one side of white segment AB to the superior nasal artery, and at (E) vision had also returned in the inferior temporal quadrant. After (H), when the column of blood had been completely restored, vision returned to normal. (From Fisher, 1959;109 by kind permission of the author and Neurology.)

(H) 10:00

B D

without evidence for cardiac arrhythmias as an additional factor. During carotid endarterectomy, fresh and friable thrombi are seen adherent to atheromatous plaques in the carotid bifurcation, especially in patients with recent attacks.114 In patients with ocular as well as cerebral attacks, the two kinds of attack almost never occur at the same time.114 Manual compression of the carotid artery may lead to dislodgement of atheromatous emboli to the cerebral circulation.115 If patients continue to have TIAs after occlusion of the ipsilateral internal carotid artery, there is often an additional atheromatous lesion in the common carotid or external carotid artery, these vessels being important collateral channels, supplying the hemisphere via retrograde flow through the ophthalmic artery.106 Asymptomatic emboli have been seen to flash up during angiography,116 while fibrin thrombi have been seen to pass through a cortical artery during craniotomy for a bypass procedure.117 Transcranial Doppler monitoring has uncovered an ongoing stream of highintensity transient signals (HITS), probably small emboli, in patients with symptomatic carotid lesions.118 The HITS disappear after carotid endarterectomy,119 the rate depending on the interval since operation.120

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 20

Chapter 2 Development of knowledge about cerebrovascular disease

Whilst artery-to-artery thromboembolism from atheromatous plaques may seem the most important factor in explaining TIAs and ischaemic strokes, it is not necessarily the only one, not even in a single individual patient. For example, it is probable that emboli have especially damaging effects in vascular beds that are chronically underperfused.

2.7 Intracerebral haemorrhage

Extravasation of blood into the brain parenchyma was recognized as early as 1658 by Wepfer,31 although we commented above that he saw the clot as an obstruction of ‘vital spirits’ rather than as the disease in itself, and subsequently by Morgagni.40 The cause remained obscure, and to a large extent it still is. In 1855, before blood pressure could be measured, Kirkes observed hypertrophy of the heart in 17 of 22 patients with fatal brain haemorrhage.121 Charcot and Bouchard in 1868 examined the brains of patients who had died from intracerebral haemorrhage and immersed these in running water; they found multiple, minute outpouchings of small blood vessels, so-called miliary aneurysms.122 The irony of these two names being joined is that Bouchard, once Charcot’s pupil, in later years generated much hostility between himself and his former chief, because he wanted to found a school of his own and to be considered the most influential man in the faculty of

medicine.123 It was in this adversarial atmosphere that in 1892 Bouchard, as president of the jury that had to judge the competition for the rank of professeur agrégé, did not admit Charcot’s pupil Babinski.124 Babinski subsequently left academic medicine by becoming chief of the Pitié hospital, where he devoted much time to the study of clinical signs, including the now famous ‘toe sign’.125 The aneurysms described by Charcot and Bouchard were white or brownish-coloured nodules about 0.5–2.0 mm in diameter, attached to a small arteriole, most often in the basal ganglia. At the beginning of the 20th century, Charcot and Bouchard’s theory came under attack and some proposed that the primary lesion in intracerebral vessels was atherosclerosis, that most of these dilatations were not true aneurysms at all but false aneurysms caused by intramural dissection, while rupture could also occur by weakening of the vessel wall without previous aneurysm formation;126 also some ‘miliary aneurysms’ may in fact have been clots in perivascular (Virchow-Robin) spaces. Alternative explanations for the pathogenesis of primary intracerebral haemorrhage included primary necrosis of brain tissue or its vessels. Some assumed that arteries dilate and rupture only when a previous infarct had occurred, thus depriving the feeding vessel of its normal support.127,128 The frequent coexistence of hypertension led to several theories other than plain rupture. Rosenblath postulated that a renal toxin caused necrosis of vessel walls,129 Westphal that arterial spasm was an intermediate factor130 and Schwartz that a multitude of terminal arterial branches became permeable.224 In the 1960s injection techniques revived the notion

Fig. 2.9 Godfrey N. Hounsfield (1919– 2004), the British engineer who received the Nobel prize in medicine in 1979 for the development of computed tomography (together with the American physicist A.M. Cormack).

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 21

2.8 Subarachnoid haemorrhage

2.8.1 Diagnosis The first unequivocal description of an aneurysm, though unruptured, was by Franciscus Biumi in 1765, who saw it not on the circle of Willis but in the cavernous sinus (at the time called Vieussens’ receptacle).8 Morgagni had also mentioned dilatations of arteries that may have been aneurysms.40 In 1812 John Cheyne provided the first illustration of lethal subarachnoid haemorrhage at the base of the brain as a result of ‘rupture of the anterior artery of the cerebrum’ (Fig. 2.11), but the aneurysm that must have been the source of the haemorrhage was not recognized at the time.45 One year later Blackall reported a postmortem observation in which the haemorrhage as well as the offending aneurysm (of the basilar artery) were identified in a 20-year-old woman.142 The observation was coincidental, because Blackall was primarily interested in her ‘anasarca’ (generalized oedema, or ‘dropsy’). The brain was also examined by Hodgson, who in his book on diseases of blood vessels Fig. 2.10 CT scan of an intracerebral haemorrhage, from the early 1970s.221

of microaneurysms,131,132 although some still suspect that the injection pressures can artifactually distend or rupture vessel walls.133 Amyloid angiopathy was first recognized as a cause of primary intracerebral haemorrhage in the first half of the 20th century.134–136 This type of haemorrhage occurs especially at the border of white and grey matter and not in the deep regions of the brain that are the most common sites of haemorrhages associated with microaneurysms. The first series of such patients appeared in the 1970s.137,138 The invention of computed tomography by Hounsfield (1919–2004; Fig. 2.9) in the 1970s made it possible to distinguish intracerebral haemorrhage quickly and reliably from cerebral infarction (Fig. 2.10).139,140

2.8 Subarachnoid haemorrhage

The history of ‘meningeal apoplexy’ is relatively short. The disorder was not recognized until three years before the battle of Waterloo; in the following 125 years numerous accounts appeared that combined a few personal cases with attempts to review the entire world literature up to that time, the last being a heroic overview of 1125 patients.141

Fig. 2.11 The first anatomical illustration of subarachnoid haemorrhage, from Cheyne (1812).45 A probe has been passed into the proximal end of the internal carotid artery and emerges at the presumed site of rupture; the offending aneurysm was not recognized at the time but presumably it was at the origin of the posterior communicating artery from the carotid artery, or at the anterior communicating artery complex.

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 22

Chapter 2 Development of knowledge about cerebrovascular disease

Fig. 2.12 Pea-sized aneurysmal dilatation of one of the branches of the middle cerebral artery (opened), containing a clot; the abnormality was surrounded by a large, fresh haemorrhage, which had caused the death of the 19-year-old patient, 8 days after a first episode with sudden headache (from Bright, 1831).61

made the point that the extravasated blood was contained under the arachnoid membrane.143 Serres, not aware of these books, published two similar observations in a French periodical.144 Parenthetically, it should be pointed out that medical journals, with articles about a variety of observations, did not emerge until the beginning of the 19th century, whereas the first scientific journals in general date from the middle of the 17th century.145 In England, Richard Bright, one of the champions of the movement of ‘organ-based medicine’ that had started in Italy and France,146 included an illustration of a pea-sized aneurysm on a branch of the middle cerebral artery in his richly illustrated book that appeared in 1831 (Fig. 2.12).61 Series of other fatal cases were reported in the next few decades.147–150 The erroneous notion that aneurysms are congenital malformations, caused by a defect in the muscular layer of the arterial wall, was first put forward in 1887,151 and subsequently adopted by other writers,152,153 to be perpetuated into contemporaneous textbooks and students’ minds. Turnbull also pointed out, correctly, that syphilis was an extremely rare cause of cerebral aneurysms.153 Series of aneurysms diagnosed post mortem were often biased towards those measuring several cm,154 or included septic aneurysms, associated with endocarditis.155 It took a long time before the clinical features were sorted out. In 1852 Brinton observed that fatal rupture was not the only possible presentation of aneurysms, and that other manifestations were local pressure,

convulsive attacks or ‘inflammation’ (a rather fuzzy notion at the time).147 The sudden onset of the headache led Lebert and Bartholow to suppose, in the 1860s and 70s, that the diagnosis might be made during life.149,150 Lebert also observed the characteristic paralysis of the oculomotor nerve in patients before they died from rupture of an aneurysm at the origin of the posterior communicating artery from the internal carotid artery.149 Indeed the diagnosis of ruptured aneurysm was made on two occasions in a patient with sudden headache and oculomotor palsy, by Hutchinson in England and by Bull in Norway,156,157 but apparently these two observations had little impact. The introduction of lumbar puncture, in 1891 by Quincke,158 initially only for therapeutic purposes in hydrocephalic patients, led to the diagnosis of ‘meningeal apoplexy’ in patients who survived a subarachnoid haemorrhage.159–161 It took another three decades before the connection with rupture of a cerebral aneurysm was made. After all, the condition was supposed to be invariably fatal since it had been recognized only after death. The ‘selection bias’ seems obvious in retrospect, but in our own time the same error was made with intraventricular haemorrhage – until CT scanning showed it in those who survived. That sudden headache and meningeal haemorrhage as diagnosed by lumbar puncture could be caused by a ruptured aneurysm without fatal outcome received widespread attention only after Charles Symonds (1890– 1978; Fig. 2.13) had published two landmark articles about the subject in 1923 and 1924.162,163 It had all started in 1920, when Symonds spent some time abroad as a temporary resident in the service of the neurosurgeon Harvey Cushing (1869–1939), who had moved not long before from Baltimore to Harvard University and the Peter Bent Brigham hospital in Boston. A 52-year-old woman had been admitted with repeated episodes of headache and unconsciousness; on examination she had a right oculomotor palsy and blurring of the optic discs. A right subtemporal decompression for a suspected tumour showed recently clotted blood extending over the entire hemisphere, apparently coming from the base of the skull.162 Apparently Symonds suggested a ruptured aneurysm as the cause.164 When the diagnosis was confirmed at autopsy (the patient had died the day after the operation) Cushing ordered Symonds to spend his remaining time in the library to review everything on the subject: ‘Either this was a fluke or there was reason in it’.165 The next advances were neuroradiological. The first angiographic visualization of a cerebral aneurysm during life was reported by Egas Moniz in 1933,166 six years after the technique had been first applied.74 In those days angiography was a hazardous procedure (involving

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 23

2.8 Subarachnoid haemorrhage

Fig. 2.14 Norman Dott (1897–1973).

Fig. 2.13 Sir Charles Symonds (1890–1978).

surgical dissection of the carotid artery), to such an extent that someone like Cushing only rarely had his patients undergo it before neurosurgical exploration. Even today, in the era of selective catheterization, the risks are far from negligible. Fortunately, minimally or non-invasive techniques of angiography by means of computed tomography or magnetic resonance have largely replaced catheter angiography, at least for diagnostic purposes. The greatest leap forward in our times was the advent of computed tomography;139 this technique made it possible to localize the extent of the haemorrhage in a precise fashion, to separate aneurysmal haemorrhage from nonaneurysmal haemorrhage and, by serial investigations, to detect and distinguish the most important complications: rebleeding, delayed ischaemia and hydrocephalus.

2.8.2 Surgical treatment Ligation of the carotid artery has been practised since the times of Ambroise Paré (1510–1590) as a method to stop arterial bleeding in patients with neck wounds. Once

aneurysms were recognized as the cause of subarachnoid haemorrhage it was a logical step to consider this procedure as a method to decrease the risk of rebleeding.157 Hutchinson would actually have carried out the operation in 1864 had the patient not declined at the last moment, going on to survive for another 11 years.156 Around 1886 Horsley was one of the first who actually ligated the (common) carotid artery in the neck, for a tumorous aneurysm.154 For decades carotid ligation remained the only surgical intervention possible, but most patients were managed conservatively because the complications of surgery were considerable.167 In 1931 the Edinburgh neurosurgeon Norman Dott (1897–1973; Fig. 2.14), at that time only 33 years old, carried out the first intracranial operation for a ruptured aneurysm.168 It was a more or less desperate attempt because the aneurysm had already rebled twice, leaving the patient comatose for some hours after the last episode, also with some degree of right-sided hemiparesis and aphasia. To complicate matters further, the patient was a well-known Edinburgh solicitor, 53 years old and chairman of the board of governors of the Royal Hospital for Sick Children. But, both the patient and the young neurosurgeon were prepared to take the risk.169 About the operation Dott wrote:170 A left frontal approach was employed and it was a difficult matter to elevate the tense and oedematous

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 24

Chapter 2 Development of knowledge about cerebrovascular disease

twelve minutes. As the retaining instrument was then cautiously withdrawn, no further bleeding occurred. The vessel was further cleared and thin strips of muscle were prepared and wound around it until a thick collar of muscle embedded the aneurysm and adjacent arterial trunks (Fig. 2.15). The patient recovered well and a few weeks later Dott wrote, his sense of triumph carefully hidden: ‘Mr Colin Black’s tibialis anticus seems to have stuck well to his internal carotid – he has gone for a holiday’.169 In later years Dott and his patient went fishing together on a number of occasions and Mr Black’s neurological condition remained good until he died from myocardial infarction 11 years after the momentous operation. Unfortunately, on later occasions the outcome with a direct approach to the aneurysm was often disappointing, if not fatal, and Dott reverted to ligating the internal carotid artery in the neck or the proximal anterior cerebral artery intracranially. In 1937 Dandy was the first to use a clip to occlude the neck of the aneurysm that had bled (Fig. 2.16).171 Yet in some patients a clip could not be secured, and in those

Fig. 2.15 Norman Dott’s drawing of the first intracranial operation for aneurysm. The proximal middle cerebral artery aneurysm was exposed and wrapped with muscle through a left frontal flap. (From Todd et al., 1990;168 by kind permission of the authors and the Journal of Neurology, Neurosurgery and Psychiatry.)

brain and identify the basal structures, which were bloodstained and largely embedded in clot. The left optic nerve was found and the internal carotid artery was defined at its outer side. This vessel was closely followed upwards, outwards and backwards to its bifurcation into the middle and anterior cerebral arteries. As this point was being cleared of tenacious clot a formidable arterial haemorrhage filled the wound. With the aid of suction apparatus, held closely to the bleeding point, we were able to see the aneurysm. It sprang from the upper aspect of the bifurcation junction; it was about 3 mm in diameter; blood spurted freely from its semidetached fundus. Meanwhile a colleague was obtaining fresh muscle from the patient’s leg. A small fragment of muscle was accurately applied to the bleeding point and held firmly in place so that it checked the bleeding and compressed the thin walled aneurysmal sac. Thus it was steadily maintained for

Fig. 2.16 Illustration from Dandy’s 1944 monograph on intracerebral arterial aneurysms.222 The legend is: ‘Typical aneurysm of the intracranial internal carotid artery, showing the narrow neck of the sac and the bulging aneurysm; also the point of rupture. The inset shows the clip placed on the neck of the aneurysm, and the aneurysm itself shrivelled with the electric cautery.’

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 25

2.9 Treatment and its pitfalls

cases he often had to have recourse to so-called trapping, by clipping the parent vessel on either side of the aneurysm. Decades later, Drake devised a technique for approaching basilar artery aneurysms, notoriously difficult until then, and managed to apply clips to them.172 In the 1960s, spring clips, which could be removed when placement was less than optimal, came into use and replaced the silver clips used by Dandy. Nevertheless, the direct operation of aneurysms remained dangerous and controlled trials of the efficacy of aneurysm operations were equivocal. Attempts to increase the safety of the operation included temporary cardiac arrest, hypotension and deep hypothermia, all without much success, although no formal trials were done. In the 1980s, a consensus developed amongst neurosurgeons that direct operation of the aneurysm should best be delayed until 12–14 days after the initial haemorrhage. This regimen meant, of course, that a proportion of patients rebled or suffered other complications in the meantime. The gradual introduction of the operating microscope for aneurysm surgery in the 1970s made early operation (within 3 days) not only feasible but also fashionable, despite the dearth of evidence from controlled clinical trials. The medical management of patients with ruptured aneurysms has also improved in recent years, especially the prevention of delayed ischaemia. In the 1980s the Italian neuroradiologist Guglielmi developed an endovascular method for occluding aneurysms by means of detachable platinum coils, initially only for aneurysms for which a surgical approach was hazardous or impossible.173,174 In the last few years ‘coiling’ has largely replaced the surgical approach, provided the method is feasible for a particular aneurysm.

2.9 Treatment and its pitfalls

Doctoring has always implied treatment. In the past, medical management was almost invariably based on what later turned out to be erroneous pathophysiological concepts, and the treatments were almost invariably ineffective, if not actually harmful. Such pitiful situations are often repeated in present times, much more often than physicians and surgeons care to realize. Anyone who finds it amusing to read about 19th-century regimens, including measures such as bleeding, mustard poultices, castor oil and turpentine enemas as treatments for apoplexy, should read post-1950 treatises about the efficacy of vasodilator drugs or about transplantation

of omentum to the intracranial cavity, as a chastening experience.

2.9.1 The numerical method Before different treatments could ever be compared, it was necessary to find methods for grouping patients together and also somehow to convert disease outcomes into numbers. The Paris physician Pierre Charles Alexandre Louis (1787–1872; he survives eponymously in the angulus Ludovici of the sternum) is generally credited with the introduction of the numerical method in medicine. In fact his contribution was more a credo than a practical method.175 True enough, there is the famous example of his empirical criticism of bloodletting: of 47 patients with pneumonia treated with bloodletting 18 died, against only nine of the 36 patients in the untreated group.176 But Louis did not have the mathematical training to estimate the likelihood that a difference of this magnitude might arise by chance. It was a mathematician, Jules Gavarret (1809–1890), who criticized the analysis and conclusions of Louis’s studies, although he agreed with the design.177 Even more purely mathematical was the notion of the ‘average human’, an approach proposed by Adolphe Quetelet (1796–1874). Groups and averages, these were notions that evoked not merely resistance but outright revulsion in the ranks of the established medical professionals. How on earth could one ever ignore the unique characteristics of each single individual by forcing these together into an artificial ‘mean’? And how could one ever believe in a standard treatment, any more than in a standard shoe? The advent of experimental physiology intensified the opposition. The famous Claude Bernard (1813–1878) warned that one will never encounter an ‘average’ in nature, and that grouping of observations will obscure the true relationships between natural phenomena.178 And the equally legendary Lord Lister (1827–1912) relied more on the theoretical basis of his antiseptic method than on the actual death rates.179 Until the 20th century, counting disease events was limited to population studies.180 The beginnings of epidemiology can be traced to Sir William Petty (1623– 1687), one of the founders of the Royal Society, and John Graunt (1620–1674). They worked together in collecting numerical data to describe patterns of mortality. A century and a half later, E. Blackmore reported not only on deaths but also on incident cases of disease in Plymouth.181,182 Victorian counterparts took this further. William Farr (1807–1883), who had trained under Louis in Paris, linked self-devised classifications of diseases and occupations to population statistics at the General Registry Office. John Snow (1813–1858) mapped the

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 26

Chapter 2 Development of knowledge about cerebrovascular disease

occurrence of cholera cases in the streets of London and related these to the positions of the local water pumps; these studies culminated in the famous act of Snow removing the handle from the pump in Broad Street, during the 1854 cholera epidemic. Incidentally, he later specialized in chloroform anaesthesia.183 The first epidemiological studies of stroke were not performed until after World War II. An early study of stroke incidence in the community was done in the UK.184 Population-based studies addressing risk factors specifically for stroke were subsequently reported from the US (the Framingham cohort), Japan and Finland.185–187

2.9.2 Clinical trials The introduction of the randomized controlled clinical trial heralded the era of ‘evidence-based medicine’ or rather ‘organized empiricism’, since medicine is not and probably will never be a positivist science like physics or chemistry.188 Randomization in a therapeutic experiment slowly gained acceptance after the landmark UK Medical Research Council (MRC) trial of streptomycin in pulmonary tuberculosis with random assignment to treatment groups.189 Some forerunners had already used parallel control groups. Louis (1787–1872) had been preceded by James Lind (1716–1794) in 1753 (lemons and oranges to prevent scurvy in sailors). A further step was the introduction of chance to obtain an equal balance between the experimental group and the control group. In 1898 Fibiger (1867–1928) used assignment on alternate days (injection of serum for diphtheria),190,191 and in 1931 Amberson et al. flipped a coin to divide patients with pulmonary tuberculosis into those who received gold treatment and controls.192 Blinding (or masking, as ophthalmologists prefer to say) of patients was also practised by Amberson’s group, as had been done four years earlier by Ferguson et al. in a test of vaccines for the common cold.193 Masking of those who were to assess outcome was advocated in 1944 by the pulmonary physicians Hinshaw and Feldman,194 and eventually carried out in the MRC streptomycin trial of 1948. Allocation in that historic trial took place by means of randomization. An important advantage of random allocation, applied by R.A. Fisher in agriculture in the 1920s, is that it ensures equal and unbiased balancing between the two groups.195 But the main reason why Sir Austin Bradford Hill (1897–1991), the trial’s principal investigator, chose randomization is that at the same time it ensured concealment of the allocation schedule from those involved in entering patients in the trial.196,197 Clinical trials in cerebrovascular disease were no exception to the rule that most methodological errors have to

be committed before they are recognized, as the correct solutions are often counter-intuitive. In the 1950s anticoagulant drugs seemed a rational form of treatment to prevent further strokes in survivors of (presumed) brain infarction. The same Bradford Hill who had pioneered the tuberculosis trial took the initiative for two such trials, the first in 142 and the second, with exclusion of hypertensives, in 131 patients.198,199 There was no significant difference in the rate of non-fatal stroke between the treatment groups and controls, while there was some excess of fatal strokes, possibly haemorrhages, in patients on anticoagulants. From that time onwards anticoagulants were largely abandoned for the prevention of stroke, unless for specific indications such as a source of embolism in the heart. However, it took at least two decades before it dawned on the neurological community that the trials of anticoagulants in brain ischaemia had been too small, separately as well as collectively, to detect even large protective effects – apart from other shortcomings.200 The same applied to an early secondary prevention trial with dipyridamole.201 The first intervention trial in acute stroke was with corticosteroids, by Dyken and White in 1956. They did not use randomization but stratified patients according to their clinical characteristics, and found a trend towards a higher death rate in the treated group (13/17 against 10/19 in controls), and ended up by identifying many of the methodological problems in this type of trial.202 The first trial of carotid endarterectomy excluded surgical mishaps from the analysis;203 subsequently the operation boomed to worrying levels, until checked by methodologically sound trials. The first large trial of aspirin in stroke prevention evoked much controversy,204 for one thing because its initiators had chosen ‘stroke or death’ as the outcome event instead of stroke alone;205 it took time for neurologists to realize that they treat whole patients rather than only their brains! Also, the initial conclusion that aspirin was ineffective for women is now a classical example of the dangers of subgroup analysis.

2.9.3 Measuring outcome: the ghost of Gall One of the stumbling blocks in trials of acute stroke used to be the babel of tongues with regard to the measurement of outcome. Initially, so-called ‘stroke scales’ were applied for this aim, analogous to scales for other specific neurological conditions, such as Parkinson’s disease or multiple sclerosis. Although the stated purpose of ‘stroke scales’ is to measure outcome, these scales are nothing but codifications of the neurological examination, while of course that examination has no other purpose than localizing lesions within the nervous system. With such a diagnostic approach, different functions of the nervous

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 27

2.9 Treatment and its pitfalls

system are separately assessed: power of limbs, speech, visual fields, etc. This reductionist, mechanistic notion of brain function reflects the localizationists’ position in the scientific battle that raged in the second half of the 19th century, the opposing party believing in so-called equipotentiality. The ‘equipotentialists’ believed that the brain worked as a unitary system, brain tissue being omnipotent and flexible in its function. Consequently, brain damage would result in a decrease in the overall level of performance, but not in loss of specific functions. The champion of this camp was the French physiologist Flourens, who supported his views with experiments on dogs and pigeons.206 The alternative notion, that of localization of specific functions, was propounded in a somewhat bizarre fashion by the anatomists Gall and Spurzheim.207 They believed that every intellectual and moral property had its own position on the surface of the brain (Fig. 2.17) and that the degree of development of these dispositions could be identified by locating overlying protuberances on the skull (Fig. 2.18). However, the theory of localization gained respectability after the stimulation experiments of Fritsch and Hitzig on anaesthetized dogs, in which they found that weak electrical currents applied through platinum electrodes to the anterior regions of

Fig. 2.18 The pseudo-science of phrenology lived on well beyond the 19th century. The Lavery Electric Phrenometer of 1907 was intended to lend modern accuracy to the measurements of bumps on the skull.223 (Reproduced by kind permission of Cambridge University Press.)

Fig. 2.17 Phrenology head (Fowler); each region of the skull is supposed to represent a mental faculty, such as ‘mirthfulness’, ‘perception of form’ or ‘ideality’.

the brain surface produced muscle contractions in the opposite half of the body.208,209 The clash between the two opposing factions culminated in 1881 at the Third International Congress of Medicine, held in London.210 The equipotentialists were represented by the German physiologist Goltz, who showed the audience a dog in which a substantial portion of the brain had been removed by means of a hose, but who could still move all four limbs, trunk and tail, and who had retained all his senses. Later, it would turn out that the lesions were less extensive than had been claimed. On the same afternoon Ferrier showed two chimpanzees, one deaf after removal of the auditory cortex, the other limping with a hemiplegic gait after extirpation of the contralateral motor area (the sight of which led Charcot to jump up and exclaim: ‘Mais c’est un patient!’). The localizationists had won the day, but they won too completely. The greater part of the brain has no ‘primary’ motor, sensory or cognitive tasks, and serves to

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 28

Chapter 2 Development of knowledge about cerebrovascular disease

Table 2.2 The state of an individual cannot be constructed from separate components. Imagine that you met ‘the boy next door’ from your childhood days after an interval of 30 years, and that your question ‘How are you?’ was answered with a list of details, instead of by a general statement (‘fine’ for example). Profession Civil state Bank account Car Holidays Sport

dentist married positive Volvo Tuscany golf

for 15 years wife 1.68 m, 59 kg a9634.92 240S 3 weeks handicap 8

Total ?

connect and integrate the separate ‘functions’. Similarly, everyday life consists of a multitude of tasks that are integrated and difficult to separate. Mood, initiative and speed of thinking are some of the essential features of human life that can be severely affected by stroke but are sadly ignored in ‘stroke scales’. It is therefore naive to try and rebuild an entire human being from separate ‘building blocks’ (Table 2.2; Fig. 2.19). Patients are more than the sum of their signs. A higher, more integrated level of measurement is needed; that is, scales should measure function not at the level of the organ but at the level of the person (disability scales), or even at the level of social interaction (handicap scales). What really counts for patients is what they can do in life, compared with what they want to do or were once able to do.

2.9.4 Meta-analysis and systematic reviews In the last quarter of the 20th century, Richard Peto and his colleagues Tom Chalmers and Iain Chalmers developed a method to overcome the problem that single studies may or may not show a significant difference in treated patients compared with controls, but that the magnitude of the difference can only be expressed as a confidence interval, which is usually wide. They collated all related trials in a given field by which the differences between the treatment group and the control group in each trial could be combined.211 The key assumption is that, if a given treatment has any material effect on the incidence or outcome of disease, then the direction, although not necessarily the size, of this effect tends to be similar in different circ*mstances. If all available studies are combined, the confidence interval can be narrowed considerably and reviewer bias is avoided. There clearly was a pressing need for up-to-date systematic reviews of all the available evidence regarding the various aspects of care of stroke patients – indeed, of all medical interventions. This need led to the Cochrane Collaboration, which includes a stroke review group.212

Fig. 2.19 The librarian (1566), by Giuseppe Arcimboldo (1530–1593). Oil on canvas, 97 × 71 cm. (By kind permission of Skokloster Castle, Sweden.)

The graphic representation of systematic reviews started in 1978, with simple lines to depict 95% confidence intervals.213 In 1982 Lewis and Clarke had the idea to combine the separate estimates into an overall estimate, at the bottom of the figure.214,215 Subsequently Richard Peto’s group solved the paradox that small trials were most conspicuous because of their large confidence intervals, by putting a square at the site of the point estimate, the size of the square being proportional to the power of the trial (Fig. 2.20).216 These graphs have since become known as ‘forest plots’, probably because the many lines might be seen as trees.215

2.10 Epilogue

Despite the many advances in the knowledge about stroke that we have highlighted, our story could not but

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 29

References

Ratio of crude death rates (99% Cl) Beta-blocker: control

remain anachronistic and fragmented. It is extremely difficult to try and stand in the shoes of one’s forebears, because to achieve this the mind should be cleared from all knowledge obtained since their time.217 For those of us who can think back as small a time span as three decades, what diagnosis did we make, in those times, in patients we now know to have survived carotid dissection or intracranial venous thrombosis, to name but two examples? Heaven only knows. In the same way, not so much longer ago, it was impossible to distinguish haemorrhage from infarction; or haemorrhage from some mysterious other condition that mimicked haemorrhage but in which the brain looked practically normal; or stroke from other brain diseases; or even stroke from heart disease. Necessarily our account has been anecdotal. In reality the progress of science is slow and continuous, not a succession of breakthroughs. This also applies to the few decades we have witnessed during our own careers. We do not expect a sensational novelty when we walk into hospital tomorrow, but a lot has changed since we were medical students. This refers not only to the body of medical knowledge but also to the methods of medical research. Empirical testing has gained ascendancy over pathophysiological theory, for the treatment as well as the prevention of disease. The rate of change is a bit like the shifting position of the sun across the sky: one cannot see it move, but there is a dramatic sweep between dawn and sunset. We expect to see many more dawns in stroke research.

References 0

0.5

1.0

1.5

2.0

Beta-blocker better Beta-blocker worse Treatment effect P < 0.0001 Fig. 2.20 ‘Forest plot’. Figure redrawn after Lewis and Ellis’s original plot from 1982, which for the first time combined 99% confidence intervals of different placebo-controlled clinical trials of beta blockers after myocardial infarction.214 This modern variant shows the results of each component study as a square centred on the point estimate of the result of each study; the size of each square is proportional to the amount of information provided in that trial. A horizontal line runs through the square to show its confidence interval (CI). The overall estimate from the meta-analysis and its confidence interval are put at the bottom, represented as a diamond.215 (Reproduced by kind permission of the BMJ Publishing Group.)

1 Oppenheim H. Lehrbuch der Nervenkrankheiten für Ärtzte und Studierende. 6th ed. Berlin: S. Karger, 1913. 2 Brain WR. Diseases of the Nervous System. 6th ed. Oxford: Oxford University Press, 1968. 3 Dechambre A. Mémoire sur la curabilité du ramollissem*nt cérébral. Gaz Med Paris 1838; 6:305–14. 4 Durand-Fardel CLM. Mémoire sur une altération particulière de la substance cérébrale. Gaz Med Paris 1842; 10:23–38. 5 Fisher CM. Lacunes: small, deep cerebral infarcts. Neurology 1965; 15:774–84. 6 Fisher CM, Curry HB. Pure motor hemiplegia of vascular origin. Arch Neurol 1965; 13:30–44. 7 Fisher CM. The arterial lesions underlying lacunes. Acta Neuropathol (Berl) 1969; 12:1–15. 8 Biumi F. Observatio V: Carotis ad receptaculum Vieusenii aneurysmatica etc. In: Sandifort E, editor. Observationes anatomicae, scholiis illustratae (thesaurus dissertationum). Leiden: S. & J. Lichtmans, 1765, pp. 373–9.

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 30

Chapter 2 Development of knowledge about cerebrovascular disease

9 Schiller F. Concepts of stroke before and after Virchow. Medical History 1970; 14:115–31. 10 McHenry LC. A history of stroke. Int J Neurol 1981; 15:314–26. 11 Karenberg A. Johann Jakob Wepfers Buch uber die Apoplexie (1658). Kritische Anmerkungen zu einem Klassiker der Neurologie. Nervenarzt 1998; 69:93–8. 12 McHenry LC. Garrison’s History of Neurology. Springfield: Charles C. Thomas, 1969. 13 Clarke E. Apoplexy in the Hippocratic writings. Bull Hist Med 1963; 37:301–14. 14 Karenberg A, Hort I. Medieval descriptions and doctrines of stroke: preliminary analysis of select sources. Part I: The struggle for terms and theories – late antiquity and early middle ages (300–800). J Hist Neurosc 1998; 7:162–73. 15 Galenus. Opera Omnia. Leipzig: Cnobloch, 1824. 16 Karenberg A, Hort I. Medieval descriptions and doctrines of stroke: preliminary analysis of select sources. Part II: Between Galenism and Aristotelism – Islamic theories of apoplexy (800–1200). J Hist Neurosc 1998; 7:174–85. 17 Jardine L. Worldly Goods: a new history of the Renaissance. London: Macmillan, 1996. 18 Karenberg A, Hort I. Medieval descriptions and doctrines of stroke: preliminary analysis of select sources. Part III: Multiplying speculations – the high and late middle ages (1000–1450). J Hist Neurosc 1998; 7:186–200. 19 Vesalius A. De humani corporis fabrica. Basle: J. Oporini, 1543. 20 Copernicus N. De revolutionibus orbium coelestium. Nuremberg: J. Petreius, 1543. 21 Vesalius A. Tabulae anatomicae. Venice: D. Bernardini, 1538. 22 Clarke E, Dewhurst K. An Illustrated History of Brain Function: imaging the brain from antiquity to the present. 2nd ed. San Francisco: Norman, 1996. 23 Berengario da Carpi J. Isagogae breves, perlucide ac uberrime, in anatomiam humani corporis etc. Venice: Benedictum Hectoris, 1535. 24 Harvey W. Exercitatio anatomica de motu cordis et sanguinis in animalibus. Frankfurt: G. Fitzer, 1628. 25 Dewhurst K. Thomas Willis’s Oxford Lectures. Oxford: Sandford Publications, 1980. 26 Willis T. Cerebri Anatome. London: Martyn & Allestry, 1664. 27 Meyer A, Hierons R. Observations on the history of the ‘Circle of Willis’. Medical History 1962; 6:119–30. 28 Fallopius G. Observationes anatomicae. Venice: Marcus Antonius Ulmus, 1561. 29 Casserio G. Tabulae Anatomicae (edited by D. Bucretius). Venice: E. Deuchinum, 1627. 30 Vesling J. Syntagma anatomicum, locis pluribus actum, emendatum, novisque iconibus diligenter exornatum. Patavii: Pauli Frombotti Bibliopolae, 1647. 31 Wepfer JJ. Observationes anatomicae, ex cadaveribus eorum, quos sustulit apoplexia, cum exercitatione de ejus loco affecto. Schaffhausen: J.C. Suteri, 1658. 32 Tatu L, Moulin T, Monnier G. The discovery of encephalic arteries. From Johann Jacob Wepfer to Charles Foix. Cerebrovasc Dis 2005; 20:427–32. 33 Willis T. Dr. Willis’s Practice of Physick. London: Dring, Harper & Leigh, 1684.

34 Kidd M, Modlin IM. Frederik Ruysch: master anatomist and depictor of the surreality of death. J Med Biogr 1999; 7:69–77. 35 Moore W. The Knife Man: the extraordinary life and times of John Hunter, father of modern surgery. London: Bantam Press, 2005. 36 van Eems J. Hermanni Boerhaave Praelectiones Academicae de Morbis Nervorum. Leiden: Petrus van der Eijk and Cornelius de Pecker, 1761. 37 Bayle F. Tractatus de apoplexia. Toulouse: B. Guillemette, 1677. 38 Aegineta P. The Seven Books (translated by Francis Adams). London: The Sydenham Society, 1844. 39 Mistichelli D. Trattato dell’apoplessia. Roma: A. de Rossi, 1709. 40 Morgagni GB. De sedibus et causis morborum per anatomen indigatis libri quinque. Venice: ex typographica Remondiana, 1761. 41 Portal A. Observations sur l’apoplexie. Histoire de l’Académie des Sciences 1781; 83:623–30. 42 Baillie M. The Morbid Anatomy of Some of the Most Important Parts of the Human Body. London: J. Johnson & G. Nicol, 1793. 43 Hall M. Lectures on the Nervous System and its Diseases. London: Sherwood, Gilbert & Piper, 1836. 44 Burrows G. On Disorders of Cerebral Circulation and on the Connection between Affections of the Brain and Diseases of the Heart. London: Longman, Brown, Green & Longmans, 1846. 45 Cheyne J. Cases of Apoplexy and Lethargy with Observations on Comatose Patients. London: Underwood, 1812. 46 King LS. Transformations in American Medicine: from Benjamin Rush to William Osler. Baltimore: Johns Hopkins University Press, 1991. 47 Serres ERA. Nouvelle division des apoplexies. Ann Med Chir 1819; 1:246–363. 48 Abercrombie J. Pathological and Clinical Researches on Diseases of the Brain and Spinal Cord. Edinburgh: Waugh & Innes, 1828. 49 Hope J, Bennett JH, Pritchard JC, Taylor RH, Thomson T. Dissertations on nervous diseases. In: Tweedie A, editor. Library of Practical Medicine. Philadelphia: Lea & Blanchard, 1840. 50 Vicq d’Azyr F. Traité d’anatomie et de physiologie. Paris: F.A. Didot, 1786. 51 Blum F. Der Formaldehyd als Härtungsmittel: vorläufige Mitteilung. Z wiss Mikr mikr Technik 1893; 10:314–5. 52 Harbison J, Hossain O, Jenkinson D, Davis J, Louw SJ, Ford GA. Diagnostic accuracy of stroke referrals from primary care, emergency room physicians, and ambulance staff using the face arm speech test. Stroke 2003; 34:71–6. 53 Heckmann JG, Stadter M, Dutsch M, Handschu R, Rauch C, Neundorfer B. Einweisung von NichtSchlaganfallpatienten auf eine Stroke Unit [Hospitalization of non-stroke patients in a Stroke Unit]. Dtsch Med Wochenschr 2004; 129:731–5. 54 Ronning OM, Thommessen B. Nar hjerneslagdiagnosen er feil [Stroke: when the diagnosis is wrong]. Tidsskr Nor Laegeforen 2005; 125:1655–7.

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 31

References

55 Foucault M. Naissance de la clinique. Paris: Presses Universitaires de France, 1963. 56 Bynum WF. Science and the Practice of Medicine in the Nineteenth Century. Cambridge: Cambridge University Press, 1994. 57 Rostan L. Recherches sur le ramollissem*nt du cerveau. Ouvrage dans lequel on s’efforce de distinguer les diverses affections de ce viscère par des signes caractéristiques. 1st ed. Paris: Béchet, 1820. 58 Lallemand F. Recherches anatomo-pathologiques sur l’encéphale et ses dépendances. Paris: Béchet, 1824. 59 Kuhn TS. The Structure of Scientific Revolutions. Chicago: Chicago University Press, 1962. 60 Abercrombie J. Pathological and Practical Researches on Diseases of the Brain and Spinal Cord. 2nd (from 3rd Brit.) ed. Philadelphia: Carey, Lea & Blanchard, 1836. 61 Bright R. Reports of Medical Cases, selected with a view of illustrating the symptoms and cure of diseases by a reference to morbid anatomy. London: Longman, Rees, Orme, Brown & Green, 1831. 62 Carswell R. Pathological Anatomy: illustrations of the elementary forms of disease. London: Longman & Co., 1838. 63 Cruveilhier J. Anatomie pathologique du corps humain; descriptions avec figures lithographiés et coloriés; des diverses altérations morbides dont le corps humain est susceptible. Paris: J.B. Baillière, 1842. 64 Durand-Fardel CLM. Traité du ramollissem*nt du cerveau. Paris: J.-B. Baillière, 1843. 65 Virchow RLK. Ueber die akute Entzündung der Arterien. Archiv Pathol Anat 1847; 1:272–378. 66 Virchow R. Thrombose und Embolie: Gefässentzündung und septische Infection. In: Virchow R, editor. Gesammelte Abhandlungen zur wissenschaftlichen Medizin. Frankfurt: Meidinger, 1856, pp. 219–732. 67 Lobstein JFM. Traité d’anatomie pathologique. Paris: Levrault, 1829. 68 Cohnheim J. Untersuchungen ueber die embolischen Processe. Berlin: Hirschwald, 1872. 69 van Swieten GLB. Commentaria in Hermanni Boerhaave Aphorismos De Cognoscendis et Curandis Morbis. Leiden: J. & H. Verbeek, 1755. 70 Chiari H. Über das Verhalten des Teilungswinkels des Carotis Communis bei der Endarteritis chronica deformans. Verh Ddtsch path Ges 1905; 9:326–30. 71 Hunt JR. The role of the carotid arteries, in the causation of vascular lesions of the brain, with remarks on special features of the symptomatology. Am J Med Sci 1914; 147:704–13. 72 Fisher CM. Occlusion of the internal carotid artery. Arch Neurol Psych 1951; 65:346–77. 73 Fisher CM. Transient monocular blindness associated with hemiplegia. Arch Ophthalmol 1952; 47:167–203. 74 Moniz E. L’encéphalographie artérielle, son importance dans la localisation des tumeurs cérébrales. Rev Neurol (Paris) 1927; 48:72–90. 75 Moniz E. Die cerebrale Arteriographie und Phlebographie. Berlin: Julius Springer, 1940.

76 Seldinger SI. Catheter replacement of the needle in percutaneous arteriography. Acta Radiol 1953; 39:368–78. 77 Rothwell PM, Eliasziw M, Gutnikov SA, Fox AJ, Taylor DW, Mayberg MR et al. Analysis of pooled data from the randomised controlled trials of endarterectomy for symptomatic carotid stenosis. Lancet 2003; 361:107–16. 78 Sacco RL, Prabhakaran S, Thompson JL, Murphy A, Sciacca RR, Levin B et al. Comparison of warfarin versus aspirin for the prevention of recurrent stroke or death: subgroup analyses from the Warfarin-Aspirin Recurrent Stroke Study. Cerebrovasc Dis 2006; 22:4–12. 79 Jood K, Ladenvall C, Rosengren A, Blomstrand C, Jern C. Family history in ischemic stroke before 70 years of age: the Sahlgrenska Academy Study on Ischemic Stroke. Stroke 2005; 36:1383–7. 80 Soloway HB, Aronson SM. Atheromatous emboli to central nervous system. Arch Neurol 1964; 11:657–67. 81 Amarenco P, Duyckaerts C, Tzourio C, Henin D, Bousser MG, Hauw JJ. The prevalence of ulcerated plaques in the aortic arch in patients with stroke. N Engl J Med 1992; 326:221–5. 82 Amarenco P, Cohen A, Tzourio C, Bertrand B, Hommel M, Besson G et al. Atherosclerotic disease of the aortic arch and the risk of ischemic stroke. N Engl J Med 1994; 331:1474–9. 83 Hachinski VM. Transient cerebral ischemia: a historical sketch. In: Clifford Rose F, Bynum WF, editors. Historical Aspects of the Neurosciences (Festschrift for M. Critchley). New York: Raven Press, 1982, pp. 185–93. 84 Benton AL, Joynt RJ. Early descriptions of aphasia. Arch Neurol 1960; 3:205–22. 85 Wood GB. Treatise on the Practice of Medicine. Philadelphia: Lippincott, 1852. 86 Jackson JH. A lecture on softening of the brain. Lancet 1875; ii:335–8. 87 Hammond WA. Diseases of the Nervous System. New York: D. Appleton, 1881. 88 Gowers WR. A Manual of Diseases of the Nervous System. 2 ed. London: J&A Churchill, 1893. 89 Osler W. Transient attacks of aphasia and paralysis in states of high blood pressure and arteriosclerosis. Can Med Assoc J 1911; 1:919–26. 90 Raynaud M. De l’asphyxie locale et de la gangrène symmétrique des extrémités. Paris: L. Leclerc, 1862. 91 Peabody GL. Relation between arterial disease and visceral changes. Trans Assoc Am Physicians 1891; 6:154–78. 92 Russel W. A post-graduate lecture on intermittent closing of the cerebral arteries: its relation to temporary and permanent paralysis. Br Med J 1909; 2:1109–10. 93 Payer L. Medicine and Culture: notions of health and sickness in Britain, the US, France and West Germany. London: V. Gollancz, 1989. 94 Pickering GW. Transient cerebral paralysis in hypertension and in cerebral embolism with special reference to the pathogenesis of chronic hypertensive encephalopathy. J Am Med Assoc 1948; 137:423–30. 95 Denny-Brown D. The treatment of recurrent cerebrovascular symptoms and the question of ‘vasospasm’. Med Clin North Am 1951; 35:1457–74.

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 32

Chapter 2 Development of knowledge about cerebrovascular disease

96 Burger SK, Saul RF, Selhorst JB, Thurston SE. Transient monocular blindness caused by vasospasm. N Engl J Med 1991; 325:870–3. 97 Call GK, Fleming MC, Sealfon S, Levine H, Kistler JP, Fisher CM. Reversible cerebral segmental vasoconstriction. Stroke 1988; 19:1159–70. 98 Kendell RE, Marshall J. Role of hypotension in the genesis of transient focal cerebral ischaemic attacks. Br Med J 1963; 2:344–8. 99 Reed RL, Siekert RG, Merideth J. Rarity of transient focal cerebral ischemia in cardiac dysrhythmia. J Am Med Assoc 1973; 223:893–5. 100 De Bono DP, Warlow CP. Potential sources of emboli in patients with presumed transient cerebral or retinal ischaemia. Lancet 1981; i:343–6. 101 Eastcott HHG, Pickering GW, Robb CG. Reconstruction of internal carotid artery in a patient with intermittent attacks of hemiplegia. Lancet 1954; ii:994–6. 102 Warlow C. Carotid endarterectomy: does it work? Stroke 1984; 15:1068–76. 103 Barnett HJM, Plum F, Walton JN. Carotid endarterectomy: an expression of concern. Stroke 1984; 15:941–3. 104 Klijn CJM, Kappelle LJ, Tulleken CAF, van Gijn J. Symptomatic carotid artery occlusion: A reappraisal of hemodynamic factors. Stroke 1997; 28:2084–93. 105 Caplan LR, Sergay S. Positional cerebral ischaemia. J Neurol Neurosurg Psychiatry 1976; 39:385–91. 106 Bogousslavsky J, Regli F. Delayed TIAs distal to bilateral occlusion of carotid arteries: evidence for embolic and hemodynamic mechanisms. Stroke 1983; 14:58–61. 107 Ross Russell RW, Page NGR. Critical perfusion of brain and retina. Brain 1983; 106:419–34. 108 The EC/IC Bypass Study Group. Failure of extracranialintracranial arterial bypass to reduce the risk of ischemic stroke. Results of an international randomized trial. N Engl J Med 1985; 313:1191–200. 109 Fisher CM. Observations on the fundus oculi in transient monocular blindness. Neurology 1959; 9:333–47. 110 Ross Russell RW. Observations on the retinal blood-vessels in monocular blindness. Lancet 1961; 11:1422–8. 111 Witmer R, Schmid A. Cholesterinkristall als retinaler arterieller Embolus. Ophthalmologica 1958; 135:432–3. 112 Hollenhorst RW. Significance of bright plaques in the retinal arterioles. J Am Med Assoc 1961; 178:23–9. 113 Archie JP, Feldtman JP. Critical stenosis of the internal carotid artery. Surgery 1981; 89:67–70. 114 Gunning AJ, Pickering GW, Robb-Smith AHT, Ross Russell RW. Mural thrombosis of the internal carotid artery and subsequent embolism. Q J Med 1964; 33:155–95. 115 Beal MF, Park TS, Fisher CM. Cerebral atheromatous embolism following carotid sinus pressure. Arch Neurol 1981; 38:310–12. 116 Watts C. External carotid artery embolus from the internal carotid artery ‘stump’ during angiography: case report. Stroke 1982; 13:515–17. 117 Barnett HJM. The pathophysiology of transient cerebral ischemic attacks: therapy with antiplatelet antiaggregants. Med Clin North Am 1979; 63:649–80.

118 Markus H. Transcranial Doppler detection of circulating cerebral emboli. A review. Stroke 1993; 24:1246–50. 119 Siebler M, Sitzer M, Rose G, Bendfeldt D, Steinmetz H. Silent cerebral embolism caused by neurologically symptomatic high-grade carotid stenosis. Event rates before and after carotid endarterectomy. Brain 1993; 116:1005–15. 120 van Zuilen EV, Moll FL, Vermeulen FE, Mauser HW, van Gijn J, Ackerstaff RG. Detection of cerebral microemboli by means of transcranial Doppler monitoring before and after carotid endarterectomy. Stroke 1995; 26:210–13. 121 Kirkes WS. On apoplexy in relation to chronic renal disease. Med Times Gaz 1855; 11:515–16. 122 Charcot JM, Bouchard C. Nouvelles recherches sur la pathogénie de l’hémorrhagie cérébrale. Arch Physiol norm pathol 1868; 1:110–27, 643–65, 725–34. 123 Iragui VJ. The Charcot-Bouchard controversy. Arch Neurol 1986; 43:290–5. 124 Satran R. Joseph Babinski in the competitive examination (agrégation) of 1892. Bull N Y Acad Med 1974; 50:626–35. 125 van Gijn J. The Babinski sign: the first hundred years. J Neurol 1996; 243:675–83. 126 Ellis AG. The pathogenesis of spontaneous intracerebral hemorrhage. Proc Pathol Soc Philadelphia 1909; 12:197–235. 127 Hiller F. Zirkulationsstörungen im Gehirn, eine klinische und pathologisch-anatomische Studie. Arch Psychiat Nervenkr 1935; 103:1–53. 128 Globus JH, Epstein JA, Green MA, Marks M. Focal cerebral hemorrhage experimentally induced. J Neuropathol Exp Neurol 1949; 8:113–16. 129 Rosenblath L. Über die Entstehung der Hirnblutung bei dem Schlaganfall. Dtsch Z Nervenkr 1918; 61:10–143. 130 Westphal K. Über die Entstehung und Behandlung der Apoplexia sanguinea. Dtsch med Wschr 1932; 58:685–90. 131 Ross Russell RW. Observations on intracerebral aneurysms. Brain 1963; 86:425–42. 132 Cole FM, Yates PO. The occurrence and significance of intracerebral micro-aneurysms. J Pathol Bacteriol 1967; 93:393–411. 133 Challa VL, Moody DM, Bell MA. The Charcot-Bouchard aneurysm controversy: impact of a new histologic technique. J Neuropathol Exp Neurol 1992; 51:264–71. 134 Fischer O. Die presbyophrene Demenz, deren anatomische Grundlage und klinische Abgrenzung. Z gesamte Neurol Psychiatr 1910; 3:371–471. 135 Scholz W. Studien zur Pathologie der Hirngefässe. II. Die drusige Entartung der Hirnarterien und -capillaren. Z Gesamte Neurol Psychiatr 1938; 162:694–715. 136 Pantelakis S. Un type particulier d’angiopathie sénile du système nerveux central: l’angiopathie congophile: topographie et fréquence. Monatsschr Psychiatr Neurol 1954; 128:219–56. 137 Torack RM. Congophilic angiopathy complicated by surgery and massive hemorrhage: a light and electron microscopic study. Am J Pathol 1975; 81:349–65. 138 Jellinger K. Cerebrovascular amyloidosis with cerebral hemorrhage. J Neurol 1977; 214:195–206.

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 33

References

139 Hounsfield GN. Computerised transverse axial scanning (tomography): I. Description of system. Br J Radiol 1973; 46:1016–22. 140 Hayward RD, O’Reilly GV. Intracerebral haemorrhage. Accuracy of computerised transverse axial scanning in predicting the underlying aetiology. Lancet 1976; 1:1–4. 141 McDonald CA, Korb M. Intracranial aneurysms. Arch Neurol Psych 1939; 42:298–328. 142 Blackall J. Observations on the Nature and Cure of Dropsies. 5th ed. London: Longman & Co., 1813. 143 Hodgson J. A Treatise on the Diseases of Arteries and Veins, containing the pathology and treatment of aneurisms and wounded arteries. London: T. Underwood, 1815. 144 Serres ERA. Observations sur la rupture des anévrysmes des artères du cerveau. Arch gén Méd 1826; 10:419–31. 145 Pyenson L, Sheets-Pyenson S. Reading: Books and the Spread of Ideas. Servants of nature: a history of scientific institutions, enterprises, and sensibilities. New York: W.W. Norton & Company, 1999: 211–35. 146 Berry D, Mackenzie C. Richard Bright (1789–1858): physician in an age of revolution and reform. London: Royal Society of Medicine Services Ltd., 1992. 147 Brinton W. Report on cases of cerebral aneurism. Trans Pathol Soc London 1852; 3:47–9. 148 Gull W. Cases of aneurism of the cerebral vessels. Guy’s Hosp Rep 1859; 5:281–304. 149 Lebert H. Über die Aneurysmen der Hirnarterien. Eine Abhandlung in Briefen an Herrn Geheimrat Professor Dr. Frerichs. Berl klin Wochenschr 1866; 3:209–405 (8 instalments). 150 Bartholow R. Aneurisms of the arteries at the base of the brain: their symptomatology, diagnosis and treatment. Am J Med Sci 1872; 44:373–86. 151 Eppinger H. Pathogenesis (Histogenesis und Aetiologie) der Aneurysmen einschliesslich des Aneurysma equi verminosum. Arch Klin Chir 1887; 35 (suppl. 1):1–563. 152 Wichern H. Klinische Beiträge zur Kenntnis der Hirnaneurysmen. Dtsch Zschr Nervenheilk 1912; 44:220–63. 153 Turnbull HM. Alterations in arterial structure, and their relation to syphilis. Q J Med 1914; 8:201–54. 154 Beadles CF. Aneurisms of the larger cerebral arteries. Brain 1907; 30:285–336. 155 Fearnsides EG. Intracranial aneurysms. Brain 1916; 39:224–96. 156 Hutchinson J. Aneurism of the internal carotid artery within the skull diagnosed eleven years before the patient’s death: spontaneous cure. Trans Clin Soc London 1875; 8:127–31. 157 Bull E. Akut Hjerneaneurisma-Okulomotoriusparalysem*ningealapoplexi. Norsk Magasin for Laegevidenskapen 1877; 7:890–5. 158 Quincke H. Die Lumbalpunktion des Hydrocephalus. Berl klin Wochenschr 1891; 28:929–33 and 965–8. 159 Froin G. Les hémorrhagies sous-arachnoidiennes et le méchanisme de l’hématolyse en général. Paris: G. Steinheil, 1904. 160 Guillain G. L’albuminurie massive dans le diagnostic des hémorrhagies méningées. Presse Méd 1915; 54:441–2.

161 Goldflam S. Beiträge zur Aetiologie und Symptomatologie der spontanen subarachnoidalen Blutungen. Dtsch Zschr Nervenheilk 1923; 76:158–82. 162 Symonds CP. Contributions to the clinical study of intracranial aneurysms. Guy’s Hosp Rep 1923; 73:139–58. 163 Symonds CP. Spontaneous subarachnoid haemorrhage. Quart J Med 1924; 18:93–122. 164 Cushing H. Contributions to the clinical study of cerebral aneurysms. Guy’s Hosp Rep 1923; 73:159–63. 165 Symonds CP. Autobiographical introduction. In: Symonds CP, editor. Studies in Neurology. London: Oxford University Press, 1970, pp. 1–23. 166 Moniz E. Anévrysme intra-cranien de la carotide interne droite rendu visible par l’artériographie cérébrale. Rev Oto-Neuro-Ophthal 1933; 11:198–203. 167 Schorstein J. Carotid ligation in saccular intracranial aneurysms. Br J Surg 1940; 28:50–70. 168 Todd NV, Howie JE, Miller JD. Norman Dott’s contribution to aneurysm surgery. J Neurol Neurosurg Psychiatry 1990; 53:455–8. 169 Rush C, Shaw JF. With Sharp Compassion: Norman Dott – freeman surgeon of Edinburgh. Aberdeen: Aberdeen University Press, 1990. 170 Dott N. Intracranial aneurysms: cerebral arterioradiography: surgical treatment. Trans Med Chir Soc Edinb 1932; 47:219–40. 171 Dandy WE. Intracranial aneurysm of internal carotid artery, cured by operation. Ann Surg 1938; 107:654–7. 172 Drake CG. Bleeding aneurysms of the basilar artery; direct surgical management in four cases. J Neurosurg 1961; 18:230–8. 173 Guglielmi G, Vinuela F, Sepetka I, Macellari V. Electrothrombosis of saccular aneurysms via endovascular approach. Part 1: Electrochemical basis, technique, and experimental results. J Neurosurg 1991; 75:1–7. 174 Guglielmi G, Vinuela F, Dion J, Duckwiler G. Electrothrombosis of saccular aneurysms via endovascular approach. Part 2: Preliminary clinical experience. J Neurosurg 1991; 75:8–14. 175 Matthews JR. Quantification and the Quest for Medical Certainty. Princeton: Princeton University Press, 1995. 176 Louis PCA. Recherches sur les effets de la saignée. Paris: de Mignaret, 1835. 177 Gavarret J. Principes généraux de statistique médicale. Paris: Librairies de la Faculté de Médecine de Paris, 1840. 178 Bernard C. Introduction à l’étude de la médicine expérimentale. Paris: J.-B. Baillière, 1865. 179 Lister J. Effect of the antiseptic system of treatment on the salubrity of a surgical hospital. Lancet 1870; i:4–6; 40–2. 180 Stolley PD, Lasky T. Investigating Disease Patterns: the science of epidemiology. New York: W.H. Freeman & Company, 1995. 181 Blackmore E. Reports on the diseases of Plymouth I. Edinburgh Medical and Surgical Journal 1829; 31:266–87. 182 Blackmore E. Reports on the diseases of Plymouth II. Edinburgh Medical and Surgical Journal 1829; 32:1–20. 183 Snow SJ. Operations Without Pain: the practice and science of anaesthesia in Victorian Britain. Houndmills, Basingstoke: Palgrave Macmillan, 2006.

9781405127660_4_002.qxd 10/13/07 11:14 AM Page 34

Chapter 2 Development of knowledge about cerebrovascular disease

184 Acheson J, Acheson HW, Tellwright JM. The incidence and pattern of cerebrovascular disease in general practice. J R Coll Gen Pract 1968; 16:428–36. 185 Kannel WB, Dawber TR, Cohen ME, McNamara PM. Vascular disease of the brain – epidemiological aspects: the Framingham study. Am J Public Health 1965; 55:1355–66. 186 Hirota Y, Katsuki S, Asano C. A multivariate analysis of risk factors for cerebrovascular disease in Hisayama, Kyushu Island, Japan. Behaviormetrika 1975; 2:1–11. 187 Salonen JT, Puska P, Mustaniemi H. Changes in morbidity and mortality during comprehensive community programme to control cardiovascular diseases during 1972–7 in North Karelia. Br Med J 1979; 2:1178–83. 188 Montgomery M. How Doctors Think: clinical judgment and the practice of medicine. New York: Oxford University Press, 2005. 189 Medical Research Council. Streptomycin treatment of pulmonary tuberculosis. Br Med J 1948; ii:769–82. 190 Fibiger J. Om Serumbehandling af Difteri. Hospitalstidende 1898; 6:309–25, 337–50. 191 Hróbjartsson A, Gøtzsche PC, Gluud C. The controlled clinical trial turns 100 years: Fibiger’s trial of serum treatment of diphtheria. Br Med J 1998; 317:1243–5. 192 Amberson JB, McMahon BT, Pinner M. A clinical trial of sanocrysin in pulmonary tuberculosis. Am Rev Tuberc 1931; 24:401–35. 193 Ferguson FR, Davey AFC, Topley WWC. The value of mixed vaccines in the prevention of the common cold. J Hyg 1927; 26:98–109. 194 Hinshaw HC, Feldman WH. Evaluation of chemotherapeutic agents in clinical tuberculosis. Am Rev Tuberc 1944; 50:202–13. 195 Fisher RA. The arrangement of field experiments. Journal of the Ministry of Agriculture 1926; 33:503–13. 196 Doll R. Controlled trials: the 1948 watershed. Br Med J 1998; 317:1217–20. 197 Chalmers I. Why transition from alternation to randomisation in clinical trials was made. Br Med J 1999; 319:1372. 198 Hill AB, Marshall J, Shaw DA. A controlled clinical trial of long-term anticoagulant therapy in cerebrovascular disease. Quart J Med 1960; 29:597–609. 199 Hill AB, Marshall J, Shaw DA. Cerebrovascular disease: a trial of long-term anticoagulant tharapy. Br Med J 1962; ii:1003–6. 200 Jonas S. Anticoagulant therapy in cerebrovascular disease: a review and meta-analysis. Stroke 1988; 19:1043–8. 201 Acheson J, Danta G, Hutchinson EC. Controlled trial of dipyridamole in cerebral vascular disease. Br Med J 1969; 1:614–15. 202 Dyken ML, White PT. Evaluation of cortisone in treatment of cerebral infarction. J Am Med Assoc 1956; 162:1531–4. 203 Fields WS, Maslenikov V, Meyer JS, Hass WK, Remington RD, Macdonald M. Joint study of extracranial arterial occlusion. V. Progress report of prognosis following surgery or nonsurgical treatment for transient ischemic attacks and cervical carotid artery lesions. J Am Med Assoc 1970; 211:1993–2003.

204 Canadian Cooperative Study Group. A randomized trial of aspirin and sulfinpyrazone in threatened stroke. N Engl J Med 1978; 299:53–9. 205 Kurtzke JF. Controversy in neurology: the Canadian study on TIA and aspirin – a critique of the Canadian TIA study. Ann Neurol 1979; 5:597–9. 206 Flourens MJP. Recherches expérimentales sur les propriétés et les fonctions du système nerveux, dans les animaux vertébrés. Paris: Crevot, 1824. 207 Gall FJ, Spurzheim JC. Anatomie et physiologie du système nerveux en général, et du cerveau en particulier, avec des observations sur la possibilité de reconnaître plusieurs dispositions intellectuelles et morales de l’homme et des animaux, par la configurations de leurs têtes. Paris: Schoell, 1819. 208 Fritsch GT, Hitzig E. Ueber die elektrische Erregbarkeit des Grosshirns. Arch Anat Physiol wiss Med 1870; 37:300–32. 209 Hitzig E. Untersuchungen über das Gehirn. Berlin: A. Hirschwald, 1874. 210 Thorwald J. Das Weltreich der Chirurgen. Stuttgart: Steingrüben, 1957. 211 Yusuf S, Peto R, Lewis J, Collins R, Sleight P. Beta blockade during and after myocardial infarction: an overview of the randomized trials. Prog Cardiovasc Dis 1985; 27:335–71. 212 Counsell C, Warlow C, Sanderco*ck P, Fraser H, van Gijn J. The Cochrane Collaboration Stroke Review Group. Meeting the need for systematic reviews in stroke care. Stroke 1995; 26:498–502. 213 Freiman JA, Chalmers TC, Smith H, Jr., Kuebler RR. The importance of beta, the type II error and sample size in the design and interpretation of the randomized control trial. Survey of 71 ‘negative’ trials. N Engl J Med 1978; 299:690–4. 214 Lewis JA, Ellis SH. A statistical appraisal of postinfarction betablocker trials. Prim Cardiol 1982; suppl. 1:317. 215 Lewis S, Clarke M. Forest plots: trying to see the wood and the trees. Br Med J 2001; 322:1479–80. 216 Antiplatelet Trialists’ Collaboration. Secondary prevention of vascular disease by prolonged antiplatelet treatment. Br Med J 1988; 296:320–31. 217 Temkin O. The historiography of ideas in medicine. In: Clarke E, editor. Modern methods in the history of medicine. London: The Athlone Press, 1971, pp. 1–21. 218 Hippocrates. The Genuine Works of Hippocrates. Baltimore: Williams & Wilkins, 1939. 219 Rokitansky C. Handbuch der pathologischen Anatomie. Wien: Braumüller und Seidel, 1842. 220 Todd EM. The Neuroanatomy of Leonardo da Vinci. Park Ridge: American Association of Neurological Surgeons, 1991. 221 New PJF, Scott WR. Computed Tomography of the Brain and Orbit (EMI scanning). Baltimore: Williams & Wilkins, 1975. 222 Dandy WE. Intracranial Arterial Aneurysms. Ithaca, New York: Comstock Publishing Company, 1944. 223 Blakemore C. Mechanics of the Mind. Cambridge: Cambridge University Press, 1977. 224 Schwartz P. Arten der Schlaganfälle des Gehirns. Berlin: Julius Springer, 1930.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 35

Is it a vascular event and where is the lesion? Identifying and interpreting the symptoms and signs of cerebrovascular disease

3.1 Introduction

3.2 Definitions of transient ischaemic attack, stroke and acute stroke syndrome (‘brain attack’ or ‘unstable brain ischaemia’) 35 3.3 The diagnosis of a cerebrovascular event

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset 3.5 Differential diagnosis of transient monocular blindness 3.6 Improving the reliability of the clinical diagnosis 3.7 Is it a subarachnoid haemorrhage?

3.1 Introduction

This chapter deals with the first two of several questions that need to be answered in the assessment of patients

103

108

111

presenting with a suspected transient ischaemic attack of the brain or eye (TIA) or stroke – is it a vascular event, and where is the lesion? (Table 3.1). We will begin by defining what is meant by a TIA and stroke, and follow by describing our approach to distinguishing TIA and stroke from other differential diagnoses.

Table 3.1 The assessment process in the diagnosis and subsequent management of a vascular event of the brain or eye. Is it a stroke, a transient ischaemic attack, or a brain attack? Which part of the brain has been affected? Which arterial territory has been affected? Is there a recognizable clinical syndrome? What pathological type of cerebrovascular event is it? What disease process caused the cerebrovascular event? What, if any, are the functional consequences? What treatment will improve survival free of handicap?

Chapter 3 Chapter 3 Chapter 4 Chapter 4 Chapter 5 Chapters 6–9 Chapters 10,11 Chapters 12–16

3.2 Definitions of transient ischaemic attack, stroke and acute stroke syndrome (‘brain attack’ or ‘unstable brain ischaemia’)

3.2.1 The definition of transient ischaemic attack A standard definition of a TIA is ‘a clinical syndrome characterized by an acute loss of focal cerebral or monocular function with symptoms lasting less than 24 h and which is thought to be due to inadequate cerebral or ocular blood supply as a result of low blood flow, thrombosis or embolism associated with disease of the arteries, heart or blood’.1,2 By definition therefore, TIAs are caused by transient reduction in blood flow to a part of the brain as a result of thromboembolic disease of the arteries, heart and blood. However, it is conceivable that TIAs may also arise from disease of the veins, causing transient reduction 35

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 36

Chapter 3 Is it a vascular event and where is the lesion?

in venous return from the brain or eye, but we are not aware of any patients with cerebral venous thrombosis who have presented with symptoms of TIA.3,4 The main features of TIAs are: • sudden onset; • symptoms of loss of focal neurological function (Table 3.2); • symptoms maximal at onset (more or less); • symptoms resolve within 24 h; • brain imaging (CT, MRI and DWI) may or may not show a relevant focal ischaemic lesion in the brain. Focal neurological symptoms Focal neurological symptoms are those that arise from a disturbance in an identifiable and localized area of the brain – for example, unilateral weakness from a lesion of the corticospinal tract (Table 3.2). However, there are some focal neurological symptoms which, when they occur in isolation, should probably not be considered as TIAs, because they occur more commonly in nonvascular conditions. Moreover, when present in isolation, Table 3.2 Focal neurological and ocular symptoms. Motor symptoms Weakness or clumsiness of one side of the body, in whole or in part (hemiparesis, monoparesis and sometimes only just the hand) Simultaneous bilateral weakness* Difficulty in swallowing* Imbalance* Speech/language disturbances Difficulty in understanding or expressing spoken language Difficulty in reading (dyslexia) or writing Difficulty in calculating Slurred speech* Sensory symptoms Altered feeling on one side of the body, in whole or in part Visual symptoms Loss of vision in one eye, in whole or in part Loss of vision in half or quarter of the visual field Bilateral blindness Double vision* Vestibular symptoms A sensation of movement* Behavioural/cognitive symptoms Difficulty in dressing, combing hair, cleaning teeth, geographical disorientation (visuospatial–perceptual dysfunction) Forgetfulness* *As an isolated symptom, this does not necessarily indicate focal cerebral ischaemia unless there is an appropriately sited acute infarct or haemorrhage, or there are additional definite focal symptoms.

they have not been shown to be associated with the high risk of future serious vascular events that characterizes TIAs. Such symptoms include rotational vertigo (caused by benign paroxysmal positional vertigo, vestibular neuritis, etc.), transient amnesia (transient global amnesia, psychogenic amnesia, etc.), and diplopia (myasthenia gravis, superior oblique myokymia, ocular myotonia, near-reflex accommodation spasm, etc.). Further research, based on imaging and long-term prognostic studies, of patients presenting with isolated focal neurological symptoms such as vertigo and dysarthria should help ascertain if, and how often, these symptoms are vascular in origin, and if so, what may help distinguish them from non-vascular causes (possibly older age and presence of other vascular risk factors). Some focal neurological symptoms should probably not be considered as transient ischaemic attacks if they occur in isolation (e.g. rotational vertigo, transient amnesia, and diplopia). Transient ischaemic attacks (and ischaemic stroke) are diagnosed with confidence only if the symptoms can be attributed to ischaemia of a focal brain area, and if there is no better explanation. Non-focal neurological symptoms Non-focal neurological symptoms (Table 3.3) are not usually caused by focal cerebral ischaemia; there are many other more common non-vascular neurological and non-neurological causes: • Generalized weakness caused by depression, acute inflammatory demyelinating polyradiculoneuropathy, myasthenia gravis, botulism, periodic paralysis, tick paralysis, acute toxic motor neuropathy, porphyric polyneuropathy, Miller Fisher syndrome and acute myelopathy; • Light-headedness caused by cardiac arrhythmias, aortic stenosis, cardiac tamponade, myocardial infarction, vasovagal syncope, reflex syncope and carotid sinus syncope; Table 3.3 Non-focal neurological symptoms. Generalized weakness and/or sensory disturbance Light-headedness Faintness ‘Blackouts’ with altered or loss of consciousness or fainting, with or without impaired vision in both eyes Incontinence of urine or faeces Confusion Ringing in the ears (tinnitus)

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 37

3.2 Definitions of transient ischaemic attack, stroke and acute stroke syndrome

(b) 60

Percentage of transient ischaemic attacks

(a)

50 40 30 20 10 0

720

45 40 35 30 25 20 15 10 5 0

720

Amaurosis fugax (n = 158) Transient ischaemic attacks of the brain (n = 311) Project;6 (b) 469 TIA patients in a hospital-referred series (from Hankey & Warlow, 19942 by kind permission of the authors and W.B. Saunders Co Ltd).

• Postural hypotension secondary to dehydration, blood pressure-lowering drugs (e.g. antihypertensives, dopamine agonists) and autonomic neuropathy; • Drop attacks (see below); • Confusion/delirium caused by metabolic/toxic encephalopathy. Therefore, non-focal symptoms should not be interpreted as being caused by a TIA or stroke unless accompanied by focal neurological symptoms. This is because – in isolation – they have other causes and do not seem to predict an increased risk of future serious vascular events.

patients with focal ischaemic neurological symptoms lasting longer than 24 h (ischaemic stroke) have similar causes and prognosis for future vascular events as patients with focal ischaemic neurological symptoms lasting less than 24 h (TIA).5 The duration of symptoms of TIA and ischaemic stroke are a continuum (Fig. 3.1);2,6 the only ‘relevance’ of longer duration TIAs and minor ischaemic stroke is that a relevant ischaemic lesion is more likely to be demonstrated by brain imaging (Fig. 3.2)7 and the risk of (recurrent) stroke increases.8–11 Indeed, increasing duration of focal neurological symptoms is one of the five main predictors of early recurrent stroke, along with Age, Blood pressure, Clinical features and Diabetes (and Duration of symptoms) in the ABCD2 prognostic model of early stroke risk after TIA10 (section 16.2.1). It has been suggested recently that the lower limit for the duration of symptoms should be 10 min.8,12 We would agree that TIAs lasting less than 10 min are associated with a lower risk of stroke than longer duration TIAs,10 and that focal neurological symptoms of the brain lasting seconds, and perhaps even a few minutes, are seldom likely to be TIAs of the brain (but TIAs of the eye – amaurosis fugax – may last only seconds). However, we remain uncertain just how short focal episodes can be and still be a TIA.

Neurological signs The standard definition of TIA allows abnormal but functionally unimportant focal neurological signs such as reflex asymmetry or an extensor plantar response to persist for longer than 24 h, provided the symptoms have resolved within 24 hours; this occurs in about 5% of patients.1,2

6–30

Fig. 3.1 Histogram showing the duration of the longest transient ischaemic attack (TIA) before presentation amongst (a) 184 TIA patients in the Oxfordshire Community Stroke

Non-focal symptoms such as faintness, dizziness or generalized weakness are seldom, if ever, likely to be due to focal cerebral ischaemia (i.e. seldom a transient ischaemic attack or stroke), but may be due to generalized brain ischaemia (e.g. syncope) as well as non-vascular neurological and non-neurological causes (e.g. hyperventilation or other manifestations of anxiety).

2–5

Duration of longest transient ischaemic attack (minutes)

It is not known how short a TIA can be (and still be a TIA).

Duration of symptoms

Brain imaging

The upper limit for the duration of symptoms (24 h) has more to do with the earth’s rotation than biology;

A substantial proportion of patients with cerebral TIA have evidence on brain imaging by CT or MRI of focal

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 38

Chapter 3 Is it a vascular event and where is the lesion?

Percentage of patients with infarct on CT

Transient ischaemic attack

49%

50 39%

40 30% 30

23%

20 11% 10 0

1–30 minutes

13% 8%

31–60 minutes

1–4 hours

5–24 hours

1–7 days

Duration of attack

areas of altered signal intensity, suggestive of ischaemia, in an area of the brain relevant to the transient symptoms.13–15 Diffusion-weighted MR imaging (DWI) is more sensitive than CT or MRI T2-weighted images, and identifies a relevant abnormality in up to two-thirds of patients with TIAs (Fig. 3.3).14,16–20 As a result some have proposed a revised definition of TIA which is restricted to patients who experience a brief episode of focal neurological dysfunction, presumptively caused by focal brain ischaemia, but without neuroimaging evidence of acute ischaemia or infarction. It has been further proposed that all other episodes with transient focal neurological symptoms with relevant lesion(s) (presumed infarction) on brain imaging are called ischaemic stroke.12,21–24 However, as stated in previous editions of this book, we do not accept such a definition, based on brain imaging, for the following reasons: • A CT or MR (including DWI) scan then becomes essential for the diagnosis of TIA, precluding up to one-fifth of patients who cannot undergo brain imaging (e.g. because they have an intraocular or intracerebral metallic foreign body or pacemaker, or are claustrophobic). Further, as the time to scanning decreases, and technology advances, the definition of a TIA would be constantly changing. • A positive CT, MRI or MRI-DWI may be a false positive (i.e. the ‘relevant’ abnormality on brain imaging may not represent recent infarction or ischaemia occurring at the time of the TIA (Table 3.4). Further, the limitations of DWI in diagnosing TIA cannot all be overcome (at present) by performing a complete stroke MRI examination, including T2-weighted imaging, angiography and perfusion imaging.25 • There is a gradual increase in the proportion of patients with a ‘relevant’ abnormality on brain imaging as the duration of symptoms increases, with no distinct change at 24 h7,12,18,23,24 (Fig. 3.2). This argues against there being any significantly different underlying pathophysiology between TIA and minor ischaemic stroke. There are also no significant

1–6 weeks

Persisting

Fig. 3.2 Histogram showing the relationship between the duration of focal neurological symptoms due to TIA and ischaemic stroke and the percentage of patients with an appropriately sited abnormality on brain imaging with CT (from Koudstaal, van Gijn, Frenken et al., 19927 by kind permission of the authors and Journal of Neurology, Neurosurgery and Psychiatry).

Table 3.4 Causes of an inaccurate diffusion-weighted MRI (DWI) diagnosis of TIA. False positive DWI diagnosis of TIA (at least 10%) Non-ischaemic cause of DWI hyperintensity Neurologically asymptomatic cause of DWI hyperintensity Chronic, persistent cause of DWI hyperintensity (i.e. old lesion) False negative DWI diagnosis of TIA (at least 10%) Very early imaging after onset of ischaemia (i.e. too early) Short-duration TIA (insufficient time to cause early ischaemic changes on DWI) Penumbral ischaemia (ischaemia is sufficient to cause neurological symptoms but insufficient to cause failure of the sodium-potassium ATPase pump in the neuronal cell membranes) Ischaemic region too small to image Ischaemic region too difficult to image (e.g. in brainstem) Confounding background lesion(s) Late imaging after onset of ischaemia (e.g. > 2 weeks after resolution of symptoms)

differences in the clinical features and prevalence of vascular risk factors among TIA patients with and without a relevant infarct on brain imaging.5 However, there may be a difference in prognosis if the results of initial studies showing the presence of such a ‘TIA scar’ on brain imaging is associated with an increased risk of future stroke are confirmed.11,15,26 • A new diagnostic category would have to be made for patients with a clinically definite stroke (i.e. symptoms for more than 24 h) but who have normal imaging, which could hardly be called a ‘TIA’. • There would be substantial implications of a tissuebased definition of TIA, in which patients with a TIA and a positive DWI would be reclassified as having a stroke, for epidemiological research and insurance policies.27 For example, studies of secular trends in stroke incidence and outcome would be confounded by the inclusion of patients with ‘TIA and brain imaging evidence of infarction’ as a ‘stroke’, whereas they

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 39

3.2 Definitions of transient ischaemic attack, stroke and acute stroke syndrome

were previously considered a TIA and not included as stroke in earlier studies. Further, patients with TIA and brain imaging evidence of infarction may now not be able to obtain life insurance, or their premiums may increase substantially.27 Critical illness insurance companies would even have to pay out for TIAs with DWI evidence of infarction that they would previously not have had to pay out for because the event would now be called a stroke; premiums would invariably rise to cover the new costs. At present there is no diagnostic test based on imaging or blood chemistry that is sufficiently sensitive, specific and widely available to reliably diagnose and exclude transient ischaemic attack and stroke.

(a)

We believe that patients with clinically definite TIAs (i.e. focal neurological symptoms lasting less than 24 hours with no other explanation other than a vascular origin), who have an appropriately sited and presumably ischaemic lesion on brain imaging, should have the fact noted (and considered to possibly be at increased risk of stroke).11,15,26 However, they should still be classified as having had a TIA, or maybe a new term for short duration symptoms of focal brain or ocular ischaemia (less than 24 h) with a relevant lesion on MRI scan (e.g. transient symptoms associated with infarction [TSI]),24 but definitely not a stroke. Our diagnostic criteria for TIA are summarized in Table 3.5. The presence of a presumed ischaemic lesion in the relevant part of the brain on CT or MR scan in a patient who presents with transient symptoms lasting less than 24 h should not change the diagnosis of a transient ischaemic attack (TIA) to a stroke.

(b) Fig. 3.3 (a) MRI (T2 weighted) of the brainstem (pons) in a patient with a recent episode of transient hemiparesis (right face, arm and leg) and conjugate gaze palsy to the left showing a very subtle region of altered signal intensity in the left paramedian pons. (b) MRI (DWI) in the same patient showing a very obvious area of high signal intensity (so-called ‘light bulb’ sign), consistent with early ischaemic change, in the left paramedian pons which was relevant to the patient’s focal neurological symptoms.

Although bright, high signal-intensity lesions (‘light bulbs’) on DWI and low apparent diffusion coefficient values have a high sensitivity (about 90%) and specificity (about 90%) for the diagnosis of acute focal cerebral ischaemia, the same findings have been reported in other diverse conditions such as focal brain haemorrhage, abscess and tumour (i.e. they are not 100% specific).

3.2.2 The definition of stroke The most widely accepted definition of a stroke is ‘a syndrome characterized by rapidly developing clinical symptoms and/or signs of focal, and at times global (applied to patients in deep coma and those with subarachnoid haemorrhage), loss of cerebral function, with symptoms lasting more than 24 h or leading to death,

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 40

Chapter 3 Is it a vascular event and where is the lesion? Table 3.5 Diagnostic criteria for transient ischaemic attack (TIA). Nature of symptoms Focal neurological or monocular symptoms Quality of symptoms ‘Negative’ symptoms, representing loss of focal neurological or monocular function (e.g. weakness, numbness, aphasia, loss of vision); rarely, ‘positive’ symptoms occur (e.g. pins and needles, limb shaking, scintillating visual field abnormality) Time course of symptoms Abrupt onset, starting in different parts of the body (e.g. face, arm, leg) at more or less the same time, without intensification or spread (‘march’); deficit maximal usually within a few seconds; symptoms resolve more gradually but completely, usually within an hour and, by definition, always within 24 h. Very brief attacks, lasting only seconds, are unusual except for transient monocular blindness (TMB) (note: we do not know how brief an attack of TMB can be and still be classified as TMB due to transient ischaemia; perhaps 10 s or so?) Associated symptoms TIAs usually occur without warning Antecedent symptoms are rare, but may reflect the cause (e.g. neck and face pain due to carotid dissection, headache due to giant-cell arteritis); otherwise, antecedent symptoms (e.g. headache, nausea or epigastric discomfort) usually suggest migraine or epilepsy Headache may occur during and after a TIA; it is to be distinguished from migraine headache Loss of consciousness is almost never due to a TIA; it usually suggests syncope or epilepsy Neurological signs Following symptomatic recovery, a few physical signs, such as reflex asymmetry or an extensor plantar response, which are not functionally significant, may be found Brain CT or MR scan The scan may show small areas of altered density, consistent with brain ischaemia or infarction, in a relevant part of the brain, or may have areas of hypodensity (on CT) or increased signal (on T2 weighted MR) remote from the symptomatic area Frequency of attacks TIAs often recur, but very frequent stereotyped attacks raise the possibility of partial epileptic seizures (sometimes due to an underlying structural abnormality such as an arteriovenous malformation, chronic subdural haematoma or cerebral tumour) or hypoglycaemia

with no apparent cause other than that of vascular origin’.28 This definition embraces stroke due to cerebral infarction (ischaemic stroke), non-traumatic intracerebral haemorrhage, intraventricular haemorrhage and some cases of subarachnoid haemorrhage (SAH). By convention, this definition does not include retinal

infarction, subdural haemorrhage, epidural haemorrhage, traumatic intracerebral haemorrhage or infarction, infection or tumour; nor does it embrace patients with intracranial venous thrombosis and SAH who are conscious and have a headache but no abnormal neurological signs. It is therefore important to be clear about what is, and what is not, included as ‘stroke’ in conversation, and in the literature. Since the clinical features, aetiology, prognosis and treatment of SAH are, for the most part, quite distinct from those of other forms of stroke,29,30 we think it is better to have a definition of stroke that does not include SAH, and so a completely separate definition for SAH. This is despite the fact that such a change in the definition of stroke would have implications for epidemiological studies of trends in incidence and outcome of stroke, just like changing the definition of TIA (see above), but in this instance would not be so dependent on the ever shifting sands of technology as changing the diagnosis of TIA to a tissue-based definition. Subarachnoid haemorrhage is therefore discussed separately below (section 3.7). The revised definition of stroke which we propose is ‘a clinical syndrome characterized by an acute loss of focal cerebral function with symptoms lasting more than 24 h or leading to death, and which is thought to be due to either spontaneous haemorrhage into the brain substance (haemorrhagic stroke) or inadequate cerebral blood supply to a part of the brain (ischaemic stroke) as a result of low blood flow, thrombosis or embolism associated with diseases of the blood vessels (arteries or veins), heart or blood’. Patients who are being assessed within 24 h of symptom onset and who still have focal neurological symptoms are temporarily classified as having a ‘brain attack’ (or something similar, such as an ‘acute stroke syndrome’ or ‘unstable brain ischaemia’). In practice, the absence of an obvious focal neurological deficit in a patient with a suspected stroke does not necessarily exclude the diagnosis. It may simply be a consequence of a delay in presentation, so that the signs have resolved, or it may be that the signs are rather subtle (but nevertheless functionally important to the patient) and have been missed. Examples include: • isolated dysphasia misinterpreted as confusion; • isolated visuospatial or perceptual disorder (e.g. dressing apraxia, geographical disorientation) which may not be noticed at the bedside examination, at least to begin with;31 • isolated amnesia or other subtle form of cognitive dysfunction;32

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 41

3.3 The diagnosis of a cerebrovascular event

• truncal and gait ataxia due to a cerebellar stroke which is only apparent (or elicitable) if the patient is asked to sit up, get out of bed and walk.33 These types of deficit may be missed (but should not be) in the emergency room, on busy ward rounds, during a hurried post-carotid endarterectomy or coronary artery bypass surgery assessment, and indeed any time.

3.2.3 The overlap between transient ischaemic attack and stroke, and the concept of an acute stroke syndrome (‘brain attack’ or ‘unstable brain ischaemia’) There is a continuum from TIA to ischaemic stroke in terms of duration of symptoms (see above).7 Moreover, patients with TIA and mild ischaemic stroke share a similar age and sex distribution, prevalence of vascular risk factors (and probably therefore pathogenesis) and long-term prognosis for serious vascular events although the short-term prognosis may differ.5 Thus, from the point of view of pathogenesis and treatment (secondary prevention), there seems no pressing need to distinguish TIA from ischaemic stroke, and indeed many trials of secondary prevention have included patients with TIA and non-disabling, mild ischaemic stroke because they are essentially the same condition. The problem in the era of increasingly rapid assessment and treatment of patients with acute cerebrovascular disease is how to use this time-based definition of TIA and stroke in patients who are being seen, and in some cases treated with potentially dangerous drugs (e.g. thrombolytics), within a few hours of the onset of symptoms. For example, if a hemiparetic patient is assessed 2·h after the onset of symptoms, an important question is whether this attack will recover and turn out to be a TIA, or not recover and become a stroke? There is no certain way of knowing, unless the patient is already recovering, but the longer the duration of symptoms of focal neurological dysfunction the more likely the deficit will persist,24 and the greater the risk of subsequent early stroke.9,10,34 Of course, for patients whose symptoms have resolved within 24 h of onset, they can be diagnosed retrospectively as having had a TIA. However, for those who still have symptoms within 24 h of onset, with or without relevant physical signs, it is appropriate to describe the acute presentation of focal cerebral ischaemia by a term such as a ‘brain attack’ or ‘acute stroke syndrome’ or ‘unstable brain ischaemia’.35 This emphasizes the need to rapidly exclude other differential diagnoses of TIA and stroke (e.g. hypoglycaemia, brain tumour), establish the pathological and aetiological subtype of the stroke and risk of recurrent stroke, and intervene with appropriate treatments that may

include reperfusion therapy for focal brain ischaemia, maintenance of physiological homeostasis, prevention of complications of stroke, early prevention of recurrent stroke and other serious vascular events, and early rehabilitation.36 ‘Time is brain’ in this setting. The anachronistic term ‘cerebrovascular accident (CVA)’ should be abandoned because it misleadingly implies that stroke is a chance event and that little can be done. Reasons to distinguish TIA and minor ischaemic stroke The are at least four reasons to distinguish TIA from minor ischaemic stroke. First, when formulating a differential diagnosis in clinical practice the differential diagnosis of focal neurological symptoms lasting minutes (e.g. epileptic seizures, migraine) is somewhat different from that of attacks lasting several hours to days (e.g. intracranial tumour, intracerebral haemorrhage), and in any event the reliability of the clinical diagnosis of stroke is much better than for TIA (section 3.6). Second, when conducting epidemiological studies of cerebrovascular disease consistency of diagnostic criteria is absolutely essential for comparing results over time and in different regions. Further, complete case ascertainment in incidence and prevalence studies is much less likely for TIA than stroke since patients who experience brief attacks are more likely to ignore or forget them, and are less likely to report them to a doctor than patients who suffer more prolonged or disabling events. Third, for case–control studies, there is less change in ‘acute phase’ blood factors related to thrombosis and tissue infarction, and there is, by definition, no survival bias amongst TIA patients compared with stroke patients. Fourth, distinguishing TIA from minor stroke can also aid assessment of case-mix in individual units and audits of management.

3.3 The diagnosis of a cerebrovascular event

The diagnosis of TIA and stroke is based on a constellation of clinical features that are thought to have a similar pathophysiology (i.e. caused by focal cerebral or ocular ischaemia or haemorrhage) (Table 3.2) and to be associated with similar outcomes (i.e. an increased risk of stroke and other major vascular events).

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 42

Chapter 3 Is it a vascular event and where is the lesion?

TIA and stroke are clinical diagnoses, based on a history of rapidly developing symptoms and signs of focal, and at times global, loss of focal brain function lasting 24 h respectively with no apparent cause other than that of vascular origin. The absence of a persistent neurological deficit does not exclude the diagnosis of TIA or stroke; such ‘negative’ findings may represent a delay in presentation, signs that have resolved, or subtle signs that have been missed (e.g. visual-spatialperceptual dysfunction). Similarly, the absence of a relevant acute lesion on brain imaging by CT, MRI or DWI does not exclude the diagnosis of TIA or stroke.

The assessment of patients with a suspected TIA or stroke depends on the time that has elapsed since the onset of symptoms. If the patient is assessed within 3– 6 h of stroke onset, the main focus is to establish the diagnosis of stroke, the pathological type and severity, and whether early reperfusion, or antiplatelet therapy and/or carotid endarterectomy may be indicated. If the patient is assessed (or reassessed) after this time, the focus is not toward reperfusion therapy but to ascertaining and minimizing the risk of recurrent stroke and the adverse sequelae and complications of the stroke. The timing of the assessment may also influence the reliability of the clinical assessment and accuracy of the diagnosis;37 focal neurological signs may resolve with time, which can make the diagnosis particularly difficult if the only signs were, for example, visual-spatialperceptual dysfunction. But it can also make the diagnosis easier if neurological signs such as drowsiness and dysphasia recover, so more history can be obtained from the patient. The first contact between clinician and patient is a crucial opportunity to conduct an appropriate history

Assessment History items Age > 45 years Absent history of seizure or epilepsy At baseline, not wheelchair bound or bedridden Blood glucose concentration between 2.8 and 22.2 mmol/L Physical assessment items Facial droop Arm drift Hand grip Speech Criteria for identifying stroke Presence of any physical assessment item All history items answered yes

FAST

LAPSS

and physical examination, and obtain relevant information from any observers, family, friends, the patient’s medical records and paramedical ambulance personnel (particularly when the patient is unable to communicate clearly due to dysphasia, depressed consciousness or knowledge only of a foreign language).38 The prior probability of a stroke among unscreened patients with neurologically relevant symptoms transported to an emergency department is about 10%.39 However, in some countries (e.g. Australia), paramedical ambulance personnel attend most patients admitted to hospital with stroke and they correctly identify about threequarters of stroke patients.40 However, because such personnel tend to overdiagnose stroke (not being aware of other conditions that mimic the symptoms of stroke) several prehospital screening tools have been developed, based on a few core clinical features (Tables 3.6, 3.7), to minimize the false positive diagnoses. These include: • the Face Arm Speech Test (FAST);41,42 • the Cincinnati Prehospital Stroke Scale (CPSS);43 • the Los Angeles Paramedic Stroke Scale (LAPSS);39 • the Melbourne Ambulance Stroke Screen (MASS).40 These tools have proved particularly helpful in rapid assessment of stroke patients ‘in the field’, and in communicating to regional stroke centres the imminent arrival of a patient with probable acute stroke.

History When a patient presents with a suspected transient ischaemic attack, ‘brain attack’ or stroke, the first question to answer is whether it really is a vascular event or not. This begins with, and depends on, a sound, carefully taken clinical history.

CPSS

X X X X

X X

X X X

X X

X X X X

X X X

X X

MASS

Table 3.6 Pre-hospital screening tools for the diagnosis of stroke: comparison of the Face Arm Speech Test (FAST), Los Angeles Paramedic Stroke Scale (LAPSS), Cincinnati Prehospital Stroke Scale (CPSS) and Melbourne Ambulance Stroke Screen (MASS).

X X X X X X

9781405127660_4_003.qxd 10/15/07 3:33 PM Page 43

3.3 The diagnosis of a cerebrovascular event Table 3.7 Motor and speech items assessed in several pre-hospital screening tools used for diagnosis of stroke.39–42 Assessment item Facial droop Patient smiles or shows teeth Arm drift Patient closes eyes and extends both arms for 10 s Hand grip Place a hand in each hand of the patient and ask him/her to squeeze hands Speech Patient repeats a sentence

Normal response

Abnormal response

Both sides move equally

One side does not move

Both arms move equally

One arm does not move or one arm drifts down, compared to the other

Both grip equally

Unilaterally weak or no grip

Normal language and articulation

Slurred or incorrect words, or unable to speak

When the patient is first assessed, take the patient and/ or eyewitness back to the onset of symptoms, recording their own words and not just your interpretation of them. This can usually be achieved by asking the three questions: • ‘When did it happen?’ • ‘Where were you when it happened?’ • ‘What were you doing when it happened?’ For clarification, it is always worth asking patients to describe their symptoms in an alternative way, particularly if the terms they use are rather vague, e.g. ‘dizziness’ or ‘heaviness’. Also it can sometimes be useful to ask patients whether they would have been able to do a specific task at the time of symptom onset; for example, if the patient describes an arm as being ‘dead’, asking whether they could lift the arm above their head would at least give a pointer as to whether the use of the word ‘dead’ was referring to a motor or just a sensory deficit. The use of certain terms is often culturally determined, and it must not be assumed that your interpretation of the term is the same as the patient’s. The most appropriate response if you are unsure is ‘what do you mean by that?’ or ‘try and describe what you mean in another way’. The history should obtain information about the following. • The nature of the symptoms and signs (sections 3.3.1– 3.3.7): – which modalities were/are involved (e.g. motor, sensory, visual)? – which anatomical areas were/are involved (e.g. face, arm, leg, and was it the whole or part of the limb; one or both eyes)? – were/are the symptoms focal or non-focal (Tables 3.2 and 3.3)? – what was/is their quality (i.e. ‘negative’, causing loss of sensory, motor or visual function; or ‘positive’, causing limb jerking, tingling, hallucinations)?

•

– what were/are the functional consequences (e.g. unable to stand, unable to lift arm)? The speed of onset and temporal course of the neurological symptoms (section 3.3.8): – what time of day did they begin? – was the onset sudden? – were the symptoms more or less maximal at onset; did they spread or progress in a stepwise, remitting, or progressive fashion over minutes/hours/days; or were there fluctuations between normal and abnormal function? Were there any possible precipitants (section 3.3.9)? – what was the patient doing at the time and immediately before the onset? Were there any accompanying symptoms (section 3.3.10), such as: – headache, epileptic seizures, panic and anxiety, vomiting, chest pain? Is there any relevant past or family history (section 3.3.11)? – have there been any previous TIAs or strokes? – is there a history of hypertension, hypercholesterolaemia, diabetes mellitus, angina, myocardial infarction, intermittent claudication, or arteritis? – is there a family history of vascular or thrombotic disorders? Are there any relevant lifestyle habits/behaviours (section 3.3.12)? – cigarette smoking, alcohol consumption, diet, physical activity, medications (especially the oral contraceptive pill, antithrombotic drugs, anticoagulants and recreational drugs such as amphetamines).

A record in the notes such as ‘no history available’ probably reflects laziness on the part of a doctor who has not tried fully to obtain it, rather than the real lack of any information.

9781405127660_4_003.qxd 10/15/07 3:33 PM Page 44

Chapter 3 Is it a vascular event and where is the lesion?

Examination The examination aims to: • confirm the presence of focal neurological signs, if any, anticipated from the history; • discover possible aetiological explanations for the event (e.g. atrial fibrillation, carotid bruits, cardiac murmurs, etc.), some of which may not be anticipated (e.g. malignant hypertension); • identify contraindications to investigation (e.g. a pacemaker – MR examination);

Symptoms

Neurological symptoms Acute onset Weakness Face Arm Leg Incoordination Limb Speech Visual disturbance Paraesthesia Face Arm Leg Vertigo Dizziness Nausea Vomiting Headache Confusion Loss of consciousness Convulsive seizures Neurological signs Weakness Face Arm Leg Visual field defect Eye movement abnormality* Dysphasia/dysarthria Visuospatial neglect Limb ataxia Hemiparetic/ataxic gait Sensory deficits Face Arm Leg

% of patients

• anticipate nursing and rehabilitation needs (e.g. impaired swallowing, urinary incontinence, immobility, pre-existing reduced visual or auditory acuity). A sophisticated knowledge of neurology is not needed to elicit and recognize the clinical features of a cerebrovascular event, as highlighted by the accurate paramedic and emergency room identification of stroke using simple stroke assessment tools (see above). However, a systematic approach is required as well as awareness of the discriminatory potential of symptoms and signs, in isolation and in groups (syndromes) (Table 3.8).44,45

Odds ratio

Stroke or TIA (n = 176)

Non-stroke (n = 167)

27.6

23 63 54

6 24 22

4.8 5.3 4.1

5 53 11

2 22 7

2.2 4.0 1.7

9 20 17 6 13 10 8 14 5 6 1

7 16 11 5 33 17 13 17 25 41 10

1.3 1.4 1.6 1.2 0.3 0.5 0.6 0.8 0.2 0.1 0.1

45 69 61 24 27 57 23 4 53

3 12 11 2 1 8 5 2 7

27.0 16.6 13.1 12.8 62.2 15.6 5.8 2.3 14.5

3 23 21

1 4 2

2.4 7.9 10.8

Table 3.8 Neurological symptoms and signs amongst patients presenting with suspected transient ischaemic attack and stroke in the Recognition of Stroke in the Emergency Room (ROSIER) study, subdivided into those occurring in patients with stroke or TIA and those with non-stroke.45

*Gaze palsy or opthalmoplegia. TIA, transient ischaemic attack.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 45

3.3 The diagnosis of a cerebrovascular event

The National Institutes of Health Stroke Scale (NIHSS) is a graded neurological examination which can be used as a template upon which to base a systematic neurological examination. It assesses neurological impairments such as level of consciousness, ocular gaze, visual fields, speech and language function, inattention, motor and sensory impairments, and ataxia, and thereby helps ensure that a reasonably thorough neurological examination is undertaken in the acute phase. It is quick to perform (taking less than a few minutes), is valid, and is reliable among neurologists, non-neurologist physicians, and non-physician coordinators in clinical trials for measuring neurological impairment and stroke severity.46 However, the scale was originally developed for use in trials of treatments for patients who had already been diagnosed with stroke (to measure impairments and grade stroke severity); it was not designed for, and has limited application in, diagnosing stroke and differentiating stroke from its mimics. A sophisticated knowledge of neurology is not needed to elicit and recognize the clinical features of a cerebrovascular event, but physicians must continually make efforts to refine their clinical abilities if the symptoms and signs are to be documented accurately and the diagnosis of stroke and its localization are to be optimized.

3.3.1 The nature of the symptoms and signs The symptoms and signs of TIA and stroke reflect the areas of the brain that are affected by the focal ischaemia or haemorrhage.45 For short duration events, such as TIA, the symptoms also reflect the activities in which the patient was engaged during the attack. For example, if the patient was not speaking or did not try to speak or read during a brief ischaemic event, it is impossible to know whether aphasia or alexia were present or not. Similarly, a weak leg may well not be noticed if the patient was sitting down. As many hours of wakefulness are spent in an alert state with eyes open, an upright posture and often speaking or reading, it is not surprising that most of the symptoms that TIA patients experience are of motor, somatosensory, visual or language function (Table 3.9).6,47,48 Other more transient activities such as swallowing and calculation are, not surprisingly, less frequently reported. Presumably TIAs, like strokes, can start during sleep, but the patient will be unaware of them if they have resolved before waking. Among patients admitted to hospital with symptoms of brain attack (defined as ‘apparently focal brain dysfunction of apparently abrupt onset’), only about twothirds are subsequently diagnosed with stroke; one-third

Table 3.9 Neurological symptoms during transient ischaemic attacks. % Unilateral weakness, heaviness or clumsiness Unilateral sensory symptoms Slurred speech (dysarthria) Transient monocular blindness Difficulty in speaking (aphasia) Unsteadiness (ataxia) Dizziness (vertigo) hom*onymous hemianopia Double vision (diplopia) Bilateral limb weakness Difficulty in swallowing (dysphagia) Crossed motor and sensory loss

50 35 23 18 18 12 5 5 5 4 1 1

From a series of 184 patients with a definite transient ischaemic attack (TIA) in the Oxfordshire Community Stroke Project.6 Many patients had more than one symptom (e.g. weakness as well as sensory loss) and no patients had isolated dysarthria, ataxia, vertigo, diplopia or dysphagia. Lone bilateral blindness was excluded from this analysis but later considered to be a TIA.48

are a stroke mimic.44 Clinical features that increase the odds of a final diagnosis of stroke include a definite history of focal neurological symptoms (odds ratio, OR = 7.2) and being able to determine the exact time of onset of symptoms (OR = 2.6) (Table 3.10).44 This is consistent with the common clinical criteria for a cerebrovascular event (i.e. abrupt onset of focal neurological symptoms or signs of a presumed vascular aetiology). A logistic regression model (Table 3.10) based on eight independent and significant predictors of the diagnosis of stroke (vs non-stroke) resulted in 83% correct classifications in the data set from which it was derived (i.e. it was internally valid).44 Other studies, using modern neuroimaging, have shown that the presence of acute facial weakness, arm drift and/or abnormal speech increased the likelihood of stroke, whereas the absence of all three decreased the odds.49 The Recognition of Stroke in the Emergency Room (ROSIER) scale has been developed and validated as an effective instrument to differentiate stroke from its mimics in the emergency room.45 It consists of seven items (total score from –2 to +5) comprising discriminating elements of the clinical history (loss of consciousness [score –1], convulsive seizures [score –1]) and neurological signs (face, arm, or leg weakness, speech disturbance, and visual field defect [each score +1]). A cut-off score >0 was associated with a sensitivity of 92%, specificity of 86%, positive predictive value of 88% and negative predictive value of 91% in the

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 46

Chapter 3 Is it a vascular event and where is the lesion?

Variable

95% CI

Known cognitive impairment An exact onset could be determined Definite history of focal neurological symptoms Any abnormal vascular findings* Abnormal findings in any other system** NIHSS = 0*** NIHSS 1–4 NIHSS 5–10 NIHSS > 10 The signs could be lateralized to the left or right side of the brain OCSP classification possible

0.33 2.59 7.21 2.54 0.44

(0.1–0.8) (1.3–5.1) (2.5–20.9) (1.3–5.1) (0.2–0.8)

1.92 3.14 7.23 2.03

(0.7–5.2) (1.03–9.6) (2.2–24.0) (0.9–4.5)

5.09

(2.4–10.7)

Table 3.10 Logistic regression model for predicting the diagnosis of stroke.44

The model gives a predicted probability of stroke (ranging from 0 to 1). *Systolic blood pressure > 150 mm Hg, atrial fibrillation, valvular heart disease, or absent peripheral pulses. **Respiratory, abdominal or other abnormal signs. ***National Institute of Health Stroke Scale (NIHSS) = 0 was entered as the reference group (therefore it does not have a coefficient). OCSP, Oxfordshire Community Stroke Project; OR, odds ratio; CI, confidence interval.

derivation data set (Table 3.8), and similar values in the test data set.45 No one symptom or sign can rule in or rule out the diagnosis of stroke and transient ischaemic attack.

3.3.2 Disturbance of conscious level Consciousness may be defined as ‘the state of awareness of the self and the environment’. Coma is the total absence of such awareness. Vascular diseases are probably the second most common cause of non-traumatic coma after metabolic/toxic disorders; up to 20% of patients with stroke – but not TIAs – may have some impairment of consciousness.44,45,47 (See also section 11.3.) Clinical anatomy Consciousness depends on the proper functioning of the ascending reticular activating system (ARAS). This is a complex functional, rather than anatomical, grouping of neural structures in the paramedian tegmentum of the upper brainstem, the subthalamic region and the thalamus (mainly the intralaminar nuclei). Focal lesions that impair consciousness tend to either disrupt the ARAS directly (i.e. mainly infratentorial lesions), or are large supratentorial lesions, which cause secondary brainstem compression or distortion (Table 3.11; Fig. 3.4).

Table 3.11 Causes of impaired consciousness after stroke. Primary damage to subcortical structures (e.g. thalamus) or to the reticular activating system in the brainstem (e.g. brainstem haemorrhage) Secondary damage to the reticular activating system in the brainstem (e.g. large supratentorial haemorrhage or infarct with transtentorial herniation and midline shift due to oedema) Coexisting metabolic derangement (e.g. hypoglycaemia, hypoxia, renal or hepatic failure) Drugs (e.g. sedatives) To be distinguished from normal consciousness but with impaired responsiveness due to: Locked-in syndrome Akinetic mutism Abulia Severe extrapyramidal bradykinesia Severe depression Catatonia Hysterical conversion syndrome Paralysis from neuromuscular disorders

Clinical assessment Conscious state is assessed by observing the patient’s spontaneous activity and their response to verbal, painful and other stimuli. The Glasgow Coma Scale (GCS) provides a structured way of describing conscious level, and is usually part of the standard ambulance and nursing observation forms (Table 3.12). Because it was developed for patients with head injury and so for more

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 47

3.3 The diagnosis of a cerebrovascular event

Septum pellucidum Crus of fornix Thalamus Corpus callosum Aqueduct Interventricular foramen

Cerebellum

Hypothalamus Midbrain Pons Fig. 3.4 Diagrammatic representation in the sagittal plane of the brainstem areas involved in consciousness (especially the ascending reticular activating system). Table 3.12 The Glasgow Coma Scale. Eye opening E1 None E2 To painful stimuli E3 To command/voices E4 Spontaneously with blinking Motor response (best response in unaffected limb) M1 None M2 Arm extension to painful stimulus M3 Arm flexion to painful stimulus M4 Arm withdraws from painful stimulus M5 Hand localizes painful stimulus in the face (reaches at least chin level). M6 Obeys commands Verbal response V1 None V2 Sounds but no recognizable words V3 Inappropriate words/expletives V4 Confused speech V5 Normal The score should be reported as: Ex, My, Vz, total score =· x·+·y·+·z·/·15.

diffuse rather than focal neurological deficits, care is needed when applying it to patients with stroke. The motor deficit must be assessed on the ‘normal’ side and not on the side with the motor deficit, and in the arm, not the leg, where the motor responses may be largely of spinal origin. The subscore of each item is probably more important than the total, since specific focal deficits, and particularly global aphasia, depress the overall score disproportionately to the level of alertness. The GCS has value as an initial prognostic indicator in acute stroke.50,51 It may also have value in monitoring

Ascending reticular activating system

IVth ventricle Medulla

the patient’s neurological status over time. Any deterioration in the GCS is a prompt to consider whether it is because of the progression of the neurological deficit or because of non-vascular factors, such as infection, metabolic disturbance, or the effect of drugs (section 11.5). However, it is important not only to document the GCS scores (and over time), but to describe the patient’s neurological impairments qualitatively and quantitatively (and over time) because the GCS measures only three of many important functions of the brain and frequently there is an obvious change in the patient on conventional neurological examination but not in the GCS. The Glasgow Coma Scale is an insensitive measure of neurological function and should not be the only measure to monitor the neurological status of the patient.

Clinical practice Almost instantaneous loss of consciousness caused by a stroke suggests either a subarachnoid haemorrhage (section 3.7) or an intrinsic brainstem haemorrhage. Loss of consciousness within a few hours of onset is usually due to brainstem compression by a large intracerebral haematoma, or cerebellar haematoma or infarct. Early impairment of consciousness after supratentorial infarction is unusual. This is because the associated cerebral oedema responsible for the mass effect, and so the midline shift and brain herniation usually takes 1–3 days to develop, although some evidence of transtentorial herniation may be present within 24·h. Postmortem studies initially showed that infarction of the complete middle

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 48

Chapter 3 Is it a vascular event and where is the lesion?

cerebral artery territory was needed before significant lateral and caudal (inferior) displacement of midline structures occurs. These findings have been confirmed in several imaging studies during life. Not surprisingly, level of consciousness is one of the best predictors of survival after stroke (section 10.2.7). Loss of consciousness during a TIA is extremely unusual and should prompt a search for alternative explanations such as hypotension (e.g. vasovagal and reflex syncope, cardiac arrhythmia), systemic disorders (e.g. hypoglycaemia) and generalized epileptic seizures.52,53 Even when transient loss of consciousness is followed by focal neurological signs, such as a hemiparesis, it is more commonly caused by an epileptic seizure resulting in a Todd’s paresis52 (section 3.4.2). When loss of consciousness does occur during a TIA, it seems to be associated with brainstem or bihemispheric ischaemia caused by either vertebrobasilar or bilateral carotid occlusive disease, respectively.53,54 A few cases may be due to ischaemia in the territory of small, perforating arteries which supply the upper brainstem including components of the reticular activating system. Impairment of consciousness must be distinguished from impaired responsiveness due to the following. The ‘locked-in’ syndrome is a state of motor deefferentation, where there is usually severe paralysis not only of the limbs but also of the neck, jaw and face. Indeed, the only muscles remaining under voluntary control may be those concerned with vertical eye movements and blinking. All this occurs with clear, and often extremely distressing, retention of awareness. The patient is unable to communicate by word or movement other than by blinking or moving their eyes up and down, but is fully aware of the surroundings and attempts to respond to them. Hearing, vision and often sensation are retained. There is usually an extensive, bilateral lesion in the ventral pons, which interrupts the descending motor tracts as well as the centre for horizontal eye movements in the pons, but the oculomotor nuclei and descending pathways for vertical eye movements are spared together with the ascending reticular activating system (Fig. 3.5). Cognitive functions are normal and so the patients must be given a full explanation of their predicament. Staff may need to be reminded to take appropriate account of the patient’s normal cognition and sensation, since prolonged survival in this state is possible.55,56

The relatives and staff caring for patients with the locked-in syndrome need reminding regularly that sensation, cognitive functions and awareness are all normal.

Fig. 3.5 A T1-weighted magnetic resonance (MR) image, sagittal plane, showing a ventral pontine infarct (arrows) in a patient with the locked-in syndrome.

Akinetic mutism and abulia are states where there is limited responsiveness to the environment, although the patients appear alert (or at least wakeful) in that their eyes are open and they follow objects. However, in contrast to the locked-in syndrome, the physical examination does not reveal evidence of a major lesion of the descending motor pathways. At its most extreme, patients with akinetic mutism may lie with open eyes, follow objects and become agitated or even say the occasional appropriate word following noxious stimuli (thus distinguishing this state from that of coma or the persistent vegetative state); but otherwise they do not respond to their environment. Occasionally, catatonic posturing may occur. If patients recover from this state, they have no recollection of it. ‘Abulia’ describes a less severe presentation of reduced spontaneous movement and speech. Such patients often appear to have a marked flatness of affect, but with adequate stimulation they can be shown to be conscious and have relatively preserved cognition. Both akinetic mutism and abulia occur with bilateral damage to the cingulate gyri, caudate nuclei and anterior limb of the internal capsules, but they can also occur with unilateral lesions of the caudate nucleus. Although these states are most commonly seen after head injury, following anterior communicating artery aneurysmal subarachnoid haemorrhage, or in multi-infarct states, they can occur with unilateral occlusion of the recurrent artery of Heubner.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 49

3.3 The diagnosis of a cerebrovascular event

3.3.3 Disturbance of higher cerebral function Higher cerebral functions can be divided into those which are ‘distributed’ – i.e. involve several areas of the cortex, such as attention, concentration, memory and higher-order social behaviour – and those which are more ‘localized’, such as speech and language, visuospatial function and praxis.32 Few tests are absolutely specific, however, for a single aspect of higher cerebral function. The nature of the cognitive assessment means that it is often appropriate to blend aspects of history taking with immediate confirmation by means of specific examination. Skilful examiners often weave their assessment into a relaxed conversation with the patients, making it more enjoyable for both. Many of the features of a brief cognitive assessment, an example of which is shown in Table 3.13,32 can be modified to suit this style of assessment. It is important to determine the handedness of an individual patient to guide which hemisphere is dominant for language. The descriptions below assume left-hemisphere dominance. Attention and concentration Attention and concentration are the ability to maintain a coherent stream of thought or action. They are not synonymous with wakefulness. clinical anatomy Attention and concentration are ‘distributed functions’ which depend on the integrated activity of the neocortex (predominantly the prefrontal, posterior parietal and ventral temporal lobes), the thalamus and brainstem. The reticular formation and other brainstem nuclei receive input from both ascending and descending pathways and there are then major ascending tracts to the thalamus, particularly its intralaminar nuclei. clinical assessment Failure of attention results in patients being unable to sustain concentration, and they are often reported to lack interest in things around them or to be tired or distractable. Another common complaint is that they have problems with memory. This may or may not be true, but from a practical point of view, if there is a significant disorder of attention, then extreme care is required when interpreting the results of tests of other higher-order functions such as memory. Attention and concentration can be assessed at the bedside (Table 3.14). clinical practice Acute stroke patients who appear inattentive should be assessed carefully to exclude an underlying focal

Table 3.13 Features of the 12-minute cognitive assessment.32 Orientation Time (day, date, month, season, year) Place Attention Serial 7s, or Months of the year backwards Language Engage in conversation and assess articulation, fluency, phonemic errors (e.g. ‘the grass is greed’ – Broca’s area lesion) and semantic errors (e.g. ‘the grass is blue’ – posterior perisylvian lesion) Naming of some low frequency items (e.g. stethoscope, nib, cufflinks, watch winder) Comprehension (of both single words and sentences) Repetition (e.g. emerald, aubergine, perimeter, hippopotamus; no ifs, ands or buts) Reading Writing Memory Anterograde: test recall of a name and address after 5 min Retrograde: ask about recent sporting or personal events Executive function Letter (F) and category fluency (animals): e.g. name as many words beginning with the letter F or A or S (>15 words per minute is normal) and animals (15 is low average, 10 is definitely impaired) Praxis Meaningful gestures (e.g. wave, salute) Luria three-step sequencing test (fist-edge-palm) Visuospatial Clock drawing, and overlapping pentagons General neurological assessment with particular attention to: Frontal lobe signs (grasp, pout, palmomental) Eye movements Presence of a movement disorder Pyramidal signs General impression Slowness of thought Inappropriateness Mood

Table 3.14 Bedside testing of attention and concentration. Digit span forwards and backwards* Recite months of the year, or days of the week, backwards Serial subtraction of 7s (although note that calculation ability needs to be intact) *The normal range is forwards: 6·± 1; backwards: 5 ± 1.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 50

Chapter 3 Is it a vascular event and where is the lesion?

neurological disturbance such as aphasia, visuospatial or perceptual disturbance, hemianopia, or amnesia. If they really are inattentive they may have a metabolic/ toxic encephalopathy, and an underlying treatable cause (e.g. hyponatraemia, hypoglycaemia, hypoxia, uraemia, dehydration, sepsis). Memory Memory function can be divided into ‘explicit’ memory (available to conscious access) and ‘implicit’ memory (relates to learned responses and conditioned reflexes). Explicit memory may be ‘episodic’ (dealing with specific events and episodes that have been personally experienced) or ‘semantic’ (dealing with knowledge of facts, concepts and the meaning of words, e.g. ‘stroke is a clinical syndrome’). Episodic memory (personally experienced events) comprises anterograde (newly encountered information) and retrograde (past events) components.32 Working memory refers to the very limited capacity which allows us to retain information for a few seconds.

Table 3.15 Bedside testing of memory. First check that patient is attentive (Table 3.14) and that language function is adequate (see below, and Table 3.16). Anterograde verbal memory Ask the patient to name three distinct objects (e.g. ‘Ball, Flag, Tree’ or ‘Boston, Car, Daisy’) Ensure that the patient has registered the information (repeat up to three times if necessary) If the patient can immediately name the objects, ask the patient to repeat the three objects three minutes later Anterograde visual memory Show the patient faces in a magazine Ensure they have recognized them Retest after 5 min Retrograde memory Ask the patient to describe recent events on the ward, or visits from relatives Ask about important historical events and major events in the patient’s life, e.g. date of marriage

The terms ‘short-term memory’ and ‘long-term memory’ are used loosely by clinicians and often rather differently by neuropsychologists. In patients with stroke, in whom the time of onset is known, it may be easier to distinguish anterograde amnesia (failure to acquire new memories) from retrograde amnesia (failure to recall previously learnt material). clinical anatomy Episodic memory depends on the hippocampaldiencephalic system, semantic memory on the anterior temporal lobe, and working memory on the dorsolateral prefrontal cortex.32 clinical assessment Patients with stroke are generally of an age when there is a natural decline in memory anyway (or have coexistent Alzheimer’s disease or vascular cognitive impairment). Therefore, many complain of memory problems before the stroke, and furthermore it is important to determine whether there really is a disturbance of memory (and not aphasia or disturbance of attention and concentration resulting in failure of registration of new information) and, if so, to identify what is a direct result of the stroke. A suggested method for assessing memory is set out in Table 3.15. clinical practice Perhaps the most frequent stroke lesion causing amnesia is infarction of the medial temporal lobe (Fig. 3.6).

Fig. 3.6 MRI DWI showing area of high signal intensity, consistent with infarction, in the left medial temporal lobe (arrow) and cerebral peduncle (arrowhead), due to occlusion of the left posterior cerebral artery at its origin.

Because the cause is usually occlusion of the posterior cerebral artery or one of its branches, the patient may have a coexistent visual disorder (e.g. hemianopia or upper quadrantanopia, colour anomia, visual agnosia). A pure amnesic syndrome may be caused by a vascular lesion involving the mammillothalamic tract or anteriomedian territory of the thalamus (supplied by the polar

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 51

3.3 The diagnosis of a cerebrovascular event

Fig. 3.7 MRI DWI showing area of altered signal intensity, consistent with haemorrhagic infarction, in the left thalamus (arrow).

and paramedian arteries) such as the anterior parts of the dorsomedian nucleus57,58 (Fig. 3.7). In general, there is relative sparing of verbal memory with right thalamic lesions and visuospatial memory with left thalamic lesions, although the amnesia may be global with unilateral lesions. Particularly severe amnesia is more likely with paramedian thalamic infarction, which is frequently bilateral because the left and right paramedian arteries arise from one stem in many people (Fig. 3.8).58 In most cases of thalamic amnesia there are also signs of upper midbrain dysfunction, such as somnolence, vertical gaze palsies and corticospinal and spinothalamic tract signs. The syndrome of transient global amnesia is described in detail in section 3.4.3. Speech and language Language is difficult to define, but may be considered as a system for the expression of thoughts and feelings by the use of sounds and/or conventional symbols. It involves the production (or expression) and comprehension (or reception) of speech, as well as reading and writing. • Aphasia/dysphasia is an acquired disorder of the production and/or comprehension of spoken and/or written language. Subtypes of aphasia include speech (verbal) apraxia, alexia, agraphia and anomia.

Fig. 3.8 A T2-weighted MR image, axial plane, showing bilateral thalamic infarction (arrows) in a patient with severe global amnesia.

• Speech apraxia/dyspraxia is a syndrome in which there is variable misarticulation of single sounds, in the absence of dysarthria (in which there is constant misarticulation). This is because of impairment of planning of movements required for speech sounds and articulation, despite the ability to articulate speech being intact (e.g. similar to gait apraxia). • Alexia/dyslexia is an inability to name or interpret previously learned printed symbols. The patient can see individual letters, but cannot decode a series of letters into a recognizable word. When it occurs in isolation, this is sometimes referred to as ‘word blindness’. • Agraphia/dysgraphia is an acquired disorder of writing, a subtype of aphasia. • Anomia/dysnomia is an inability to generate a specific name. In the context of a vascular event, it is usually a manifestation of aphasia, but can be an amnesic disorder. • Anarthria/dysarthria is a disorder of articulation of single sounds. • Dysphonia is defined as a disorder of phonation of sounds.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 52

Chapter 3 Is it a vascular event and where is the lesion?

Disorders of speech and language are not synonymous; speech involves language and articulation, whereas language involves reading and writing as well as speech.

clinical anatomy In general, non-fluent (expressive, Broca’s) aphasia is a syndrome with elements that range from initial mutism to speech apraxia and the classical pattern of agrammatism (inability to produce grammatical or intelligible speech, usually with simplified sentence structure – telegraphic speech – and errors in tense, number and gender). It is likely to be due to a lesion involving the posterior, inferior, dominant frontal lobe cortex and subcortex. The articulatory component (apraxia) is most heavily represented in the left insular region. Lesions of the dominant thalamus may also result in predominantly non-fluent aphasia. Fluent (receptive, Wernicke’s) aphasia is commonly caused by a more posterior lesion involving the temporal lobe cortex and subcortex. Most patients with stroke have a combination referred to as ‘mixed aphasia’ (or, if severe, ‘global aphasia’), due to more extensive lesions within the dominant hemisphere. Consequently, there is often an associated right hemiparesis and hemianopia. Occasionally, patients with non-fluent aphasia have preserved repetition. This is termed transcortical motor aphasia, and it is usually caused by lesions restricted to the anterior cerebral artery territory and which spare the arcuate fasciculus, between Broca’s area and Wernicke’s area (section 4.2.2). Fluent aphasia with normal repetition (transcortical sensory aphasia) occurs with strokes in the left temporo-occipital region. Dysarthria may be a result of cerebellar (ataxic), pyramidal (spastic), extrapyramidal (hypokinetic), or facial nerve (flaccid) dysfunction. Anarthria may occur as part of a pseudobulbar palsy caused by bilateral lesions of the internal capsule (not necessarily at the same time) or with a single lesion involving both sides of the brainstem. Alexia, with or without agraphia, may result from strokes that involve the medial aspect of the left occipital lobe and the splenium of the corpus callosum. There is usually a right visual field defect but no hemiparesis, and it is thought that the lesion in the splenium interrupts the transfer of visual information from the normal left visual field (right occipital lobe) to the damaged left hemisphere language areas. Gerstmann syndrome is the combination of aphasia, agraphia, right-left disorientation, and acalculia, due to a lesion in the region of the angular gyrus of the dominant hemisphere.59

clinical assessment The first distinction to make is between aphasia/ dysphasia (a disorder of language), anarthria/dysarthria (a disorder of articulation) and dysphonia (a disorder of phonation). If the patient’s speech sounds ‘like a drunk’, and if the ability to understand and express spoken and written language is preserved, then the problem is dysarthria (or dysphonia). If the main difficulty is understanding or expressing spoken or written language – such as difficulty in reading (the patient can see the letters but cannot make sense of them); difficulty in writing, even though the use of the hand is otherwise normal (often not the case); or difficulty in producing sentences, with words not being in their proper place or even non-words being used – then the problem is aphasia. Table 3.16 sets out a scheme of bedside testing that will detect most speech and language problems. There is a general tendency in everyday clinical practice to underestimate the receptive component of aphasia, particularly if the examiner does not go beyond questions requiring a yes/no answer, or simple social conversation. clinical practice Beware labelling a patient as dysphasic when a lack of other symptoms and signs suggest isolated non-dominant hemisphere dysfunction. ‘Crossed’ aphasia is a disturbance of language which occurs from a right hemisphere lesion in a right-hand dominant patient and is seen in about 4% of such patients. It is presumed that some right-handed patients have mixed cerebral dominance for language, but other causes include bilateral strokes (including the thalamus) and previous strokes. It is worth noting that many righthanded patients with right hemisphere strokes show subtle alterations in the affective aspects of speech, such as intonation (aprosody). Isolated dysarthria may be the only manifestation of a lacune at the genu of the internal capsule or in the corona radiata. In such cases, there is specific impairment of corticolingual fibres.60 Foreign accent syndrome is a rare, acquired disorder of speech in which native speakers listening to a patient speaking their language describe hearing a foreignsounding accent – yet the patient may never have been exposed to any other language or dialect before the stroke. It is probably due to an inability to make the normal phonetic and phonemic contrasts of the native language. The syndrome has most often been associated with small, subcortical infarcts in the left cerebral hemisphere.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 53

3.3 The diagnosis of a cerebrovascular event Table 3.16 Bedside testing of language function. First ensure any hearing aid has a battery, is switched on and that appropriate, clean spectacles are worn Also check that you are using the patient’s native language – if not, use an interpreter Spontaneous speech Consider output (whether fluent or non-fluent), articulation and content: during history-taking and for a structured task (e.g. ‘describe your surroundings’) Auditory comprehension Simple yes/no questions (e.g. Is Russia the capital of Moscow? Can dogs fly? Do you put your shoes on before your socks?) Give commands (being careful not to use non-verbal cues) of one, two and three steps using common objects, such as the manipulation of three different-coloured pens (care not to require the use of limbs with significant weakness or apraxia) Naming Ask the patient to name objects, parts of objects, colours, body parts, famous faces (certain groups, particularly the naming of people, may be more severely affected) If visual agnosia is present, use auditory/tactile presentation, e.g. bunch of keys Repetition ‘West Register Street’ (difficult if dysarthric) ‘No ifs, ands or buts’ (difficult if aphasic) Reading Aloud, e.g. from a book or newspaper Comprehension of the same piece Writing Spontaneous (‘why have you come into hospital?’) Dictation (‘the quick brown fox jumped over the lazy black dog’) Copying Articulation Ask the patient to say: p/p/p/p/p/p (labial sounds, which test the orbicularis oris) t/t/t/t/t/t (lingual sounds, which test the anterior tongue) k/k/k/k/k/k (palatal sounds, which test the posterior tongue and palate) p/t/k/p/t/k (tests the overall coordination of sounds)

Visuospatial dysfunction Many patients with stroke fail to respond to stimulation of, or to report information from, the side contralateral to the cerebral lesion. There are two broad categories of neglect: intrapersonal (i.e. with respect to the patient’s own body) and extrapersonal or topographical (i.e. with respect to the surrounding environment). There are a number of different types and/or degrees of severity of neglect in patients with stroke, and in many, a combination of somatic sensory deficits and disturbed visual perception contribute to the clinically apparent ‘neglect’,

Table 3.17 Glossary of terms describing disorders of visuospatial function. Hemi-inattention: where the patient’s behaviour during examination suggests an inability to respond appropriately to environmental stimuli on one side, e.g. people approaching, noise, or activity in the ward Sensory or tactile extinction: where the patient fails to register a tactile stimulus (light touch) of adequate intensity on one side of their body when both sides are stimulated simultaneously and adequately (i.e. double simultaneous stimulation) but where the stimulus has been registered when each side was stimulated separately Visual inattention or extinction: where the patient fails to register a visual stimulus (e.g. finger movement) in one hom*onymous visual field (half field or quadrant) when the same stimulus is presented to both fields simultaneously, but where the patient had no field defect on normal testing Allaesthesia: where the patient consistently attributes sensory stimulation on one side to stimulation of the other; this is related to right/left confusion, where the patient consistently moves the limbs on one side when requested to move the limbs on the other Anosognosia: denial of a sensorimotor hemisyndrome Anosodiaphoria: indifference to/unconcern about a sensorimotor hemisyndrome Asomatognosia: lack of awareness of a body part Somatoparaphrenia (non-belonging): lack of ownership of a paralysed limb Experience of supernumary phantom limbs: reduplication of limbs on the affected side of the body Personification: nicknaming a limb and giving it an identity of its own Misoplegia: the morbid dislike or hatred of paralysed limbs in patients with hemiparesis61 Related phenomena seen in parietal lobe dysfunction: Astereognosis: unable to recognize objects placed in the affected hand yet cutaneous sensation is preserved Agraphaesthesia: unable to identify a number drawn on the palm of the affected hand yet cutaneous sensation is preserved Geographical disorientation: where the patient becomes lost in familiar surroundings despite being able to see Dressing apraxia: unable to dress, or dresses inappropriately, despite having no apparent weakness, sensory loss, visual or neglect problems; this is occasionally seen in an isolated form and probably occurs because of a combination of disordered body image, sensory and visual inattention rather than being a true apraxia

hence the use of the broader term ‘visuospatial dysfunction’. Table 3.17 provides a glossary of the terms that are used,61 Fig. 3.9 shows an example of anosognosia; this patient denies the presence of weakness of the left leg62

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 54

Chapter 3 Is it a vascular event and where is the lesion?

Fig. 3.9 Neglect and anosognosia: the photograph on the left shows how someone may normally view their hands and legs (before a stroke). The photograph on the right shows how the

same person may view their hands and legs after a major right hemisphere stroke, causing neglect and anosognosia for the left hand and left leg.

Fig. 3.10 Somatoparaphrenia (non-belonging): the photograph on the left shows how someone may normally view their hands and legs (before a stroke). The photograph on the right shows how the same person may view their hands and legs after a major right hemisphere stroke causing

somatoparaphrenia (i.e. the man denies ownership of the paralysed arm and leg on the left side of the body, and even thinks the left arm and leg belong to another person, such as his wife – note the different left hand, wedding ring on the left hand, and the different left leg).

and Fig. 3.10 shows an example of somatoparaphrenia (non-belonging); this patient denies ownership of the paralysed leg on the left side of the body, and even attributes the left leg to another person.

artery occlusion.63 Although it can occur with dominant hemisphere lesions, when it does so detection is often hindered by coexistent language disturbance and inability to use the dominant hand.

clinical anatomy Visuospatial dysfunction is most severe with posterior parietal lesions of the non-dominant hemisphere, particularly those that extend to the visual association areas. Among patients with ischaemic stroke, the aetiology may be middle cerebral or posterior cerebral

clinical assessment Relatives may report little more than ‘confusion’ or ‘difficulty in dressing’. If visuospatial problems are suspected from the history, they should be carefully sought in the examination. Simply observing how patients respond to their environment and carry out tasks, such

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 55

3.3 The diagnosis of a cerebrovascular event

Fig. 3.11 A letter from a patient with left visual neglect, showing neglect of the left side of the page.

as writing (Fig. 3.11), copying flowers (Fig. 3.12) and drawing a clock face (Fig. 3.13) can be revealing. An obvious example would be if a patient (without a hemianopia) does not register the doctor’s presence when approached from one side, even when spoken to. Or they might be unable to find their way back to their hospital bed after being taken to the toilet, suggesting geographical disorientation. The nurses and therapists are often better placed than the doctor to identify visuospatial problems, so it is important that staff are trained to recognize and report them to other members of the team. Table 3.18 sets out a bedside examination that should detect significant visuospatial dysfunction. Of the many cancellation tasks available, the star cancellation test is easy to use and probably the most sensitive (Fig. 3.14). Using two or three different tests increases the sensitivity of detecting visual neglect, but this may not always be practical in the setting of acute stroke.64 Many aspects of these assessments of visuospatial function require subjective judgements to be made by the physician – which probably accounts for the relatively poor inter-observer reliability. Although many other tests to identify and quantify visuospatial dysfunction have been described, the ‘gold standard’ against which the tests are evaluated is often regarded as the functional assessment by an occupational therapist.

Fig. 3.12 Abnormalities of copying flowers. Five patients with right hemisphere lesions were asked to copy pictures (a) and (b) in panel 1. The figures illustrate the variation seen in copying tasks. Panel 2: this patient has mainly neglected the information on the left side of the page. Panel 3: this patient has omitted the left-hand components of the objects, but has shifted attention to the right-hand side of another object placed to the left of the neglected space. Panel 4: this patient has drawn the right side of both flowers in the pot, but has completely neglected the left of the two separate flowers. Panel 5: this patient has transposed objects in the left field to the right field, i.e. both flowers are drawn on a single stem. Panel 6: this patient has produced a ‘hallucinatory’ rabbit in the left field when copying the separate flowers; this has been termed ‘metamorphopsia’. Table 3.18 Bedside tests of visuospatial function. Is the patient aware, and reacting appropriately to their deficit? Observe the patient’s response to the environment Observe the patient’s ability to carry out a specific task Check for sensory and visual extinction Copy a simple picture, e.g. a flower (Fig. 3.12) Draw a clock face and put the numbers in (Fig. 3.13); this may not be specific for visual neglect but rather reflect other cognitive problems, such as dementia Perform the star cancellation test (Fig. 3.14)

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 56

Chapter 3 Is it a vascular event and where is the lesion?

Fig. 3.13 (a) A drawing of a clock face by a patient with left visuospatial disturbance after a stroke, showing crowding of the digits on the right and neglect of the left side of the clock face. (b) A drawing of a clock face by a confused, elderly patient who has not had a stroke. The crowding of digits on the right occurs

TEN GET

DAY

E J

HER

MAN

READ

N O

STAR M

LEG

ARE Fig. 3.14 Star cancellation test. The chart is placed in front of the patient, who is asked to cross out all the small stars while ignoring the large stars and letters (with permission from Thames Valley Test Company, 7–9 The Green, Flempton, Bury St Edmunds, Suffolk IP28 6EL, UK).

An elderly man who lived alone reported that he woke one morning and thought that there was ‘something in bed with me’. He said it felt warm, and was pressing against the left side of his body. He thought his cat had got into bed with him, but when he touched it with his right hand he realized that it was his left arm. He had had a right parietal infarct during the night. One can often deduce the presence of visuospatial problems by simply observing how the patient responds to the environment and carries out tasks around the ward.

because of failure of planning. Many normal people insert 12, 3, 6 and 9 before other numbers. (c) Rather bizarre drawing of a clock face by a patient with strokes affecting both cerebral hemispheres.

clinical practice When there is an isolated problem with visuospatial function – i.e. when it is not accompanied by a more easily recognized ‘stroke’ deficit such as weakness – the patient’s behaviour may seem extremely bizarre, and even be interpreted as psychiatric disease. An elderly man was escorted to hospital by his concerned passenger after he drove his car along a street unaware that he was scraping along a whole row of other cars on the left-hand side. Examination revealed normal visual fields (when each eye was tested in turn) but visual inattention to the left (when the visual fields of each eye were tested simultaneously). There was also somatosensory extinction on the left when the arms were touched simultaneously. Brain CT scan showed a small right parietal haemorrhage. A middle-aged single lady was flying home from holiday when she became ‘confused’. On disembarking from the aircraft, she was staggering to the left, appeared unable to follow the signs to the customs point, and could not find her passport in her left-hand jacket pocket. She was held initially by the police on suspicion of alcohol or drug intoxication, but then admitted to a psychiatric hospital. It was only a week later, when she had a transient ischaemic attack affecting power in her left hand, that a right parietal infarct and severe stenosis of the right internal carotid artery were discovered. Perhaps not surprisingly, there is some evidence that patients who have varying degrees of indifference to their stroke are more likely to delay seeking medical

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 57

3.3 The diagnosis of a cerebrovascular event

attention.31,33 Visuospatial problems are a major cause of disability and handicap, and impede the patient’s functional recovery (section 11.28).

d b

Disorders of praxis

Apraxia is defined as inability to perform learned movements that cannot be explained by weakness, sensory loss, incoordination, inattention and other perceptual disorders, or by failure to understand the command. Dressing and constructional apraxias are best considered as disorders of visuospatial function rather than true apraxias (see above). Although the terms ‘verbal apraxia’ or ‘speech apraxia’ may be used by speech and language therapists when there are repeated phonemic substitutions, in practice such patients usually also have evidence of aphasia and/or dysarthria. clinical anatomy It is thought that the programmes of learned movements (engrams) are maintained predominantly in the left temporoparietal cortex. Messages then pass to the left premotor frontal cortex and finally, via the anterior corpus callosum, to the right premotor frontal cortex (Fig. 3.15). Lesions in the posterior left hemisphere may result in bilateral apraxia (because the message is not transmitted); those in the premotor areas are usually associated with a hemiparesis and therefore apraxia may only be apparent in the non-paralysed limbs; and finally, a lesion of the corpus callosum may cause isolated apraxia of the left limbs, with normal function of the right limbs. clinical assessment Apraxia should always be considered as a potential explanation for disparity between the degree of deficit as tested at the bedside (when one often gives the patient relatively simple commands) and much more severely impaired functional abilities when the patient is observed around the ward (e.g. dressing, swallowing, speaking). Patients have difficulty with miming actions, imitating how an object is used, and even making symbolic gestures. However, at other times they may be observed making the individual movements that would be needed to perform the action. In general, they will have most difficulty in miming the action, less difficulty in imitating the examiner, and least difficulty when actually given the object to use. The more sequences there are to the action, the more difficult it is and so the more sensitive the test. These relatively common problems are sometimes referred to as ‘ideomotor apraxias’, and can be distinguished from ideational apraxias when the patient has difficulty in performing a sequence of

Right

Left

Fig. 3.15 Diagrammatic representation of the lesions that can result in apraxia. (a) The dominant temporoparietal cortex is probably the site of the programmes of learned movements (engrams). Lesions here result in bilateral apraxias, due to failure to transmit the information to both frontal lobes. The clinical signs may be difficult to identify because the interpretation of the command may be affected by receptive aphasia. (b) Although lesions of the dominant frontal lobe are often associated with expressive dysphasia, comprehension is usually relatively spared. The apraxia may only be apparent in the left limbs, since there will usually be a right hemiparesis. (c) Lesions in the anterior corpus callosum may result in isolated apraxia of the left limbs, because of failure of transmission of the motor information to the right frontal lobe, while the right arm and leg move normally. (d) Lesions of the non-dominant frontal lobe are not normally associated with clinically apparent apraxias, because there is usually a left hemiparesis.

movements even though the individual movements can be performed normally. However, the latter probably occurs very rarely in a pure form, and the distinction between ideomotor and ideational apraxias is of little value to clinicians. Table 3.19 suggests ways of screening for apraxia. clinical practice In patients with stroke, the main problem is being sure that the patient has understood the command, because lesions of the relevant areas will often result in aphasia. Nevertheless, 80% of patients with aphasia also have evidence of apraxia with an imitation test (i.e. no verbal command). Because apraxic patients may perform actions reflexly that they are unable to do when asked, this should not be misinterpreted as a sign of a hysterical conversion disorder.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 58

Chapter 3 Is it a vascular event and where is the lesion? Table 3.19 Bedside testing of praxis. Limbs Ask the patient to: Mime the use of a pen, comb and toothbrush Imitate the examiner’s use of the same objects Use the actual objects Orofacial Ask the patient to: Whistle Put tongue out Blow out cheeks Cough Serial actions Ask the patient to: Mime putting the address on a letter Then seal it Then put a stamp on it

3.3.4 Disturbance of the motor system Clinical anatomy Particular areas of the motor cortex, when stimulated, result in movement of a particular body part. This localization of function is traditionally portrayed by the homunculus (or manikin).65 Although the absolute neuroanatomical relationships may be incorrect, it

Ha nd Jaw T Sw ongue allo wi ng

Vocalization

Lips

Sal ivat ion

s kle An

e dl x id M Inde b um Th eck N w Bro ll a yeb e and e d i l Fac Eye

ca tio n

Wrist

Elbow

Shoulder Trunk Hip ee Kn

To es

is still useful as an aide-memoire (Fig. 3.16). The corticospinal tract (Fig. 3.17) descends from the primary and more anterior supplementary motor cortex, the fibres converging in the corona radiata. The fibres then pass through the internal capsule. The traditional view is that those relating to the head pass through the anterior limb; those relating to the mouth, larynx and pharynx are in the genu; those relating to the arm are in the anterior part of the posterior limb; while those relating to the leg lie more posteriorly. In fact, the fibres almost certainly follow an oblique course through the capsule, becoming progressively more posteriorly placed in the caudal (inferior) segments of the capsule. The fibres then pass into the brainstem. Here, the fibres that originate in the precentral gyrus lie in the cerebral peduncles of the midbrain and the base of the pons before entering the medullary pyramids (i.e. the pyramidal tracts). The facial nerve nucleus in the pons has a rostral portion from which fibres innervate the muscles of the upper face, while the more caudal portion of the nucleus supplies fibres to the muscles of the lower face. The caudal loop of the fibres to the facial nerve descends as far as the medulla and explains why lesions of the medullary pyramid or medial medulla can be associated with contralateral upper motor neurone-type facial weakness.66,67 Most fibres in the corticospinal tract decussate in the lower medulla and come to lie in an anterolateral

Lit t Ri le ng

ti as M

Fig. 3.16 Topographic organization of the motor cortex in the cerebral hemisphere, coronal view (after Penfield & Rasmussen, 195065).

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 59

3.3 The diagnosis of a cerebrovascular event

Hand

Patients use terms such as ‘heaviness’ and ‘numbness’ interchangeably; further questioning is needed to distinguish between motor and sensory deficits.

Leg Face

Internal capsule

Corticospinal tract Corticobulbar tract Midbrain Cerebral peduncle

Pons

Pyramid

Spinal cord

Medulla oblongata

Lateral corticospinal tract (crossed) Anterior corticospinal tract (uncrossed)

Fig. 3.17 Diagrammatic representation of the corticospinal and corticobulbar tracts.

position in the spinal cord, although a variable proportion remain uncrossed. These uncrossed fibres project to motor neurones in the medial part of the ventral horns, subserving axial and proximal muscles, corresponding with movements of the trunk, or of two limbs together. The uncrossed corticospinal fibres cannot be invoked to explain residual function in distal parts of an otherwise plegic limb, nor deficits in contralateral limbs. Clinical assessment Motor symptoms are usually described as ‘weakness’, ‘heaviness’ and ‘clumsiness’. They are often accompanied by sensory symptoms of some sort, which can lead to diagnostic confusion because a purely weak limb may be described by the patient as ‘numb’ or ‘dead’. Descriptions such as ‘heaviness’ and ‘numbness’ should not simply be accepted as evidence of motor and sensory disturbance, respectively. In our experience, the terms are used interchangeably (and are often culturally determined), and a little more questioning is often required.

In patients with suspected TIA, unilateral facial weakness is probably under-reported, because they do not realize they have had facial weakness unless they have seen themselves in the mirror, or were seen by someone else. If there is a clear history of slurred speech, but no symptoms of cerebellar or bulbar dysfunction, it is reasonable to suspect facial weakness, because this may cause dysarthria. However, care should be taken before accepting a patient’s or relative’s description of the side of a facial weakness. They should be asked, ‘Which side dropped?’ and ‘Did saliva trickle from one side of the mouth?’ Upper motor neurone facial weakness affects the lower half of the face, while function of the forehead muscles is relatively preserved because of bilateral innervation of the forehead muscles. Mild weakness may only be apparent by observing asymmetry of the nasolabial folds. If the examiner is uncertain whether a facial weakness is present, or whether it is simply the normal sideto-side asymmetry, it may be useful to ask the patient to attempt to whistle, an action that requires fine control of the facial muscles. A mild upper motor neurone facial weakness can be overcome during emotionally generated movements, e.g. a smile. Patients sometimes complain of being ‘generally weak’. This should be viewed as a non-focal neurological symptom, as it is rarely described when they strictly mean motor weakness. It is sometimes used as a term for fatigue, tiredness, lethargy and just occasionally loss of balance. For the clinician, the difficulties with the physical examination lie not with the densely hemiplegic patient with increased tone, brisk deep tendon reflexes and an extensor plantar response, but rather with the patient with a mild neurological deficit. Subtle abnormalities of motor function may be detectable in the hand at a time when there is no objective weakness. Impairment of fine finger movements (or rapid alternating hand movements) is a sensitive clinical test of corticospinal function. This equates with functional problems reported by patients who, in the presence of normal power, often have difficulty with delicate motor tasks such as doing up buttons or controlling a pen and so may describe the problem as ‘clumsiness’. Of course, they are much more likely to notice this in their normally dominant hand. Impairment of fine finger or rapid hand movements is probably the most sensitive clinical test of corticospinal function.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 60

Chapter 3 Is it a vascular event and where is the lesion?

Drift from the horizontal of the outstretched arm with the eyes closed, and rapid tapping of the foot against the examiner’s hand, are both good screening tests of motor function in the arm and leg, respectively, but neither test is very specific.

Fig. 3.18 Downward drift of the outstretched left arm in a patient with a right corticospinal tract motor deficit causing a mild left hemiparesis.

Drift of the outstretched pronated arm with the eyes closed is a good screening test of motor function (Fig. 3.18). However, there are several other potential causes including loss of proprioception, when the fingers tend to move independently – so-called ‘piano-playing’ or ‘pseudoathetosis’; neglect, when there tend to be much larger-amplitude movements, including upwards; or cerebellar dysfunction, when there tend to be largeramplitude oscillations, particularly if sharp downward pressure is applied to the arm. Thus, as a screening test for motor dysfunction, it is quite sensitive but not very specific and should be used in conjunction with the examination of fine finger movements. Some clinicians favour performing the test with the patient’s palms facing upwards, and if asymmetrical internal rotation occurs without downward drift this is taken as a sign of very subtle motor dysfunction. Minor motor deficits affecting the leg are probably best detected by drift of the leg (when flexed at the hip against gravity in a supine patient) and rapid tapping of the foot against the examiner’s hand. Additionally, the patient’s gait should be carefully observed.

The pattern of weakness within an individual limb is traditionally taught to be of localizing value. In particular, when the antigravity muscles (i.e. shoulder abductors; elbow, wrist, and finger extensors; hip and knee flexors; and ankle and toe dorsiflexors and ankle everters) are weaker than their counterparts (i.e. shoulder adductors; elbow, wrist, and finger flexors; hip and knee extensors; and ankle and toe flexors and ankle inverters), this is often described as a ‘pyramidal distribution’ of weakness. It has been suggested that this pattern is simply a function of the intrinsically greater strength in antigravity muscles together with the effects of hypertonia – the actual pattern of weakness being equally common in patients with central or peripheral lesions causing muscle weakness.68 The deep tendon reflexes were considered of more value as a localizing feature,68 although the inter-observer reliability of two standard scales for grading tendon reflexes is probably no better than ‘fair’ (kappa < 0.35).69 While the anatomical extent of the motor deficit is important for clinicoanatomical correlation, the severity is helpful in the acute phase for determining prognosis, the potential risks and benefits of interventions (such as thrombolysis) and the functional management and rehabilitation of the patient. There are several methods of quantifying the severity of motor weakness, such as the Medical Research Council (MRC) scale, NIHSS, and Scandinavian Neurological Stroke Scale which have an operational definition of the grades of weakness, and moderately good inter-observer reliability.46 In addition a description of some actions the patient can and cannot perform (e.g. holding a cup of water, combing hair) is helpful when trying to understand the problems that the patient is having and for appropriate goal setting during rehabilitation. And, for assessing change, it can be helpful to ask the patient, or note, what they can just do (e.g. can just extend the fingers against gravity, walk 10 m in 15 s); and worsening or improvement can then be easily observed, even if the weakness appears to have the same MRC grade. Attention should focus on the anatomical extent of the weakness and the functional consequences, rather than solely trying to grade the severity with a motor scale. An extensor plantar response is only one part of a nociceptive spinal flexion reflex, which in its complete

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 61

3.3 The diagnosis of a cerebrovascular event

(a)

Fig. 3.19 The Babinski sign: (a,b) evoked in this case by stroking the lateral part of the dorsum, rather than the sole, of the foot, in order to avoid voluntary withdrawal. The Babinski sign (c) involves contraction of the extensor hallucis longus simultaneously with other muscles that shorten the leg: tibialis anterior, the hamstrings (arrow), and the tensor fasciae latae (from van Gijn, 199570, by kind permission of the author and Postgraduate Medical Journal).

(c)

form (the sign of Babinski) involves flexion at the hip, knee and ankle as well as extension of the great toe70 (Fig. 3.19). Failure to appreciate this perhaps explains, in part, the rather poor reliability of the sign. While the presence of a Babinski response signifies a lesion of the corticospinal tract, it is not invariable, particularly if there is no weakness of the foot. Although swallowing involves both the motor and sensory systems, mention will be made of it here. The

(b)

gag reflex alone is not an adequate examination of the ninth and tenth cranial nerves, nor is it a good indicator of swallowing ability71 (section 11.17). Sensation on the two sides of the soft palate should be tested separately with an orange-stick, elevation of the palate should be observed, and the patient should be asked to cough. Failure to oppose the vocal cords adequately will result in some air escaping, which should alert the physician that the patient may have swallowing difficulties.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 62

Chapter 3 Is it a vascular event and where is the lesion?

Clinical practice Most patients with stroke have motor symptoms or signs. Weakness usually affects one side of the body: the face, arm, or leg in isolation (monoparesis or monoplegia), each limb as a whole or in part, or a combination of these (hemiparesis or hemiplegia). Lesions in the internal capsule and ventral pons tend to result in a hemiparesis/ hemiplegia that is equally severe (proportional) in the arm and leg and is seldom accompanied by other neurological symptoms and signs – a pure motor stroke (section 4.3.2). When there is a brachial monoparesis (isolated weakness of the arm), or the weakness affects predominantly the face, hand or fingers even, it is more likely to be due to a cortical (where function is distributed anatomically) rather than a subcortical (where function is concentrated) lesion.72 When weakness involves the hand only, it is often referred to as a ‘cortical hand’. This can be mistaken for a peripheral nerve lesion, but closer analysis reveals that this would require simultaneous involvement of the median, ulnar and radial nerves, a most unlikely occurrence. It is generally considered to occur because of the large cortical representation of the hand. Isolated upper motor neurone facial weakness, however, seems to be of less localizing value, and it can certainly occur with very small infarcts in the genu of the internal capsule and in the pons. When weakness is confined to, or predominates in, the leg, the lesion is most likely to, but not invariably, involve the territory of the anterior cerebral artery or the sagittal venous sinus73 (section 4.2.2). Crossed weakness (i.e. weakness of one side of the face and the contralateral limbs) indicates a brainstem lesion or bilateral lesions (e.g. both hemispheres, or one hemisphere and contralateral brainstem). Paraplegia, triplegia and tetraplegia all occur more commonly from spinal than brain disorders (although see ‘locked-in syndrome’, section 3.3.2). When bilateral motor signs develop simultaneously – particularly if a cranial nerve palsy or crossed sensory disturbance (pointing to a brainstem lesion) are not present – and there is no sensory or reflex level to suggest a spinal cord lesion, cardiogenic embolism (i.e. causing two or more lesions), some kind of multifocal arteriopathy (such as vasculitis), abnormalities of the circle of Willis (section 4.2.2), or systemic hypotension (resulting in bilateral boundary-zone infarcts; section 4.2.4) must be considered. Very rarely hypotension results in paralysis predominantly of both arms (the ‘man in the barrel’ syndrome), with bilateral infarction in the boundary between the anterior and middle cerebral arteries. Although most strokes causing facial weakness result in a typical upper motor neurone pattern, there are some exceptions, which can lead to the erroneous diagnosis

of Bell’s palsy if there is minimal associated limb weakness. The most obvious exception is that of a brainstem stroke affecting the facial nerve nucleus. Occasionally, patients with very severe lower facial weakness from a supranuclear lesion also have some weakness of the upper face as well, particularly in the first few days after a stroke. This may reflect individual variation in the bilateral innervation of the upper facial (forehead) muscles. Lesions of the seventh cranial nerve nucleus involve both the upper and lower parts of the face; this pattern is not always due to Bell’s palsy. Sometimes patients seem at one moment to have (or are reported to have) a dense hemiplegia (usually leftsided) and yet, very soon afterwards, move the ‘paralysed’ limbs. This may be misinterpreted as a hysterical conversion disorder.74 However, this pattern can also be seen with the so-called capsular warning syndrome, or crescendo small vessel TIAs,75 although in such cases the episodes seem much more discrete (section 6.7.3). It can also occur in patients with a haemodynamically significant internal carotid artery stenosis, presumably due to subtle changes in distal perfusion pressure. However, the majority are patients in whom this seems to be a manifestation of inattention/neglect, or even apraxia. In patients who are recovering from what appears to be an extensive non-dominant hemisphere stroke, what seems to be a dense hemiplegia may improve very rapidly as the inattention/neglect begins to resolve – a fact that needs to be borne in mind when predicting the eventual functional outcome. It is always important to see whether a patient can sit up, get off the bed and walk, provided there is no risk to the patient or physician, whatever the motor deficit when tested on the bed. A severe deficit may be due to neglect and not weakness, and profound ataxia of gait may be associated with no motor deficit at all. Dysphagia is a common feature of acute stroke and an important cause of complications if not appropriately managed (section 11.17), but it is of limited neuroanatomical localizing value. Movement disorders such as hemiballismus, unilateral asterixis, hemichorea and focal dystonia occur in about 1% of patients with TIA and acute stroke due to contralateral, and rarely ipsilateral, small deep vascular lesions of the subthalamic nucleus, striatum or thalamus. Transient cerebral ischaemia may also masquerade as paroxysmal dyskinesia, and involuntary tonic limb spasms may arise contralateral to ventral pontine brainstem infarction. The abnormal movements usually regress spontaneously.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 63

3.3 The diagnosis of a cerebrovascular event

(a) Fig. 3.20 (a) T2-weighted MR scan showing an extensive infarct in an area usually supplied by the middle cerebral artery (arrows) in a patient who initially presented with several episodes of jerking of the left arm (which were initially thought

to be epileptic) and who awoke three days later with a left hemiparesis. (b) Catheter angiography showing occlusion of the ipsilateral internal carotid artery (arrow).

Occasionally, patients describe jerking movements of the limbs just before the onset of a stroke, or during a TIA. The distinction from focal motor epilepsy may be difficult (section 3.4.2).52,76 However, in contrast to epileptic seizures, these attacks may be provoked by postural change (from lying to sitting or standing up), hyperextension of the neck, walking, coughing, or starting or increasing antihypertensive therapy, and they may be alleviated promptly by sitting or lying down, all of which suggest they are due to ‘low flow’ rather than embolism (section 6.7.5). There is an association with severe internal carotid artery stenosis or occlusion (Fig. 3.20),76 and the attacks usually stop after carotid endarterectomy. This pattern has also been reported with internal boundary-zone infarcts (section 4.2.4). Other disorders that sometimes need to be considered in the differential diagnosis of stroke are the GuillainBarré syndrome, mononeuropathies, drop attacks, cataplexy and motor neurone disease (section 3.4.11).

(also known as cutaneous or exteroceptive) includes light touch, pain and temperature modalities. Deep and proprioceptive sensation refers to deep pressure and joint position sense, respectively. Synthesis and appreciation of these sensory inputs occurs at a cortical level. Discriminative sensation refers to stereognosis, two-point discrimination and graphaesthesia. Paraesthesiae are positive sensory phenomena (e.g. pins and needles) that are presumed to occur because of partial damage to the sensory tracts or posterior horn cells, which become hyperexcitable (perhaps akin to brisk reflexes), such that ectopic impulses are generated either spontaneously or after a normal stimulus-evoked volley of impulses.

3.3.5 Disturbance of the somatic sensory system There are broadly two types of sensory message passing from the periphery to the brain. Superficial sensation

(b)

Clinical anatomy The main sensory pathways are shown in Fig. 3.21. Impulses for superficial sensation are conveyed in the spinothalamic tracts, which synapse in the dorsal horn, cross the midline at about the same spinal level and then ascend through the lateral spinal cord and brainstem. Fibres carrying similar sensory impulses from the face enter the ipsilateral, descending (or spinal) trigeminal

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 64

Chapter 3 Is it a vascular event and where is the lesion?

Hand Leg

Face

Thalamus

Medial geniculate nucleus Red nucleus Substantia nigra Spinothalamic tract (light touch, pain, temperature) Vth nerve

postero-medial nuclei (face, tongue, fingers). Lesions of the thalamus often involve all sensory modalities, although deep sensation may be more affected than superficial sensation. From the thalamus there are probably two main projections. The first is to the postcentral or primary somatic cortex where there is somatotopic representation, with the leg uppermost and the face lowermost. The areas of sensory representation in the cortex match those of the motor homunculus (Fig. 3.16). A second projection is to the area adjacent to the upper part of the sylvian fissure and insula. Here there is less discrete localization, but in general, the face is rostral and the leg caudal. Interestingly, stimulation of this area may result in bilateral symptoms. Lesions confined to the parietal lobe generally affect higher-level ‘discriminatory’ functions (i.e. proprioception, two-point discrimination, astereognosis) rather than primary modalities of sensation (e.g. pain, temperature), but sometimes the primary modalities are affected, giving rise to a ‘pseudothalamic’ syndrome.

Pons

Clinical assessment

Spinal tract of V Dorsal root ganglion

Medial lemniscus (deep sensation, joint position sense) Olive Medulla

Spinal cord

Fig. 3.21 Diagrammatic representation of the main sensory pathways between the entry of the dorsal root to the spinal cord and the sensory cortex.

nucleus and cross the midline in the upper cervical spinal cord. They then ascend through the medulla medially, close to the medial lemniscus, and separate to join the medial part of the spinothalamic tract in the pons. The relevant fibres for deep sensation are primarily in the ipsilateral posterior columns of the spinal cord. Decussation (crossing) occurs in the caudal medulla, after which the fibres ascend through the brainstem in the medial lemniscus. Fibres carrying similar sensory impulses from the face enter the primary trigeminal nucleus in the pons and cross the midline at this level to form the trigeminal lemniscus, which lies adjacent to the medial lemniscus. All these ascending fibres converge towards the midbrain and project, in the main, to the posterior group of thalamic nuclei and in particular to the ventropostero-lateral nuclei (trunk and legs) and ventro-

Somatosensory symptoms due to stroke are usually described by the patient as numbness (‘like the numbness I have after going to the dentist’), tingling, or a dead sensation; occasionally as loss of temperature sensation when in the bath or shower; and very rarely as pain (at least at onset). Often, patients find it difficult to describe unusual sensations in a manner that allows accurate classification, and the descriptions seem to vary between cultures. Nevertheless, it is the distribution of the sensory disturbance, usually involving the face and arm, or hand and leg, or face and arm and leg, that stamp the disturbance as involving the central nervous rather than the peripheral nervous system. The difficulty lies with isolated sensory disturbances that are limited to a part of a limb or a part of the face, which can be central or peripheral in origin. It is widely recognized that formal testing of the sensory system is one of the most unreliable parts of the neurological examination (see below). Quite frequently there is no detectable sensory loss. In general, therefore, sensory symptoms even in the absence of a deficit on examination should be noted. Indeed, on the assumption that the patients are able to communicate and do not have neglect, the only situation in which they will have sensory loss without sensory symptoms is when there is a restricted problem of discriminatory rather than primary sensory function (due to a parietal lesion), something that is uncommon in clinical practice. Conversely, care needs to be taken over the interpretation of very transient sensory symptoms, which can

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 65

3.3 The diagnosis of a cerebrovascular event

be within the range of normal experience, although the clinician should not accept uncritically the patients’ common interpretation of sensory symptoms as being due to a ‘trapped nerve’ or ‘lying in a draught’. Formal testing of the sensory system is the most unreliable part of the neurological examination. However, the physician should always take due note of sensory symptoms, even when there is no deficit on examination. Superficial sensation should ideally be tested in the standard manner, using a wisp of cotton wool (light touch), the side of a tuning fork (cold temperature), and an appropriate pin (i.e. not a hat pin or hypodermic needle) or other sharp object that can be disposed of after it has been used (pain). Proprioception may be assessed with the patient’s arms outstretched, the fingers spread and the eyes closed. Look for drift of the patient’s arm in a ‘pseudoathetoid’ or ‘piano-playing’ manner. This can be amplified by asking the patient to touch the tip of his or her nose with the forefinger while the eyes remain closed; those with disturbed proprioception will repeatedly miss the target. This screening test will assess proprioception around proximal as well as distal joints, but drift can also be due to a motor deficit or neglect/ inattention. Therefore, if possible, one should always attempt the traditional method of testing joint position sense of the distal interphalangeal joints. If there is expressive aphasia, it is sometimes worth asking the patient to indicate with gesture the direction of movement. Romberg’s test can be useful to assess position sense in the legs if there are no other deficits affecting them (e.g. cerebellar ataxia). Clinical practice The lacunar syndrome of pure sensory stroke is typically caused by a lateral thalamic infarct or haemorrhage.77 The deficit may involve all modalities, or may spare pain and temperature sensation. If there is extension to the internal capsule, a sensorimotor stroke may occur (section 4.3.2). The Déjerine–Roussy syndrome is caused by more extensive lateral thalamic infarction, and consists of a mild contralateral hemiparesis, marked hemianaesthesia, hemiataxia, astereognosis and frequently paroxysmal pain/hyperaesthesia and choreoathetotic movements. The original cases had extension of the infarcts into the internal capsule and towards the putamen, although most of the features of this syndrome result from involvement of the ventroposterior nuclei of the thalamus57 (section 4.2.3).

Central pain alone may be caused by lesions at any level of the somatosensory pathways. However, in our experience, most patients with a diffuse pain in the body at the onset of the event often turn out to have a nonorganic/functional disorder, and patients with severe localized limb pain tend to have other disorders, such as nerve root compression in the neck or low back, or even myocardial infarction if the pain is down the arm or in the hand. Nevertheless, we have seen the occasional patient with thalamic infarction or haemorrhage develop unpleasant sensation in the contralateral limb within 2–3 days of onset. Some patients have restricted sensory syndromes that affect unusual combinations of body parts. The most frequent is the cheiro-oral syndrome, where there is a sensory abnormality over the perioral area (sometimes bilaterally) and ipsilateral palm. In some patients, the foot may also be involved (the cheiro-oral-pedal syndrome) and, in both the hand and foot, certain digits may be affected while others are spared (a pseudoradicular distribution). Although lesions at most levels of the sensory pathways can give these patterns, lesions in the ventro-posterior nuclei of the thalamus are the most likely. It is thought that sensory projections from the face (with a particularly large representation for the lips), hand and foot are somatotopically arranged in the ventral portion of this nucleus, and the fingertips have particularly large representation areas, with that for the thumb more medial and the little finger more laterally. The projection areas for the trunk and proximal limbs are relatively small and sited more dorsally. However, a similar proximity of projection areas through the corona radiata and in the sensory cortex may also occur, and consequently lesions in these areas may also result in the cheiro-oral syndrome. An isolated deficit in a pseudoradicular distribution (most often involving the thumb and forefinger) is probably more often caused by a cortical lesion, because a stroke of any given size would there affect fibres from a more restricted anatomical area than in the thalamus. Bilateral symptoms may occur from midpontine lesions, because the sensory fibres from the mouth, arm and leg are once again arranged somatotopically in the medial lemnisci. A defect of pain and temperature sensation below a certain level on the trunk on one side is usually a sign of spinal cord disease, but this may occur from ischaemic lesions in the lateral medulla on the contralateral side. This is thought to be due to the orientation of fibres from different body parts within the lateral spinothalamic tract. Although patients usually have contralateral facial sensory loss as well, during the recovery phase this can disappear, leaving only the pseudospinal sensory loss. It is also worth noting that infarction of the cervical

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 66

Chapter 3 Is it a vascular event and where is the lesion?

spinal cord causing a partial Brown–Séquard syndrome can be due to bilateral vertebral artery dissection.78

3.3.6–Disturbance of the visual system When trying to assess the visual disturbances that occur in patients with cerebrovascular disease, one needs to consider: the reception of visual stimuli by the eyes; the transmission of the visual information from the eyes to the occipital cortex and visual association areas; and the interpretation of the visual information in the occipital cortex. Additionally, we have included in this section information about pupillary reactions and eye movements. Vision clinical anatomy Lesions at different sites in the visual pathway give highly characteristic abnormalities (Fig. 3.22). Amaurosis fugax (meaning literally ‘fleeting blindness’) and transient monocular blindness (TMB) are terms used interchangeably to describe temporary loss of vision in one eye. TMB may be caused by transient ischaemia in the distribution of the ophthalmic, posterior ciliary, or central retinal artery. Vascular lesions affecting the optic chiasm symmetrically (when one might detect a bitemporal hemianopia) are rare. Indeed, the only vascular

Retina

lesion of note would be a large aneurysm of the circle of Willis. hom*onymous visual field deficits (i.e. loss of vision in the corresponding part of the visual fields in both eyes) signify a retrochiasmal lesion (Fig. 3.23). Lesions of the lateral geniculate nucleus may result in a hom*onymous horizontal sectoranopia (i.e. a segmental defect that respects the vertical but not the horizontal meridian) due to topographical organization within the lateral geniculate nucleus, but these rarely occur in isolation. The optic radiations pass from the lateral geniculate nucleus as the most posterior structures of the internal capsule. Involvement at this level is probably one of the causes of hemianopia in extensive middle cerebral artery territory infarction. The radiations do not seem to be affected by occlusion of a single perforating artery. Restricted lesions of the inferior optic radiation between the lateral geniculate nucleus and the calcarine cortex, where the fibres swing over the temporal horn of the lateral ventricle and deep into the temporal lobe (Meyer’s loop), result in a hom*onymous superior quadrantanopia, while lesions of the superior optic radiation in the parietal lobe result in a hom*onymous inferior quadrantanopia. On purely anatomical grounds, one might anticipate encountering an inferior quadrantanopia fairly frequently, because the middle cerebral artery is traditionally considered to supply the area through which the superior but not the inferior optic Visual field deficit Optic nerve

Anterior cerebral artery Middle cerebral artery Internal carotid artery Posterior cerebral artery

Ophthalmic artery Optic chiasm

Optic tract

Optic nerve

Lateral geniculate nucleus

Hypothalamus Optic tract Tectal plate

Superior optic radiation

Lateral geniculate nucleus

Optic radiation

Inferior optic radiation

Combined optic radiation

Visual field projection

Right

Calcarine cortex Visual cortex

Left

Fig. 3.22 Diagram of the visual pathways and their vascular supply, and the visual field defects that may result from vascular lesions at various sites along the visual pathway. The visual fields represented by the arrows on the retina correspond with the arrowheads and tails superimposed on the visual cortex. The dark purple represents visual pathways carrying vision from the left hom*onymous half fields and the light purple those from the right hom*onymous half fields. Arteries are shown in solid black.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 67

3.3 The diagnosis of a cerebrovascular event

hom*onymous hemianopia is a loss of corresponding field of vision in both eyes

When it affects the left visual field it looks like this:

f corresponding field of vision in both eyes

When it affects the right visual field it looks like this:

hom*onymous hemianopia is a loss o

Fig. 3.23 hom*onymous hemianopia.

radiation passes. However, in some patients, the area of supply of the middle cerebral artery extends much more posteriorly than is apparent on standard ‘maps’ (section 4.2.2), and an middle cerebral artery lesion thus produces a hom*onymous hemianopia by interrupting the optic radiations as they converge from the temporal and parietal lobe on their way to the occipital lobe.79 Also, perforating arteries from the carotid system supply the optic tracts and lateral geniculate nucleus (section 4.2.2). In other patients, the posterior cerebral artery may derive its blood supply via the posterior communicating artery, and therefore, like the middle cerebral artery, be affected by embolism from the internal carotid artery. Perhaps the most likely explanation, however, is that in many patients there is a mixture of a true visual field loss and visual inattention. If the entire calcarine cortex or optic tract on one side is damaged, there will be a complete hom*onymous hemianiopia, including macular vision. When this is an isolated feature, it is most likely due to a lesion in the occipital lobe, and when this is an infarct it will most often be due to occlusion of the posterior cerebral artery. Embolism is probably the most frequent cause, but giant-cell arteritis and migraine need to be considered. Sparing of macular vision with an otherwise complete hemianopia does occur in posterior cerebral artery territory infarction. The conventional explanation is that the infarction is confined to the lateral cortical

surface and that the macular cortex is spared because it receives sufficient collateral supply from the terminal branches of the middle cerebral artery (section 4.2.2). In theory, if the deficit in the two eyes is incongruous, it is most likely to be due to a lesion of the optic tracts, whereas if the deficits are entirely congruous, the lesion is likely to be in the calcarine cortex. However, it can be difficult to make this distinction in routine clinical practice. Visual agnosia is when primary visual perception is intact but the patient is unable to identify an object without resorting to the use of other sensory modalities such as touch. Prosopagnosia refers to the inability of patients to recognize familiar faces, even though they can describe them. In its pure form it is very rare, but it can result in enormous distress if patients deny recognizing their close family. Most lesions causing visual agnosia are in the anterior part of the dominant occipital lobe (the so-called visual association areas) and the angular gyrus, although most cases of prosopagnosia have occurred with bilateral lesions. clinical assessment Many patients have great difficulty in describing visual symptoms, particularly when they have resolved. It is therefore very important to be clear what a patient means. For example, the term ‘blackout’ may be used to mean bilateral blindness and also loss of consciousness,

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 68

Chapter 3 Is it a vascular event and where is the lesion?

and some patients who report ‘blurred vision’ in both eyes are actually trying to describe double vision. We find it useful to establish the functional severity of the visual disturbance – e.g. were they unable to find their way around, or to recognize faces (neither of which would occur with either TMB or an hom*onymous hemianopia). In vascular TMB the symptoms usually occur without provocation, but they may occasionally be precipitated by bright or white light, a change in posture, exercise, a hot bath or a heavy meal, particularly in patients with severe ipsilateral carotid disease. Generally, the visual loss during vascular TMB is painless (although some patients do complain of a dull ache or numbness in or above the eye) and very rapid. It can be described as if a blind or shutter had come down from above or, less often, up from below. The visual loss may be restricted to either the upper or lower half of the visual field and, less frequently, to the peripheral nasal and/or temporal field (in which case, be suspicious that the visual loss is or was binocular; i.e. a hom*onymous hemianopia). A pattern of diffuse, constricting or patchy loss may also occur. TMB may recur, usually in a stereotyped fashion, but the area of visual impairment may vary from one episode to the next, depending on which part of the retina is ischaemic. A frequent and, from a vascular anatomical point of view, very important clinical problem is trying to differentiate a transient hom*onymous hemianopia from TMB. The single most important question is to ask whether the patient covered each eye in turn during the episode. For patients without any residual visual disturbance, it can be useful to cover one eye and ask whether that reproduces the previously experienced effect. After covering the ‘good eye’, patients with TMB (of the ‘bad eye’) will have seen nothing, whereas patients with a hom*onymous hemianopia will have still seen something in the remaining half of the visual field. On the other hand, when patients cover the affected eye during TMB, they tend not to notice any visual disturbance because of normal vision from the other eye. In our experience, asking patients whether they saw half of everything often seems to cause confusion, and it may be better to ask them to look at your face and describe what they might have seen if they had been having an attack

When a patient complains of transient loss of vision in one eye, do not assume that the visual loss was monocular; it may have been a hom*onymous hemianopia. Ask the patient if they covered each eye in turn during the episode of loss of vision, and if so if the visual disturbance was present in both eyes or one (and if the latter, which eye).

Patients may not recognize an isolated hom*onymous hemianopia, or they may simply describe it as ‘blurred vision’ or ‘a shadow’. Even if they covered each eye in turn during the symptoms, it still may not be possible to be really confident, because hom*onymous hemianopia does not necessarily split macular vision and may be interpreted by the patient as loss of vision in one eye only. The presence of other symptoms may be helpful – for example, if there are ipsilateral visual and brain symptoms the visual problem is likely to be a hemianopia, whereas if they are contralateral it points to TMB. It may be difficult to test visual acuity formally (e.g. with a Snellen chart) in patients with acute stroke, because of drowsiness, aphasia, or the fact that they are bed-bound, but at least testing with a hand-held acuity chart or simply using everyday written material should be attempted. Many stroke patients are elderly, and concomitant eye diseases such as glaucoma, senile macular degeneration, cataracts and diabetic retinopathy are therefore common. It is important to identify these conditions at an early stage, because they may well make rehabilitation significantly more difficult. Indeed, the improved visual acuity that follows a cataract extraction may make the difference between being able to live independently (and safely) or not for a patient who has residual disability from a stroke. Because of the very large representation of the macula area in the occipital cortex, even patients who have a complete hom*onymous hemianopia do not have significantly reduced visual acuity per se. Elderly patients rarely attempt, let alone accomplish, tests of visual acuity without their own clean spectacles. If the visual loss is persistent (i.e. beyond several hours) and ophthalmoscopy reveals pallor of all or a section of the retina (due to cloudy swelling of the retinal ganglion cells), then the diagnosis is retinal infarction (Fig. 3.24). Additional findings may include an afferent pupillary defect, embolic material in the retinal arteries or arterioles, and a cherry-red spot over the fovea (due to accentuation of the normal fovea, which is devoid of ganglion cells, against the abnormally pale retina) in cases of central retinal artery occlusion (Fig. 3.25). If the eye is red and painful with a fixed, semidilated, oval pupil and cloudy/steamy cornea, then acute glaucoma is the likely diagnosis (section 3.5.1) (Fig. 3.26). Episcleral vascular congestion, a cloudy cornea, neovascularization of the iris (rubeosis iridis) and a sluggishly reactive mid-dilated pupil indicate chronic anterior segment ocular ischaemia which may be due to carotid occlusive disease, or small vessel disease particularly in a diabetic. This is so-called ischaemic oculopathy

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 69

3.3 The diagnosis of a cerebrovascular event

Fig. 3.24 An ocular fundus photograph of a patient with inferior temporal branch retinal artery occlusion, showing pallor of the inferior half of the retina due to cloudy swelling of the retinal ganglion cells caused by retinal infarction. The inferior temporal branch arteriole is attenuated and contains embolic material (arrow). Also reproduced in colour in the plate section.

(Fig. 3.27).2 Because ischaemic oculopathy is a gradual, chronic process it is commonly also called chronic ischaemic oculopathy. If there is a visual field defect of acute onset, such as an absolute or relative inferior altitudinal hemianopia, inferior nasal segmental loss, or central scotoma, and if ophthalmoscopy reveals swelling of a segment or all of the optic disc (which may be indistinguishable from that seen with raised intracranial pressure), flame-shaped haemorrhages near the disc and distended veins

Fig. 3.26 A photograph of the eye of a patient with acute glaucoma, showing congested sclera, cloudy cornea and oval pupil. Also reproduced in colour in the plate section.

Fig. 3.25 An ocular fundus photograph of a patient with central retinal artery occlusion, showing a cherry-red spot over the fovea (arrow). The cherry red spot is the normal fovea (devoid of ganglion cells), which seems more obvious because the surrounding infarcted retina has lost its red colour and appears pale. Also reproduced in colour in the plate section.

(Fig. 3.28), then the diagnosis is likely to be anterior ischaemic optic neuropathy (section 3.5.2). Later the optic disc becomes pale. Ophthalmoscopy may reveal retinal emboli but these are not necessarily symptomatic (up to 1–2% of the population over the age of 50 may have asymptomatic retinal emboli). The most common type are the bright orange or yellow crystals of cholesterol that originate from ulcerated atheroma in proximal arteries. Although cholesterol crystals are actually white, they appear orange

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 70

Chapter 3 Is it a vascular event and where is the lesion?

(a)

Fig. 3.28 An ocular fundus photograph of a patient with anterior ischaemic optic neuropathy due to occlusion of the posterior ciliary artery as a result of giant cell arteritis. Note the oedema of the optic disc and flame-shaped haemorrhages (arrow). Also reproduced in colour in the plate section.

(b) Fig. 3.27 (a) and (b) Ischaemic oculopathy of the right eye; note episcleral vascular congestion, cloudy cornea, neovascularization of the iris (rubeosis of the iris) and middilated pupil on external examination of the eye, which indicate chronic anterior segment ischaemia due to carotid occlusive disease (from Hankey & Warlow, 19942 by kind permission of the authors and W.B. Saunders Co. Ltd). Also reproduced in colour in the plate section.

or golden because their thin, fish-scale contour permits blood to pass above and below them and thus produce their characteristic refractile appearance (Fig. 3.29). Most of the crystals, because of their small size, thin, flat structure and lack of adhesiveness, pass through the retinal arterioles rapidly and rarely occlude the larger vessels, although it is probable that large clumps of crystals briefly occlude the central artery of the retina, producing TMB, before breaking up and being flushed away. White plugs of fibrin, platelets, or fatty material are less common. They occur in all sizes and are more likely to be symptomatic. Calcium emboli are chalky white angular crystals that tend to arise from calcific aortic stenosis, and they may permanently occlude the central retinal artery (behind the cribriform plate), or one of

Fig. 3.29 An ocular fundus photograph showing golden orange cholesterol crystals (Hollenshorst plaques) in the cilioretinal artery (arrows). The cilioretinal artery is present in only about one-third of the population. It originates from a branch of the short posterior ciliary artery and supplies the macula. Also reproduced in colour in the plate section.

the branch retinal arterioles near the optic disc. Other less common types of emboli include microorganisms (septic), fat and tumour cells. Roth spots, which are very small white infarcts encircled by haemorrhage, were

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 71

3.3 The diagnosis of a cerebrovascular event

Fig. 3.30 An ocular fundus photograph showing narrowing and tortuosity of retinal arterioles, arteriovenous nipping, retinal haemorrhages and papilloedema. These are the features of hypertensive retinopathy seen in malignant hypertension. Also reproduced in colour in the plate section.

thought to be caused by septic emboli, but it now seems more likely that they are due to rupture of retinal capillaries and the extrusion of blood cells. Whilst the interobserver and intra-observer agreement for the detection of retinal emboli is quite high (kappa = 0.73 and 0.63, respectively), agreement on a range of qualitative assessments of emboli type is much poorer. Narrowing, focal irregularity/constriction and tortuosity of retinal arterioles, arteriovenous nipping and fluffy white patches of transudate (‘cotton-wool patches’ which are thought to represent small focal infarcts in the inner layers of the retina), indicate long-standing hypertension. If papilloedema and retinal haemorrhages are also present, this indicates malignant hypertension, but is now uncommon (Fig. 3.30). Retinal haemorrhages without the other changes of hypertensive or diabetic retinopathy in a patient with a non-traumatic, acute neurological event is strong evidence of a haemorrhagic stroke. They are usually caused by a very sudden increase in intracranial pressure transmitted to the distal optic nerve sheath, where it causes a temporary obstruction of retinal venous outflow. The subsequent rise in retinal venous pressure leads to secondary bleeding from retinal veins and capillaries. The appearance of the haemorrhage depends on its site. Small dot and blot haemorrhages lie in the deep retinal layers; linear haemorrhages in the superficial (nerve fibre) layer; ‘thumbprint’ haemorrhages with frayed borders are preretinal or superficial retinal; and large subhyaloid haemorrhages (large round haemorrhages

with a fluid level) lie between the retina and the internal limiting membrane. Subhyaloid haemorrhages and other types of retinal haemorrhage can be seen in about 20% of patients with subarachnoid haemorrhage (section 3.7). Assessment of the visual fields must be tailored to the patient’s overall condition. It is important that the physician has a repertoire of methods and does not give up simply because ‘formal’ testing by confrontation is impossible in a patient who is drowsy, aphasic, cognitively impaired, or just cannot sit up. Kinetic testing (i.e. using moving objects, waggling fingers, etc.) is a less sensitive way of detecting deficits than static methods such as counting fingers or comparing colours in each hemifield. If the patients can understand and communicate, one should first ask them to describe what they see in front of them, perhaps using the same text as for testing visual acuity, and ideally testing each eye individually. After that, hold up fingers sequentially in each quadrant of vision in each eye and ask the patient to count them or, for greater sensitivity, use a pin with a red ball on its end. This will detect a hemianopia or quadrantanopia. Following this, perform bilateral, simultaneous finger movement to detect evidence of visual inattention. Testing of the visual fields with automated perimetry is extremely tiring and of little value in the acute situation. Indeed, it may produce bizarre deficits that are normally associated with non-organic disorders. However, it may be necessary later when there is doubt about eligibility for driving. Although testing the visual fields using conventional confrontation methods is often impossible because the patient is drowsy, dysphasic, cognitively impaired, or cannot sit up, one can usually use other methods to determine whether or not there is an abnormality. If, for whatever reason, the patient cannot follow commands, one will need to use quite gross stimuli to be sure of eliciting and identifying a response if the visual fields are intact. Examples include observing if there is any response to moving a brightly coloured object in one hemifield, getting a colleague to approach the patient from one side, or seeing whether the patient blinks when a threatening stimulus (e.g. a quickly moving finger) is brought towards the eye (one needs to be careful that the associated air current is not simply stimulating a corneal reflex). A hemianopia is almost always associated with ipsilateral loss of this blink reflex, although the reverse is not always true. If the patient is not aphasic and there are members of the team around the bed, ask the patient to point to each one of the team in turn. An asymmetrical response to any of these tests suggests a field defect, inattention, neglect or a combination of these.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 72

Chapter 3 Is it a vascular event and where is the lesion?

It is not surprising, given these difficulties, that the inter-observer reliability of the assessment of visual fields is relatively poor (Table 3.31). clinical practice Some patients with severe stenosis or occlusion of one or both internal carotid arteries may experience a visual disturbance in one or both eyes on exposure to bright sunlight or white light. Blurring, dimming, or constriction of the visual field from the periphery to the centre of vision of the involved eye develops over minutes rather than seconds. Objects appear bleached like a photographic negative, or there may be a scotoma or complete visual loss; an altitudinal defect is most unusual. Such transient episodes of monocular or binocular blindness are presumed to be due to low flow in the choroidal circulation. They are typically less rapid in onset than the brief transient attacks of embolic origin, and sight returns more gradually. Sunglasses may be an effective symptomatic treatment. Cortical blindness is a syndrome in which the patient has no vision despite normally functioning eyes and anterior visual pathways. Cases of sudden, spontaneous and simultaneous dimming or loss of vision in all of the visual field of both eyes are presumed to be due to bilateral occipital lobe (visual cortices and optic radiations) ischaemia/infarction due to occlusion of the top of the basilar artery or both posterior cerebral arteries. If the visual symptoms occur in isolation (without associated symptoms of focal cerebral ischaemia, seizures, or reduction in consciousness) in an elderly patient and if they resolve within 24·h, they are probably due to a TIA of the occipital lobes.48 However, when the same symptom occurs in adolescents and young adults, investigations are unlikely to reveal a cause, and the long-term prognosis appears benign. Occasionally, when the deficit persists, true cortical blindness has to be distinguished from non-organic visual loss. This is best detected with an optokinetic drum or a long piece of material with vertical stripes (e.g. a scarf or tape measure) because it is impossible to suppress nystagmus voluntarily if there is visual function. Sometimes, genuine bilateral blindness is denied by the patient (Anton syndrome) which signifies involvement of the association areas adjacent to the primary visual cortex, but otherwise the pathogenesis is as unclear as that of the denial of left hemiplegia or anosognosia in general. Perhaps the right hemisphere component is crucial. Visual hallucinations can occur in patients with stroke involving the occipital, temporal and parietal cortices as well as the eye, optic pathways and cerebral peduncle (section 11.27.3). Those secondary to occipital lesions most commonly consist of elementary

(unformed) visual perceptions, sensations of light and colours, simple geometric figures, and movements. Posterior temporal lesions, involving the association cortex, result in more complex (formed) visual hallucinations, consisting of faces and scenes that may include objects, pictures and people. Lesions in the high midbrain, particularly the pars reticulata of the substantia nigra, may give rise to the so-called ‘peduncular hallucinosis’ of Lhermitte, in which the hallucinations are purely visual, appear natural in form and colour, move about like an animated cartoon, and are usually considered to be unreal, abnormal phenomena (i.e. insight is preserved). More commonly, however, visual hallucinations are due to non-vascular disorders such as migraine or partial seizures (in which case, the hallucinations are usually unformed), psychosis, or an adverse effect of a drug such as levodopa. Micropsia, which is the illusion of objects appearing smaller than normal, and palinopsia, which is the persistence or recurrence of visual images after the stimulus has been removed, can occur with parietal lobe lesions. Flashing lights, shooting stars, scintillations, or other positive phenomena in the area of impaired vision occasionally arise during retinal or optic nerve ischaemia, but are far more commonly features of migraine or glaucoma. Pupils clinical anatomy The size of the pupil is determined by the balance of tonic impulses from the pupillodilator fibres, which receive input from the sympathetic nervous system, and the pupilloconstrictor fibres, which receive input from the parasympathetic nervous system. The sympathetic fibres descend ipsilaterally from the hypothalamus, through the lateral brainstem adjacent to the spinothalamic tract. They occupy a more central position in the lateral grey column of the cervical spinal cord and exit via the first thoracic root. The fibres then pass across the apex of the lung to enter the sympathetic chain, which ascends through the neck in association with the carotid artery. The fibres associated with sweating separate in the superior cervical ganglion and then travel in association with branches of the external carotid artery. The other fibres enter the cranial cavity on the surface of the internal carotid artery. The fibres innervate the pupil via the long ciliary nerves, whilst those supplying the tarsal muscles are carried in the third cranial nerve. Following reception of light by the retina, impulses are conveyed in the optic nerve. After the optic chiasm, they are conducted in both optic tracts to both Edinger– Westphal nuclei (a distinct part of the third nerve nuclear complex). The parasympathetic nerves exit

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 73

3.3 The diagnosis of a cerebrovascular event

Lateral geniculate nucleus From hypothalamus, via Optic tract posterior commissure Optic nerve Midbrain

Retina Fig. 3.31 The neurogenic control of pupil size: an outline of the parasympathetic and sympathetic pathways involved in pupil constriction (–) and pupil dilatation (–). The third cranial nerve nuclear complex consists of: the Edinger-Westphal nuclei concerned with parasympathetic innervation of the pupils; the midline nucleus of Perlia, concerned with convergence and accommodation; and the lateral nuclei, which innervate the levator palpebrae, superior recti, inferior oblique, medial recti and inferior recti muscles. It is possible for vascular lesions to cause ischaemia of the lateral nuclei (resulting in extraocular palsy) but spare the pupilloconstrictor fibres from the Edinger-Westphal nuclei.

Pretectal nucleus Pupil

Optic chiasm

Internal carotid artery Parasympathetic pupilloconstrictor reflex Superior Sympathetic pupillocervical dilator reflex ganglion

Ciliary ganglion

alongside the third nerve and travel with it to the orbit (Fig. 3.31). There, they synapse in the ciliary ganglion, which gives rise to the short ciliary nerves that innervate the sphincter pupillae and the ciliary muscle. Lesions anterior to the lateral geniculate body result in loss of the pupillary light reflex. clinical assessment It is very uncommon for a patient to be aware of their own pupillary abnormalities, but others may notice. Just occasionally, patients notice that their pupils are unequal – they will usually think one is dilated, rather than the more common abnormality of one being constricted. With a truly dilated pupil, the patient may be distressed by abnormal brightness and difficulty focusing. The response of the pupils to light – both direct and consensual – should be tested, as well as accommodation, if possible. Interruption of the descending sympathetic pathway in the brainstem and at other sites before the carotid bifurcation results in a complete ipsilateral Horner syndrome, i.e. miosis, ptosis and loss of sweating on the side of the face. Lesions of the internal carotid artery (e.g. carotid dissection; section 7.2.1) generally spare facial sweating. Sometimes a transient Horner syndrome is the only clue to a carotid dissection. clinical practice There are many causes of anisocoria (unequal pupils) in the elderly, most are not vascular. Perhaps the commonest is the use of drops to treat glaucoma, but any local inflammatory condition (e.g. iritis) can be responsible. Furthermore, physiological anisocoria may occur in

Edinger-Westphal nucleus

IIIrd nerve

Spinal cord (C8-T1) Ascending sympathetic chain

up to 20% of the normal population. Therefore, anisocoria very much needs to be assessed in the context of any other signs (such as ptosis). Elderly patients with stroke may be using pupilloconstrictor drops for glaucoma, and this may cause unequal pupils. Because of the functional separation of fibres within the third nerve nuclear complex, it is possible for third nerve palsies from midbrain vascular lesions to spare the pupillary reaction, which remains normal to light. On the other hand, in an unconscious patient with extensive damage to the midbrain (either due to intrinsic disease or secondary to pressure from above), the pupils will both be fixed and either dilated or in mid-position (4–5·mm), depending on whether the sympathetic as well as the parasympathetic fibres are involved. Bilateral ‘pinpoint’ pupils in an unconscious patient suggests an extensive lesion in the pons if there is no evidence of drug overdose (e.g. opiates). This is thought to be due to a combination of damage to the sympathetic fibres and irritation of the parasympathetic fibres (lesions solely of sympathetic fibres do not usually result in such intense pupilloconstriction). Despite this, the pupils react to a bright light, although this may be difficult to observe. External ocular movements and eyelids clinical anatomy The external ocular muscles maintain fusion of the images from each retina. The oculomotor (third nerve) complex

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 74

Chapter 3 Is it a vascular event and where is the lesion?

in the midbrain innervates the medial, superior and inferior recti and inferior oblique muscles. The trochlear nerve (fourth nerve), also originating in the midbrain, innervates the superior oblique muscle, and the abducens nerve (sixth nerve) in the pons innervates the lateral rectus muscle. Other structures of importance include the medial longitudinal fasciculus (MLF) which effectively links the nuclei; the paramedian pontine reticular formation, sometimes known as the pontine lateral gaze centre; and the rostral interstitial nucleus of the MLF in the midbrain which generates the vertical and torsional components of eye movement (Fig. 3.32). The cerebellum and vestibular nuclei are also important for the control of eye movements. The supranuclear control of conjugate eye movement is of relevance to patients with stroke. Voluntary eye movements are initiated in the frontal eye field, which is anterior to the precentral gyrus, whilst the reflex visual pursuit movements involve the occipital cortex and visual association areas. The fibres from these areas do not directly innervate the oculomotor nuclei, but rather their input is integrated by the MLF. The levator palpebrae superioris is innervated by the superior division of the oculomotor nuclei in the midline, as are the superior rectus muscles. Because of this, movement of the eyelids is closely linked to vertical eye movements. The superior and inferior tarsal muscles receive sympathetic innervation via the third cranial nerve. They assist eye opening, and when they are paralysed (as in Horner syndrome), the palpebral fissure is narrowed (but there is not complete ptosis, as is seen with a third nerve palsy, or central nuclear third nerve palsy). Nystagmus is an involuntary, biphasic ocular oscillation that occurs with lesions of the vestibular pathway and cerebellum. In patients with vascular disease, there may be either ischaemia of the labyrinth or of the vestibular nuclei in the brainstem, and the pattern of any associated signs may be of more use in localizing the lesions than attempts to analyse the nystagmus itself – i.e. with nuclear lesions, there are likely to be other signs of brainstem disturbance. clinical assessment Hemispheric stroke involving the frontal cortex results in a failure of voluntary conjugate gaze to the contralateral side. Because of the unopposed action of the intact frontal eye field of the hemisphere not affected by the stroke, there is conjugate deviation of gaze towards the side of the stroke lesion. This normally settles over a period of 1–2 weeks. The patients are usually drowsy so that reflex pursuit movements cannot be assessed, although if the lesion was confined to the frontal lobe one might expect such movements to be retained. If

Frontal lobe Occipital lobe

III

Tectal field for vertical eye movements

III

Midbrain

Superior colliculus

IV Medial longitudinal fasiculus

Inferior colliculus

x y Pons

y VI

Pontine field for horizontal eye movement

Midline Fig. 3.32 The role of the medial longitudinal fasciculus (MLF) in the control of conjugate gaze. Conjugate gaze requires coordinated action of the third, fourth and sixth cranial nerve nuclei and nerves in the brainstem. The MLF links these nuclei and also provides a pathway for inputs from the frontal eye field (for voluntary eye movements – saccades) and the occipital lobe (for reflex movements), as well as the tectal field for vertical eye movements and the pontine field for horizontal eye movements. ‘x’ indicates the site of the lesion that results in the classical bilateral internuclear ophthalmoplegia often associated with multiple sclerosis. Vascular lesions (‘y’) more often result in unilateral internuclear ophthalmoplegia, presumably because they are more likely to respect the vascular territory of a paramedian pontine perforating artery, and thus the midline, than a plaque of demyelination. Conversely, a lesion of the lateral pons may prevent ipsilateral lateral gaze, and there may be a conjugate deviation of the eyes away from the side of the lesion.

patients have an acute pontine lesion, conjugate deviation away from the side of the lesion may occur, but this is less likely to recover. When a patient complains of double vision, the following questions help to identify the site of the problem: • is the double vision present when one eye is closed (monocular diplopia), or only when both eyes are open (binocular diplopia)?

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 75

3.3 The diagnosis of a cerebrovascular event

• are the images separated side by side (horizontal), one above the other (vertical), or at an angle to one another (oblique)? • and in which direction of gaze are the images separated maximally? If the patient is conscious, communicative and cooperative, the eye movements can be tested in the normal way. However, it is difficult to do this if patients cannot follow commands. In such cases, spontaneous movements in each direction should be observed to confirm the absence of a gaze palsy. One can also stimulate patients to look in each direction by doing something ‘interesting’ in different fields of vision – so that they follow the examiner’s face, for example, rather than a finger or a pen. This also has the advantage that the examiner can continually reinforce the command ‘watch my nose’. The patient may be stimulated to look in each direction by doing something ‘interesting’ in different fields of vision – so that he or she follows the examiner’s face, for example, rather than a finger or a pen. Nystagmus of brainstem or cerebellar origin is probably best appreciated by asking the patient to fixate on, and then follow, a moving target. A few irregular ‘jerks’ of the eyes are often seen in normal people when they move their eyes, particularly at the extremes of lateral gaze. Acquired pendular nystagmus (i.e. where there is a sinusoidal waveform) is associated with lesions of the tegmentum of the pons and medulla. Upbeat nystagmus is associated with pontine or cerebellar lesions, and downbeat nystagmus with lesions in the medulla or at the craniocervical junction. A torsional or rotatory component can occur from both central and peripheral lesions. Convergence retraction nystagmus (rhythmic oscillation in which a slow abduction of the eyes in respect to each other is followed by a quick movement of adduction, and is usually accompanied by a quick rhythmic retraction of the eyes into the orbits) is considered indicative of midbrain disease; this is actually a disorder of horizontal eye movement rather than true nystagmus. A number of related disorders, which are probably due to a disturbance of saccadic (voluntary) eye movements, are associated with cerebellar disease. These include ocular dysmetria (where there is overshoot of the eyes on attempted fixation), ocular flutter (where there are occasional bursts of rapid horizontal oscillations without an intersaccadic interval), and so-called square wave jerks. clinical practice Transient diplopia in isolation may be an indication of a brainstem ischaemic event but can also be due to

myasthenia gravis, for example. However, transient diplopia in association with other symptoms of brainstem or cerebellar dysfunction – such as unilateral or bilateral motor or sensory disturbances, vertigo, ataxia or dysarthria – usually signifies a TIA in the vertebrobasilar circulation. Monocular diplopia is usually due to intraocular disease causing light rays to be dispersed onto the retina (e.g. corneal disease, cataract, vitreous haemorrhage) or to functional (non-neurological) disturbance. It has been reported rarely after occipital lesions but it is not due to paralysis of extraocular muscles. Even though it is sometimes claimed to be a pathognomonic sign of multiple sclerosis, vascular disease can cause an internuclear ophthalmoplegia (failure of adduction in the adducting eye, with nystagmus in the abducting eye), due to involvement of the MLF on the side of the adducting eye (Fig. 3.32). A failure of conjugate horizontal gaze to one side can occur with ischaemia of the ipsilateral (the side to which the patient cannot look) paramedian pontine reticular formation. Additional involvement of the ipsilateral MLF (with failure of adduction of the ipsilateral eye on attempted gaze to the other side) may result in the so-called ‘one-and-a-half syndrome’, where the only remaining horizontal eye movement is abduction of the contralateral eye. An oculomotor (third cranial) nerve palsy at the onset of subarachnoid haemorrhage (SAH) frequently indicates a ruptured aneurysm at the origin of the posterior communicating artery from the internal carotid artery, less frequently an aneurysm of the carotid bifurcation, the posterior cerebral artery, the basilar bifurcation or the superior cerebellar artery. Third nerve palsy may also occur with unruptured aneurysms (presumably by expansion) or several days after SAH as a result of swelling of the ipsilateral cerebral hemisphere because of delayed cerebral ischaemia. Most often the pupil is dilated and unreactive, but in some patients it is spared. Abducens (sixth cranial) nerve palsies, frequently bilateral in the acute stage, may develop after SAH as a false localizing sign of raised intracranial pressure, because of traction on the nerves against the petrous temporal bone, or by downward transtentorial herniation of the diencephalon. Occasionally, posterior circulation aneurysms may cause a sixth nerve palsy as a result of direct compression.

If a patient presents following a sudden, severe headache and is found to have a third cranial nerve palsy, rupture of a posterior communicating artery aneurysm is highly likely.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 76

Chapter 3 Is it a vascular event and where is the lesion?

Occasionally, one encounters patients who are unable to open their eyes. If bilateral ptosis is associated with a vertical gaze palsy and there is no suggestion of myasthenia gravis, then it is probably due to a nuclear third nerve palsy (dorsal part, in the midline) or a massive right hemisphere lesion (‘cerebral ptosis’).80 ‘Cerebral ptosis’ must be distinguished from blepharospasm, and from apraxia of eyelid opening in which patients are unable to open their eyes on command but can do so spontaneously (i.e. the eyelid dysfunction is episodic and may be precipitated by eyelid closure, often in patients with extrapyramidal disorders). Ocular bobbing is usually only present with an impaired conscious level and extensive pontine disease. The spontaneous rapid downward movement of the eyes is followed by a slow drift back to the original position. It is thought to occur because of the tendency of such patients to have roving eye movements but, without any horizontal gaze, the only possible movements are in the vertical plane. Oscillopsia is an illusion of movement, or oscillation of the environment. The patient may complain that static objects are oscillating either from side to side or up and down. This can occur with nystagmus or any of the other tonic abnormalities of eye movement, but these can sometimes be difficult to demonstrate. This symptom, although uncommon, can be extremely distressing and disabling. Tortopia is the illusion of transient tilting or inversion of the environment. This can occur with cerebellar ischaemia.

3.3.7 Disturbance of hearing, balance and coordination Clinical anatomy ‘Dizziness’ may refer to light-headedness, imbalance, a feeling of faintness, a lack of mental clarity, or frank vertigo – the patient must be asked to explain exactly what he or she means by dizziness. Vertigo is a subjective or objective illusion of motion (usually rotation) or position. It is a symptom of dysfunction of the peripheral or central vestibulocerebellar system. Dysequilibrium is a sensation of imbalance when standing or walking due to impairment of vestibular, sensory, cerebellar, visual, or motor function, and consequently it may be due to lesions in many parts of the nervous system. Ataxia (derived from the Greek meaning ‘lack of order’) is disordered coordination of the extremities (limb ataxia), and imbalance of sitting (truncal ataxia)

or gait (gait ataxia). It is typically associated with disorders of the cerebellum or the cerebellar connections in the brainstem. However, lesions of the thalamus, particularly within the posterolateral territory supplied by the thalamogeniculate artery, may present with isolated contralateral ataxia – although more often there is an additional motor and/or sensory deficit.57 The most likely explanation is that the ventrolateral nucleus receives input from the cerebellar, vestibular and spinothalamic systems (e.g. dentato-rubro-thalamic tract). A related condition in which patients are unable to stand or even sit unsupported, in the absence of marked motor deficit, has been termed astasia and is associated with lesions in the posterolateral thalamus. Sudden unilateral hearing impairment, with or without ipsilateral tinnitus, is a symptom of dysfunction of the cochlea, vestibulocochlear nerve, or cochlear nucleus. The relationship between the ascending and descending fibre tracts and the cranial nerve nuclei in the brainstem is shown in Fig. 3.33. Clinical assessment When patients complain of an illusory sense of movement, the first step is to distinguish rotatory vertigo or tilting of the visual axis from less specific symptoms such as faintness, and then to localize the disturbance to the brainstem (central) or to the vestibulocochlear nerve or labyrinth (peripheral).81,82 To define as closely as possible the patient’s actual sensations, direct questions often have to be asked; for example, ‘is it a spinning feeling or just light-headedness?’ Descriptions that include a subjective or objective illusion of motion, such as spinning or whirling, which is usually so unpleasant that it makes the patient feel nauseated and also unable to stand, denote what is meant by vertigo. Feelings of lightheadedness, swaying, a swimming feeling, walking on air, queer head or faintness (often with accompanying feelings of panic, palpitations or breathlessness), without a feeling of motion, are non-specific and may be caused by a wide variety of systemic disturbances (usually hypotension, panic or overbreathing). Precipitating factors and premonitory symptoms may be of diagnostic value, as also may be the mode of onset (whether sudden or gradual), the duration, and the presence of any associated symptoms such as deafness, tinnitus and ear pain or fullness. Indeed, it is not so much the character of the vertigo that helps to localize the disorder as the associated features of the attacks. For example, vertigo accompanied by features of brainstem dysfunction such as diplopia and face and limb sensory disturbance, with normal hearing, points to a central cause; whereas vertigo triggered by sudden movements and positional

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 77

3.3 The diagnosis of a cerebrovascular event

Midbrain Aqueduct III nerve nucleus

Superior cerebellar peduncle

Pons Medial longitudinal fasciculus Lateral spinothalamic tract Medial lemniscus

Fourth ventricle

Medial longitudinal fasciculus

Vestibular nuclei

Cerebellar peduncle

VIIIn nucleus

Central sympathetic pathway

VIn nucleus Vn nucleus

Lateral spinothalamic tract

VIIn nucleus Corticospinal and corticobulbar tracts

(a)

Corticospinal tract

Medial lemniscus

(b) Medulla

Vestibular nuclei

XII nerve nucleus

Medial lemniscus

Central sympathetic pathway

Inferior cerebellar peduncle IXn + Xn nuclei

Fig. 3.33 Diagrammatic representation of the main anatomical structures within the brainstem: (a) midbrain; (b) pons; (c) medulla.

changes and associated with auditory or ear symptoms points to a peripheral cause.53 Unsteadiness is a fairly common symptom in stroke/ TIA patients, but unless it is associated with clearly focal symptoms or residual neurological signs of weakness or ataxia, it can be difficult to decide whether the patient means weakness, incoordination, vertigo, presyncope or anxiety, or indeed a combination of these. Sometimes it is helpful to ask patients whether they felt unsteady in the head or in the legs, or whether it was a visual problem. Patients often complain of being unsteady when they have rotational vertigo, but it is useful to ask quite specifically if there was persisting unsteadiness after the vertigo had stopped – a positive response suggesting a central rather than a peripheral problem. The most common manifestation of cerebrovascular disease involving the cerebellum or its connections is truncal ataxia. Signs in the limbs are traditionally considered to need involvement of the ipsilateral cerebellar hemisphere rather than the midline structures, and thus there are often no cerebellar signs when the limbs are

Olivary nucleus

Trigeminal spinal tract

Lateral spinothalamic tract Corticospinal tract

(c)

examined on the bed. Consequently, the problem may be overlooked if the patient’s gait is not examined. If the disorder is mild, asking the patient to turn round quickly may be the most sensitive way of detecting an abnormality. It is also worth asking about the impact of loss of visual fixation, e.g. in the dark, in a shower, or performing Romberg’s test. Most ataxic syndromes are somewhat worse when visual fixation is lost, but a marked loss of balance should make one think of disordered proprioception (i.e. sensory ataxia). Cerebellar disorders may be missed if the patient’s gait is not examined. Care must be taken when faced with a resolving hemiparesis, since at this time cerebellar-like signs can be elicited that are probably just a manifestation of impaired corticospinal control. This is most relevant when trying to distinguish the lacunar syndromes of pure motor stroke and ataxic hemiparesis from each other (section 4.3.2).

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 78

Chapter 3 Is it a vascular event and where is the lesion?

Clinical practice Many patients complain of ‘dizziness’, either around the onset of a stroke or at other times, but this term alone is too imprecise to be of localizing value, even between carotid and vertebrobasilar territories, although it is probably more frequent in the latter. Other sensory experiences, such as giddiness, light-headedness, swimminess, or a feeling of passing out may be caused by diffuse cerebral ischaemia due to postural hypotension, vasoactive medication, cardiac dysrhythmias (all common in the elderly), and also with anxiety states, panic attacks and hyperventilation. These symptoms are of no localizing value and should not be considered to be indicative of a stroke/TIA. The tendency to label any episode of ‘dizziness’, especially in the elderly, as vertebrobasilar ischaemia or – worse – insufficiency, should be strongly resisted.53 All too frequently one sees patients who have complained of ‘dizziness’ when turning their neck, have then had a cervical spine X-ray that shows some degenerative changes (extremely common in the elderly population), and are then told that they are ‘trapping the blood supply to the brain’. What little work there is to support this theory is mainly based on postmortem studies or dynamic arteriography in highly selected patients (section 7.1.5). Furthermore, many of the patients initially described under this banner had clear focal disturbances of brainstem and occipital lobe function. There is thought to be a considerable proprioceptive input from structures in and around the cervical spine to the vestibular system, and therefore for the vast majority of patients with these non-focal symptoms, the terms simply engender undue anxiety about impending strokes and divert attention from more likely and potentially treatable explanations of their dizziness. When vertigo is an isolated phenomenon, and particularly when it is induced by head movement, severe, and short-lived, it is almost always a symptom of peripheral rather than central dysfunction. The relatively common condition of benign paroxysmal positional vertigo is described in section 3.4.7. Rarely, and most commonly in patients with diabetes, occlusion of a branch of the anterior cerebellar artery or the basilar artery supplying the inner ear can cause vertigo, unilateral hearing loss or both.83 Some cerebellar strokes and lesions of more central vestibulocerebellar pathways may also mimic ‘peripheral’ symptoms. Multidirectional nystagmus that is not suppressed by visual fixation, an inability to stand without support and a negative ‘head impulse test’ in a patient with a first-ever attack of acute, spontaneous vertigo points to a central rather than a peripheral process.81,82 The head impulse test is a simple clinical

test comprising high acceleration head rotation;84 with severe unilateral vestibular weakness the normal vestibuloocular reflex is replaced by a misalignment of the eyes followed by a series of corrective saccades which are evident to the examiner.84 Sudden hearing loss may be caused by trauma, tympanic membrane rupture, viral and other infections, toxic and metabolic disorders, and ischaemia.85 How often sudden unilateral hearing loss in isolation (without vertigo or other brainstem dysfunction) is due to vascular disease is unknown, however, because histopathological examination of the temporal bone and labyrinth at postmortem is not performed routinely and, if it is, it is usually carried out so long after the onset of any hearing loss that the clinical details are unclear. Nevertheless, there is some histopathological evidence of labyrinthine infarction due to vascular disease. Even greater uncertainty prevails as to whether transient deafness in isolation is ever due to a TIA in the internal auditory artery territory, but we do not see why this should not sometimes be the case. Acute cerebellar syndromes that can mimic strokes may be caused by drug toxicity, Wernicke’s encephalopathy and Creutzfeldt–Jakob disease (sections 3.4.5 and 3.4.6).

3.3.8–The speed of onset of the symptoms and temporal course Patients with a TIA or stroke usually describe their neurological symptoms as coming on abruptly, without warning, and as being more or less maximal at onset. However, patients are often frightened by their symptoms and are probably not very accurate about the exact time (seconds vs minutes) they take to develop. If several parts of the body (e.g. face, arm and leg) are affected, the symptoms usually start in each part almost simultaneously rather than intensifying or spreading (‘marching’) from one part to another – a pattern more typical of focal epilepsy52 or a migrainous aura.86 However, the simultaneous onset of a disturbance of several body parts may not be recognized, and the event misinterpreted as ‘spread’. For example, a patient who is playing the piano at the time of a stroke will notice weakness of the hand but not weakness of the leg until he or she tries to stand up, or dysphasia until they try to speak. Because it is the suddenness of onset that stamps the event as vascular, it is useful to ask patients what they were doing at the time; if they were awake and do not remember, the onset was probably not all that sudden. Less commonly, other patterns are encountered: the symptoms may evolve steadily over minutes or hours,

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 79

3.3 The diagnosis of a cerebrovascular event

they may develop in a stuttering/stepwise fashion over several hours, and occasionally they may continue to increase over a few days. In the last case, it can be virtually impossible to decide whether this is due to the primary condition continuing to evolve, or to other factors such as the development of cerebral oedema, infection, or metabolic upset. Not surprisingly, there is no agreed definition of the widely used term ‘stroke-inprogression’ (section 11.5).87,88 Of course, in the face of a progressing deficit, it may not be possible to be absolutely certain whether the diagnosis is actually that of stroke/TIA at all, and the degree of diagnostic uncertainty will influence the extent of the investigations needed to firm up the clinical diagnosis.

3.3.9 Possible precipitants In addition to giving an impression of the suddenness of onset of the symptoms, asking the patients (or witnesses) what they were doing at the time of symptom onset can also support a diagnosis of a vascular event. • Vigorous physical activity and coitus have been associated with haemorrhagic stroke, particularly subarachnoid haemorrhage. However, apart from isolated case reports, there is no evidence that such activities precipitate TIAs and ischaemic stroke. • A change in posture, neck turning (section 7.1.5), exposure to bright or white light, bending, straining, or sneezing, exercise, a hot bath, or a heavy meal, may provoke cerebral and ocular ischaemic symptoms in people with severe carotid and vertebrobasilar occlusive disease and a compromised collateral cerebral and ocular circulation. Of course, some of these stimuli may also provoke non-vascular symptoms, such as those due to hypoglycaemia (after a large carbohydrate meal) and seizures (after exposure to bright flashing lights). • Certain circ*mstances may predispose to arterial dissection, such as neck manipulation, road traffic accidents and head injuries. There can often be a delay of days or weeks between the trauma and the first neurological symptoms (sections 7.2.1 and 8.2.13). • Drug abuse is increasingly common, it is not confined to the very youngest age groups (section 7.15), and one should therefore have a relatively low threshold for specific questioning and toxicology screening. • Symptoms beginning shortly after starting or increasing the dose of hypotensive drugs should raise the possibility of a so-called ‘low-flow’ stroke/TIA due to the combination of systemic hypotension, focal arterial stenosis/occlusion/compression and poor collateral circulation (section 6.7.5). A similar argument might also apply to symptoms present on

waking from a general anaesthetic, although other factors such as intravascular thrombosis may be present, particularly if the symptoms are referable to the posterior circulation. Obviously, patients undergoing cardiac or carotid surgery are at special risk, and this should be explained to them prior to surgery (section 7.18). • Pregnancy and the puerperium are times when otherwise healthy young women may be predisposed to stroke as a result of paradoxical embolism from the venous system of the legs or pelvis, intracranial haemorrhage due to eclampsia, or intracranial venous sinus thrombosis (section 7.14). • There is a complex relationship between migraine and cerebrovascular disease, and this is discussed in detail in sections 3.4.1 and 7.8. • There is some evidence that significant ‘life events’ in the preceding year may increase the risk of a stroke (section 6.6.18).

3.3.10 Accompanying symptoms Any accompanying symptoms may be useful in sorting out whether the pathogenesis is vascular. Headache Headache occurs in about one-sixth of patients at the onset of a transient ischaemic attack of the brain or eye, about one-quarter of patients with acute ischaemic stroke, about one-half of patients with non-traumatic intracerebral haemorrhage, and nearly all patients with subarachnoid haemorrhage.89–93 Cortical ischaemia causes headache more often than small, deep, lacunar infarcts (section 6.7.6).89,91,93 Given that ischaemic strokes are four times as common as haemorrhagic strokes, the predictive value of headache for the presence of haemorrhagic stroke is about 33%, while the predictive value of ‘no headache’ for ischaemic stroke is about 86%. The cause of the headache associated with ischaemic events is unknown. It has been suggested that it is due to the release of vasoactive substances such as serotonin and prostaglandins from activated platelets during cerebral ischaemia. Other possibilities include distortion or dilatation of collateral blood vessels, and mechanical stimulation of intracranial nociceptive afferents. Very occasionally, external carotid artery territory emboli can result in ischaemia of the scalp, and so pain. And, some headaches are probably due to anxiety and muscle tension. Headache may not only be a consequence of the vascular event, but also a marker of the underlying cause of the event. Headache in anyone over the age of 50 or so

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 80

Chapter 3 Is it a vascular event and where is the lesion?

presenting with ischaemia of the brain or eye demands immediate consideration of giant-cell arteritis (section 7.3.1). Similarly, severe pain on one side of the head, face, eye or neck before or around the time of onset is highly suggestive of carotid or vertebral arterial dissection,94 and may even be the only symptom of carotid dissection95 (section 7.2.1). Other causes of headache with focal neurological signs include carotidynia (secondary to carotid occlusion), migrainous stroke, meningitis and intracranial venous thrombosis, but usually there are other clues to these diagnoses.96

Neck stiffness Meningism means painful resistance to passive or voluntary neck flexion because of irritation of the cervical meninges by subarachnoid blood, or by inflammation. Neck stiffness caused by meningism is a common symptom and sign of blood in the subarachnoid space, but it does not occur immediately; it takes some 3–12·h, and may not develop at all in deeply unconscious patients, or in patients with minor SAH (section 3.7.1). Photophobia

Epileptic seizures About 2% of stroke patients have an epileptic seizure (sections 3.4.2, 6.7.6 and 11.8) at the onset; about half are generalized and half are partial seizures.97 These are more common with intracerebral or subarachnoid haemorrhage than with arterial ischaemic stroke.97 They are, however, a characteristic feature of intracranial venous infarction (section 7.21).3 Onset seizures probably indicate irritation or damage to the cerebral cortex by the stroke, and are associated with an increased risk of further seizures. Epileptic seizures may not only complicate stroke but also be a marker of increased risk of stroke. Looking forward, from the time of diagnosis of epilepsy, older people (>60 years of age) diagnosed with new-onset idiopathic epilepsy have an almost three-fold risk of subsequent stroke compared with others of the same age with no history of seizures.98 Looking back, since the time of diagnosis of stroke, about 3% of patients with acute stroke have a past history of epileptic seizures, one-third of these occurring for the first time in the previous year.97 Older people (>60 years of age) diagnosed with new-onset idiopathic epilepsy should therefore be screened for risk factors for stroke and treated appropriately. Vomiting Vomiting is very rare during TIA and is uncommon even in patients with stroke, at least near the onset. When it does occur, it suggests subarachnoid haemorrhage (see section 3.7.1), posterior fossa stroke (because of vertigo in some cases, and presumably because of direct involvement of the ‘vomiting centre’ in the area postrema in the floor of the fourth ventricle in others), or large supratentorial stroke causing raised intracranial pressure. Vomiting within 2·h of stroke onset is highly predictive of intracranial haemorrhage. Occasionally, patients with brainstem ischaemia can have profuse vomiting with little or no vertigo and few other clinical signs.

Patients are often photophobic and irritable for several days after SAH, presumably as a result of meningeal irritation by the blood (section 3.7.1). Hiccups and abdominal pain Hiccups are brief bursts of intense inspiratory activity, involving the diaphragm and inspiratory intercostal muscles, with reciprocal inhibition of the expiratory intercostal muscles. Glottic closure occurs almost immediately after the onset of diaphragmatic contraction, generating the characteristic sound and sense of discomfort. Hiccups usually resolve spontaneously after a few minutes. If they persist for days, they may indicate underlying structural or functional disturbances of the medulla (in the region of the vagal nuclei and tractus solitarius), or of afferent or efferent nerves to the respiratory muscles, or diaphragmatic irritation. Hiccups are well recognized in patients with lateral medullary infarction, but may also occur with a lesion of any part of the medullary region that is associated with respiratory control. Neurogenic hiccup rarely occurs in isolation; usually there are associated brainstem or long tract signs (section 11.10). If patients who have presented with a stroke also complain of abdominal pain, they may have ischaemia of the bowel or viscera – particularly if there are pointers to cardiogenic embolism being the underlying cause of the stroke (such as atrial fibrillation). Chest pain, palpitations and shortness of breath Chest pain and/or palpitations at the time of onset of the stroke/TIA suggest the possible coexistence of an acute coronary syndrome (due, among other things to simultaneous inflammation and rupture of cerebral and coronary atheromatous plaque) and cardiac dysrhythmias (as a consequence of cardiac ischaemia and as a possible source of embolism to the brain) (section 6.5). Other relevant causes of chest pain and/or shortness of

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 81

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset

breath include aortic dissection, pulmonary embolism and pneumonia. Panic and anxiety Sudden loss of limb power, speech or vision is a frightening experience which often evokes considerable anxiety and panic in patients and carers. Patients may consequently hyperventilate and in turn develop presyncopal or sensory symptoms including perioral and distal limb paraesthesia bilaterally, and even unilateral sensory symptoms. Under these circ*mstances, it is important to try to distinguish between the primary (stroke/ TIA) and secondary (panic/anxiety) symptoms. It is also important to be clear about the timing of the symptoms; for example, palpitations occurring immediately before or concurrently with stroke/TIA are less likely to be a consequence of panic and anxiety than similar symptoms that clearly follow the onset of the neurological symptoms.

3.3.11 Past medical history It is important to ask more than once about previous neurological symptoms. Many patients have told us several days after their transient ischaemic attack (TIA) or stroke about prior TIAs, having not recalled them for a number of reasons at the time of their first consultation. Sometimes this is because of anxiety, and of course a few patients are affected by altered awareness or amnesia during the event. Others may not wish to disclose such information for fear of potential repercussions for their employment or driving status. It is the doctor’s responsibility to explore these possibilities with insight, sensitivity and confidentiality, using all available sources of information. Sometimes a detailed history of previous episodes helps confirm an alternative diagnosis such as focal epilepsy or migraine, while a detailed drug history may identify previous use of aspirin, or warfarin given after an otherwise forgotten event. We would recommend working through a checklist of focal neurological symptoms such as those in Table 3.20. It is also important to ask specifically about: • common vascular risk factors (e.g. hypertension, hyperlipidaemia, diabetes, smoking); • other manifestations of vascular disease (e.g. angina, intermittent claudication of the legs); • heart disease (e.g. valvular heart disease or cardiomyopathy – many lay people think a stroke is a form of heart attack); • pointers to a thrombotic tendency (e.g. prior unexplained deep venous thrombosis);

Table 3.20 Checklist of symptoms of cerebrovascular disease. Have you ever been told that you have had a stroke, ministroke, transient ischaemic attack or brain attack? If so, when did this occur and can you describe what happened? Have you ever suddenly: Lost vision or gone blind in one eye? Had double vision for more than a few seconds? Had jumbled speech, slurred speech or difficulty in talking? Had weakness or loss of feeling in the face, arm or leg? Had clumsiness of an arm or leg? Had unsteadiness walking? Had a spinning (dizzy) sensation? Lost consciousness? How long did the symptoms last? Do you still have these symptoms? Did you see a doctor about the episode and, if so, who was it and what were you told, and were you admitted to hospital? What medications/drugs/tablets are you taking? (particularly aspirin, clopidogrel, dipyridamole, warfarin)

• and clues to vasculitis (e.g. arthralgia, skin rashes, renal problems (sections 6.7.6, 6.7.7 and 7.3).

3.3.12 Lifestyle, behaviour and family history Relevant and potentially modifiable lifestyle factors include tobacco, alcohol, saturated fat, salt, recreational drug consumption and physical activity. A family history of stroke or myocardial infarction is a risk factor for ischaemic stroke, which is attributable, at least partly, to a familial predisposition to hypertension (section 6.6.3).99,100

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset

There is a wide differential diagnosis of focal cerebral symptoms of rapid onset of short duration (Table 3.21) and long duration (Table 3.23). Tables 3.22 and 3.24 show the frequency of these possibilities in practice. Because there is a continuum of the duration of neurological symptoms caused by focal neurological ischaemia,7 patients who are being assessed within 24 hours of symptom onset and who still have focal neurological symptoms are temporarily classified as having a ‘brain attack’ (‘acute stroke syndrome’ or ‘unstable brain ischaemia’; section 3.2.3). Another way of considering

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 82

Chapter 3 Is it a vascular event and where is the lesion? Table 3.21 Differential diagnosis of transient ischaemic attack (in approximate order of frequency, depending on referral patterns). Syncope Migraine aura (with or without headache) Labyrinthine disorders Partial (focal) epileptic seizures Hyperventilation, anxiety or panic attacks, somatization disorder Intracranial structural lesion: meningioma tumour giant aneurysm arteriovenous malformation chronic subdural haematoma Transient global amnesia Acute demyelination (multiple sclerosis) Drop attacks Metabolic disorders: hypoglycaemia, hyperglycaemia, hypercalcaemia, hyponatraemia Mononeuropathy/radiculopathy Myasthenia gravis Cataplexy

Table 3.22 Final diagnosis of all 512 patients with suspected transient ischaemic attacks notified to the Oxfordshire Community Stroke Project, modified from reference 6. Confirmed transient ischaemic attacks: Incident (i.e. first-ever) Prevalent (i.e. had previous attacks) Lone bilateral blindness* Not transient ischaemic attacks: Migraine Syncope Possible TIA† ‘Funny turn’‡ Isolated vertigo Epilepsy Transient global amnesia Isolated diplopia Drop attack Intracranial meningioma Miscellaneous

209 184 11 14 303 52 48 46 45 33 29 17 4 3 2 24

*Lone bilateral blindness was later classified as a transient ischaemic attack after following up these patients and noting their similar prognosis to patients with definite transient ischaemic attack.48 †Possible transient ischaemic attack was diagnosed in patients with transient focal neurological symptoms in whom the clinical features were not sufficiently clear to make a diagnosis of definite transient ischaemic attack or of anything else. ‡‘Funny turn’ was used to describe transient episodes of only non-focal symptoms not due to any identifiable condition (e.g. isolated and transient confusion).

Table 3.23 Differential diagnosis of stroke (in approximate order of frequency, depending on referral patterns). Systemic illness, or seizure, causing apparent deterioration of previous stroke Epileptic seizure (postictal Todd’s paresis) or non-convulsive seizures Structural intracranial lesion: subdural haematoma brain tumour arteriovenous malformation Metabolic/toxic encephalopathy: hypoglycaemia non-ketotic hyperglycaemia hyponatraemia Wernicke–Korsakoff syndrome hepatic encephalopathy alcohol and drug intoxication septicaemia Functional/non-neurological (e.g. hysteria) Hemiplegic migraine Encephalitis (e.g. herpes simplex virus)/brain abscess Head injury Peripheral nerve lesion(s) Hypertensive encephalopathy Multiple sclerosis Creutzfeldt–Jakob disease Wilson’s disease

the differential diagnoses according to duration of symptoms is outlined in Table 3.25. The top five differential diagnoses of brain attack (TIA and stroke) are the ‘5 Ss’: seizures, syncope, sepsis, subdural haematoma and somatization. The clinical diagnosis of stroke (differentiating ‘stroke’ from ‘not stroke’) is accurate more than 95% of the time if there is a clear history (from the patient or carer) of focal brain dysfunction of sudden onset (or first noticed on waking), with symptoms persisting for more than 24·h. This is particularly true if the patient is elderly or has other vascular diseases or risk factors, because the risk – i.e. the prior (or pretest) probability of stroke – is greater in elderly people with vascular disease than in younger people who have no evidence of vascular disease. However, the accuracy of the diagnosis of stroke is also influenced by the timing of the assessment, the experience and confidence of the examiner, and the likelihood of other differential diagnoses in the community37,41,101–103 (Table 3.24). For example, the diagnosis of stroke (vs not stroke) can be quite difficult within the first few hours of symptom onset if it is not possible to obtain a clear history of the onset, or even the nature of the symptoms, because the patient is unconscious, confused, forgetful or dysphasic. In these patients in

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 83

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset Table 3.24 Final diagnosis in two recent studies of patients presenting with suspected TIA and stroke.44,45

Nor et al.,45 (n = 59)

Stroke ‘mimic’

Seizure Sepsis Metabolic/toxic encephalopathy Space-occupying brain lesion Tumour Subdural haematoma Syncope/presyncope Acute confusional state Vestibular dysfunction (labyrinthitis) Acute mononeuropathy Functional/somatization Dementia Migraine Spinal cord lesion Other

particular, a persistent focal neurological deficit may not be due to stroke but to epileptic seizures (postictal), sepsis, encephalitis, brain abscess, brain tumour, head trauma, or a chronic subdural haematoma.102 Symptoms or signs that are unusual in uncomplicated stroke, such as papilloedema and unexplained fever, should also call into question the diagnosis. Further, a persistent focal neurological deficit may be due to a previous stroke, and the new clinical presentation may be caused by a non-vascular problem such as pneumonia (section 11.12). In the absence of information about the rate of onset and progression of symptoms, it is helpful to search for indirect clues to their cause in the past history and physical examination, and to continue to assess the patient over time for the development of new signs such as fever, and for the improvement that characterizes most non-fatal strokes. The diagnosis of a cerebrovascular event is usually made at the bedside, not in the laboratory or in the radiology department. It depends on the history of the sudden onset of focal neurological symptoms in the appropriate clinical setting (usually an older patient with vascular risk factors) and the exclusion of other conditions that can present in a similar way.

3.4.1–Migraine Transient ischaemic attack Migraines are not considered to be TIAs, because the clinical features (Table 3.26) and prognosis for serious vascular events are very different.104 Classical migraine (migraine with aura) usually starts in younger patients who may have a family history of migraine. The attack begins with an aura that commonly consists of positive

8 (14%) 8 (14%) 7 (12%) 4 (7%) 3 (5%) 13 (22%) 3 (5%) 7 (12%)

13 (22%)

Hand et al.,44 (n = 109) 23 (21%) 14 (13%) 12 (11%) 10 (9%)

10 7 7 6 6 4 3 3 3

(9%) (6%) (6%) (5%) (5%) (4%) (3%) (3%) (3%)

symptoms of focal cerebral dysfunction that develop gradually over 5–20·min and last less than 60 min.86,105 The most common aura consists of hom*onymous, unilateral or central visual symptoms, such as flashes of light, zigzag lines, crescents, scintillations, or fortification spectra, which gradually ‘build up’, expand and migrate across the visual field. Somatosensory or motor disturbances, such as paraesthesiae or heaviness in one or more limbs, may also occur, evolving and spreading over a period of minutes in a ‘marching’ fashion (e.g. spread of tingling from hand to arm to face to tongue over several minutes). This serial progression from one accompaniment to another without delay, such as from visual symptoms to paraesthesiae and then to aphasia, is quite characteristic of migraine but can occur with cerebral ischaemia.106 Sometimes, however, the symptoms are negative and consist of ‘blind patches’ (often a hom*onymous hemianopia) and, rarely, even loss of colour vision. Headache and nausea usually follow the focal neurological symptoms immediately or in less than an hour, but in some patients they precede the neurological symptoms and in others they occur simultaneously. Often there is associated photophobia and phonophobia, which clearly help distinguish migraine attacks from TIAs. The headache usually lasts 4–72 h.86 It is not unusual for patients who have experienced classical migraine to suffer identical auras, but without headache (acephalgic migraine aura), particularly as they get older (otherwise referred to as ‘migraine aura without headache’ or ‘late-life migraine accompaniments’).107 This should not cause confusion with TIA. However, diagnostic difficulty arises when an older patient (over 40 years) with no previous history of classical migraine presents following a first-ever episode of transient symptoms of focal neurological dysfunction that are typical of a migraine aura, but without any associated headache.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 84

Chapter 3 Is it a vascular event and where is the lesion? Table 3.25 Differential diagnosis of sudden focal neurological symptoms by duration of symptoms.

Migraine Epilepsy Partial seizures Todd’s paresis Transient global amnesia Structural intracranial lesions Subdural haematoma Tumour Arteriovenous malformation/aneurysm Metabolic/toxic disorders Hypoglycaemia Hyperglycaemia Hyponatraemia Hypercalcaemia Hepatic encephalopathy Wernicke’s encephalopathy Hypertensive encephalopathy Posterior reversible encephalopathy syndrome CNS infections Encephalitis Brain abscess Subdural empyema Creutzfeldt–Jakob disease Progressive multifocal leukoencephalopathy Labyrinthine disorders Vestibular neuronitis Ménière’s disease Benign paroxysmal positional vertigo Benign recurrent vertigo Psychological disorders Hyperventilation Panic attacks Somatization/conversion disorder Head injury Multiple sclerosis Neuromuscular disorders Mononeuropathy Radiculopathy Myasthenia gravis Motor neurone disease

Transient ischaemic attack

Brain attack

Stroke

+++

+++ +++ +++

++ ++ ++

+ + –

++ + ++

++ ++ ++

++ ++ +

+++ +++ ++ ++ + + + +

++ ++ ++ ++ ++ ++ ++ ++

+ + + + ++ ++ ++ ++

+ + + – –

++ ++ ++ – –

++ ++ ++ + +

+ ++ ++ ++

++ – – +

++ – – –

+++ +++ +++ + +

– – ++ + +

– ++ ++ ++

++ ++ ++ –

++ ++ + +

+++ Common or frequent; ++ encountered regularly in a busy practice or important, treatable condition; + infrequent; – rare.

Patients with these ‘late-life migraine accompaniments’ have a favourable prognosis with a lower risk of stroke and other serious vascular events compared with TIA patients.104 A 69-year-old former general practitioner was referred by a cardiologist for suspected ocular TIAs, a diagnosis that caused him some alarm. He had unexpectedly suffered a

myocardial infarction at the age of 59, and had been receiving beta-blockers and aspirin since then. The attacks in question, of which he had now had nine, had first occurred more than 2 years previously, but had increased in frequency. Each attack began with a small, bright spot on the left or right side of the centre of vision. He verified that it occurred in both eyes by covering each eye in turn. In the course of 1–3·min, the bright spot would gradually enlarge and change

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 85

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset Table 3.26 Features of migraine aura that help distinguish it from TIA. Gradual onset over minutes (usually) Positive neurological symptoms such as visual scintillations, not blindness Intensification and ’spread’ of symptoms from point to point, or from one symptom to another symptom, over minutes Gradual resolution of symptoms over 20–60 min Headache (often unilateral and pulsatile) and nausea usually follow (but not always) Onset in young and middle-aged adults Family history of migraine is common Recurrent attacks are usually stereotyped

Fig. 3.34 A drawing by a physician who experienced ‘late-life migraine accompaniments’. He wrote the following caption: ‘These zigzag lines appear during an attack in both eyes, more often to the left than to the right of centre of vision, sometimes on both sides. The lines flicker on and off, about six times per second; they are colourless. Sometimes a dazzling white spot will appear within a line, which also turns on and off.’

into a cloudy area, until the centre of vision became obscured; the border of the opaque zone consisted of intensely bright and flickering shapes: zigzag lines, stars and sparks (Fig. 3.34). This phenomenon would remain for about 15·min; if he tried to read, he could not make out the words in the centre. Finally, the visual disturbance would disappear over a few minutes. He never had headache or other symptoms with the attacks, and was able to resume his normal activities afterwards. He could not recall that he had suffered similar attacks, or migraine, earlier in his life. This was migraine, not TIA. A detailed history eliciting the slow onset, and spread and intensification of neurological symptoms – particularly if positive and visual or referable to more

than one vascular territory during the same attack – points towards the diagnosis of migraine.86,104 Careful questioning about childhood or menstruation-related headaches, and a family history of migraine, may suggest a predisposition to migraine. If the patient is young (i.e. less than about 40 years old), with a normal heart and no other clinical manifestations of arterial disease, it is most unlikely that the symptoms are those of a TIA caused by atherothrombosis. Perhaps this explains why young female patients with suspected TIA, with or without a family history of migraine, do not seem to have any higher risk of stroke than other women of the same age. In this group, there is a remote possibility of thrombophilic and vasculitic disorders but, for the most part, these patients can be reassured that they have not ‘had a stroke’. Because the risks of stroke are very low in absolute terms in people with migraine, they do not require any potentially hazardous or costly investigations or prophylactic treatments108,109 (section 7.8). There is no convincing evidence that arteriovenous malformations or aneurysms occur more frequently than can be accounted for by chance in patients with migraine aura. There can, however, be particular concern about migraine occurring for the first time during pregnancy, because of the recognized propensity for meningiomas and possibly arteriovenous malformations to become symptomatic at this time.110 In this group of patients, we would suggest careful clinical examination looking for residual focal signs following a single attack, and a low threshold for MR scanning after the baby has been born if the attacks are multiple and stereotyped, particularly if the same side of the brain is always involved. In older patients, particularly those with their first attack of acephalgic migraine and if the neurological symptoms are other than positive visual ones, the distinction from TIAs can be much more difficult. From a purely pragmatic point of view, if there is a suspicion that the attack was a TIA, it seems prudent to modify any vascular risk factors and recommend aspirin. Although drugs used to lower blood pressure and cholesterol, and antiplatelet drugs such as aspirin, are not free of risk or expense, some of them (e.g. propranolol and aspirin) may also prevent migraines. Stroke Occasionally, the aura of a previously experienced and otherwise unremarkable attack of migraine persists for days or longer – a so-called migrainous stroke (section 7.8.1). The simultaneous occurrence of a stroke and a migraine attack may be: • coincidental because, after all, both conditions are common, the prevalence of migraine in the general

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 86

Chapter 3 Is it a vascular event and where is the lesion?

adult white population is about 10% and ischaemic cerebrovascular disease about 0.8%; • causal, either migraine may predispose to cerebral ischaemia by leading to platelet activation, arteriolar constriction and dehydration, or cerebral ischaemia may trigger a migraine attack; • a misdiagnosis (e.g. arterial dissection may cause headache and a neurological deficit due to thromboembolism that is misinterpreted as migraine); or • both are a manifestation of a disorder such as a patent foramen ovale, an arteriovenous malformation, or a mitochondrial disorder111–113 (sections 6.5.12, 7.19 and 8.2.4). The 69-year-old wife of a general practitioner had suffered from classical migraine since she was a teenager. Her attacks always began with visual fortification spectra, progressing to a hemianopia (either left or right) over the next 10·min. The visual disturbance would resolve after about 30·min, and she would then get a severe throbbing headache, nausea and vomiting. She had always recognized bright light as a potential trigger factor. Because of a family history of glaucoma, an optician recommended that she have her visual fields tested using a Humphrey perimeter. She found the flashes of white light uncomfortable, and by the end of the examination she was aware that she had her typical migrainous visual disturbance in her right visual field. The headache was unusually severe and she went straight to bed. When she woke 6·h later, the headache had resolved but the right hemianopia was still present. A brain CT scan later that day showed a recent left occipital infarct. No other explanation for the stroke was found despite detailed investigation. The hemianopia persisted, and it was presumed that this had been a migrainous stroke.

3.4.2 Epilepsy Transient ischaemic attack Partial (focal) seizures can be distinguished from TIAs because they usually cause sudden positive sensory or motor symptoms that spread or ‘march’ fairly quickly to adjacent body parts (Table 3.27).52 Although positive symptoms such as tingling and limb jerking can occur in TIAs,76 they tend to arise in all affected body parts at the same time (i.e. in the face, arm and leg together), whereas the symptoms of partial seizures spread from one body part to another over a minute or so (cf. migraine aura, where any spread is usually over several minutes; section 3.4.1). Negative motor symptoms, such as postictal or Todd’s paresis, are well recognized after a partial motor seizure or a generalized seizure with partial onset, but this should be obvious from the history unless

Table 3.27 Features of partial epileptic seizures that help distinguish them from TIAs. Young and middle-aged adults commonly, but also patients with previous cortical stroke Antecedent symptoms may occur (e.g. epigastric discomfort) which are the beginning of the seizure Onset over several seconds or 1–2 min Positive neurological symptoms usually March or ‘spread’ of symptoms over seconds Rapid resolution of symptoms over a few minutes Amnesia for the event Family history of epilepsy may be present Recurrent attacks are usually stereotyped and reduced with antiepileptic treatment (if necessary)

the patient was asleep, or is aphasic and there is no witness (section 3.3.10). Difficulty arises in the very rare patient in whom epileptic seizures actually cause transient ‘negative’ symptoms during the electrical discharge. One then has to rely on other factors, such as the patient’s age, any past history of seizures and the nature of the symptoms. For example, transient speech arrest (as opposed to aphasia with muddled language output), which is characterized by the sudden onset of cessation of speech, often accompanied by aimless staring and subsequent amnesia for the details of the episode, is usually epileptic rather than ischaemic. Transient aphasia and rarely bilateral blindness or amnesia can also sometimes be epileptic. Distinguishing between seizures and TIAs is occasionally very difficult.76 Sometimes it requires prolonged and careful observation of, and interaction with, the patient (and witnesses) over several visits. The diagnosis should not be rushed, as patients may have several types of attack with different causes. Initially, it is important to explain the diagnostic uncertainty to the patient and why it is necessary to establish the precise diagnosis; to exclude a structural intracranial lesion (with brain imaging); and to advise the patient not to drive or put themselves in a position in which they would be a danger to themselves or others if they were to have another attack. Clues to the diagnosis of epilepsy can be obtained by taking a careful, targeted history, particularly with respect to previous epileptic seizures and any symptoms that immediately preceded the onset of the neurological deficit (e.g. epigastric discomfort and olfactory or gustatory hallucinations may be due to an initial partial sensory seizure involving the temporal or frontal lobes). The interictal electroencephalogram (EEG) can be normal but not usually in patients having several seizures a day. Ambulatory or telemetered EEG requires skilled and careful analysis because of the potential for a false

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 87

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset

positive diagnosis of epilepsy. Nevertheless, concomitant video recording of an attack can be very useful. If recurrent partial seizures are suspected, brain MR should be performed to seek a focal structural lesion too small to be seen on CT, such as mesial temporal sclerosis, a hamartoma, or a tumour, usually in the frontal or temporal lobe. However, transient abnormalities that might be mistaken for areas of ischaemia on brain MR have been reported in patients with partial status epilepticus, although the abnormalities tend not to respect normal vascular distributions and to resolve with time. A 64-year-old woman described about 20 attacks of pins and needles in her right arm and leg over a period of 6 weeks. Each attack lasted for about 5·min, and there were no associated symptoms. On closer interrogation, she said that the sensation started in the right foot and then over a period of about 1·min spread ‘like water running up my leg’ to involve the whole leg and arm. Each attack was identical. A brain CT scan showed a glioma in the left parietal lobe. A diagnosis of partial sensory seizures secondary to the glioma was made. Stroke An epileptic seizure is one of the most common reasons for the misdiagnosis of stroke.41,44,45,103 The usual scenario is a patient with postictal confusion, stupor, coma or hemiparesis (Todd’s paresis) where the preceding seizure was unwitnessed or unrecognized.103 Todd’s paresis refers to a focal deficit that may follow a focal motor seizure. In general, the deficit usually resolves very quickly (within minutes), but in patients with prior cerebral damage, e.g. an earlier stroke, the deficit can last for several days, or there may be permanent deterioration, making the distinction from a recurrent stroke very difficult. It is noteworthy that involuntary convulsive-like movements sometimes occur in patients with brainstem strokes.114 These movements vary in nature, frequency and triggers, including fasciculation-like, shivering, jerky, tonic-clonic and intermittent shaking movements. The episodes usually consist of brief clonic contractions of the proximal and distal upper extremities and occur in paroxysms lasting for 3–5 s.114 Their pathogenesis is uncertain but may be related to ischaemia or haemorrhage in the corticospinal tracts, They should not be misinterpreted as decerebrate postures or seizures.

3.4.3–Transient global amnesia Transient global amnesia (TGA) is not considered to be a TIA because its prognosis is so much better. It is a very characteristic clinical syndrome that typically occurs in a middle-aged or elderly person.115 There is a sudden dis-

Table 3.28 Features of transient global amnesia that help distinguish it from TIAs. Older adults Sudden onset Anterograde and retrograde memory disturbance Repetitive questioning Able to recognize familiar individuals and places with no loss of personal identity Headache can occur No focal neurological symptoms or signs All but complete recovery within a few hours to a day or so, leaving a dense amnesic gap for the attack itself Recurrent attacks are unusual, and if they occur may be due to migraine or partial seizures Diagnosis is very difficult without a witness

order of memory, which is often reported as confusion.116 For some hours, the patient cannot remember anything new (anterograde amnesia) and often cannot recall more distant events over the past weeks or years (retrograde amnesia). There is no loss of personal identity, personality, problem-solving, language or visuospatial function, and the patient can perform complex activities such as driving a car (Table 3.28). The patient seems healthy, but repetitively asks the same questions and has to be reminded continually of what he or she has just asked or done. There are no other symptoms, apart from perhaps headache (migraine is more common in TGA patients than in controls, and there are some theoretical reasons to implicate migraine in its pathogenesis).117 In women, episodes are mainly associated with an emotional precipitating event, a history of anxiety and a pathological personality. In men, they occur more frequently after a physical precipitating event, e.g. diving into a swimming pool.115 After the attack, anterograde memory returns to normal, but the patient cannot remember anything that happened during the amnesic period. The retrograde amnesia tends to diminish with recovery, but leaves a short retrograde gap for the period of the TGA. Recurrences are not very common, about 3% per year. The early reports tended to view TGA as a type of TIA, and of course one does encounter patients with pure amnestic strokes (section 3.3.3). However, careful case–control studies have shown that both the prevalence of vascular risk factors and the rates of subsequent stroke or myocardial infarction are much lower in the TGA group and are indeed approximately similar to those in the age-matched general population.115,118,119 There is no overall difference in the prevalence of epilepsy between TGA patients and control groups, but an important minority (7%) of TGA patients go on to

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 88

Chapter 3 Is it a vascular event and where is the lesion?

develop epilepsy, usually within a year of presentation. These patients tend to have had shorter attacks of ‘TGA’, lasting less than 1·h, and to have already experienced more than one attack at the time of presentation. It must be presumed that in this minority of cases with ‘TGA’, the cause was temporal lobe epileptic seizures from the beginning. Because of this, driving is generally allowed after a single episode of TGA but not after multiple attacks. Functional imaging, such as perfusion CT, has shown transient bilateral hypoperfusion of the medial temporal lobes or basal ganglia, and unilateral left opercular and insular hypoperfusion, during the TGA attack.120 However, it is not known whether this perfusion defect is a primary event or secondary to a reduction in cerebral metabolic activity in these areas, or due to other factors. Diffusion-weighted MRI during episodes of TGA shows transient signal abnormalities, consistent with oedema, in the uncal-hippocampal region unilaterally or bilaterally in a variable proportion of patients.119,121,122 Ischaemia, due to arterial thromboemboli or venous stasis, has been proposed as a potential mechanism.119,123,124 It has also been hypothesized that diverse stressful precipitants may trigger a release of an excitotoxic neurotransmitter (such as glutamate), which then temporarily shuts down normal memory function in the medial temporal regions via spreading depression, leading in turn to a fall in cerebral perfusion.

Table 3.29 Possible explanations for transient neurological symptoms in patients with a structural intracranial lesion. Partial epileptic seizure Spreading depression of Leao (as some suppose occurs in migraine) Vascular ‘steal’, leading to focal brain ischaemia adjacent to the tumour Vessel encasem*nt or direct compression by the tumour mass Indirect compression of vessels by herniating tissue or coning (usually a preterminal event)

ischaemic stroke, if CT brain imaging is not done early, and certainly within 10–14 days after the onset of symptoms by which time the characteristic changes will have resolved (section 5.4.1).13 Fortunately, intracerebral haemorrhage is visible indefinitely on gradient echo MRI sequences as a low signal (black) ring or dot (section 5.5.1). A 77-year-old man, whilst standing in his garden, suddenly developed weakness of the right arm and unsteadiness on walking. He had no headache or vomiting. He sat down and within 3·h was ‘better’, although on closer questioning it emerged that his arm had not returned to normal for about 3 days. A CT brain scan 8 days later showed a small resolving haemorrhage in the left putamen. This was not a TIA in either its duration or pathology.

3.4.4 Structural intracranial lesions Although structural intracranial lesions often cause focal brain dysfunction, the symptoms and signs generally progress over several days or weeks, or even months, they are not abrupt like strokes or TIAs. About 5–10% of patients who are initially diagnosed as having an acute stroke on the basis of the history and examination turn out to have a structural intracranial lesion on brain imaging, such as a subdural haematoma, tumour or arteriovenous malformation.44,45 Possible explanations for transient neurological symptoms are listed in Table 3.29. Non-traumatic intracerebral haemorrhage Non-traumatic intracerebral haemorrhage almost invariably causes prolonged or permanent focal neurological dysfunction, although there are isolated reports of neurological deficits that have resolved within a few days, and even 24·h. However, one sometimes needs to probe quite hard to be sure about the duration of symptoms from the history. For example, ‘I got better in 24·h’ does not necessarily mean ‘I got completely back to normal’, which is what we are referring to when defining a TIA. Intracerebral haemorrhage can be minsterpreted as TIA, or

Subdural haematoma Transient focal neurological symptoms such as hemiparesis, hemisensory loss, aphasia and speech arrest may occur as a consequence of a chronic subdural haematoma (SDH)125 (section 8.10). And SDH may rarely present with the abrupt onset of focal neurological signs, which then persist and so mimic a stroke.45 Usually the patient has a confusional state and a history of minor neurological symptoms starting a few days to a few weeks previously. SDH should be suspected if there is any evidence of subacute onset of focal neurological symptoms and signs; persistent headache; more confusion and drowsiness than expected from the neurological deficit; or a progressive or fluctuating clinical course. Although SDH occurs in all age groups, it is more frequent among the elderly, alcoholics and patients receiving anticoagulants or who have a bleeding disorder. About 50% of patients recall sustaining a head injury, which may have been mild. SDH is a rare complication of lumbar puncture, and spontaneous or traumatic intracranial hypotension. In the acute phase, the brain CT scan usually shows unilateral hyperdensity in the subdural space, ipsilateral

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 89

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset

effacement of sulci, and a mass effect causing shift of the midline and distortion of the ventricular system (Fig. 3.35). There is a transitional stage between 7 and 21 days, during which clotted blood evolves on the CT scan from a region of hyperintensity to one of isointensity. This can easily be missed, particularly if the subdural haematomas are bilateral, when there is little if any midline shift or asymmetrical ventricular compression. Thereafter, the haematoma becomes hypodense and thus more easily visible. MRI is more sensitive than CT for detecting subdural haematomas. In a few cases, a subdural haematoma may prove to be an incidental finding, and the TIA can be attributed to more commonly recognized factors such as carotid artery disease. Tumour Occasionally – perhaps if there is bleeding into a tumour – there can be a sudden focal neurological deficit caused by a brain tumour, although this usually lasts longer than 24·h.126,127 Greater diagnostic difficulty arises when a structural lesion causes a partial and non-convulsive seizure, with or without postictal (Todd’s) paresis, or intermittent focal neurological symptoms (so-called ‘tumour attacks’) that do not seem to be epileptic in origin.44,45,49 The clinical features associated with these ‘tumour attacks’ are usually focal jerking or shaking, pure sensory phenomena, loss of consciousness and isolated aphasia or speech arrest.126 With time, a more obvious epileptic syndrome often declares itself. A 78-year-old woman complained of many attacks of weakness and clumsiness of the left arm over a period of 4 months. The weakness came on suddenly, lasted for between 10 and 45·min, and was not associated with any other symptoms. In between attacks, she had no symptoms. A diagnosis of transient ischaemic attack was made by both her general practitioner and the neurologist, and she was started on aspirin. A brain CT scan was performed later because the attacks continued, and this showed a meningioma involving the right frontal lobe (Fig. 3.36). A final diagnosis of ‘tumour attacks’ was made, but in view of the patient’s age, the neurosurgeon thought an operation was not advisable. Of course it is always important to consider intracranial tumours in the differential diagnosis of a patient with a progressing neurological deficit, particularly if the rate of progression is relatively slow (over days, weeks or months) and there is a history of recent headache or epileptic seizures, papilloedema, or any evidence of a primary extracranial source of malignancy. In practice, most patients with suspected stroke who turn out to have a cerebral tumour have actually had a focal

neurological deficit (e.g. hemiparesis) gradually evolving over several weeks to several months. Papilloedema is very uncommon in acute stroke. The brain tumours that tend to bleed are glioblastoma, choroid plexus papilloma, meningioma, neuroblastoma, melanoma, hypernephroma, lymphoma, endometrial carcincoma, bronchial carcinoma and choriocarcinoma (section 8.5.1). The CT scan may then disclose intracerebral haemorrhage in an unusual location or associated with considerable surrounding oedema, or multiple haemorrhages, or it may show other metastatic deposits (Fig. 3.37). If there has been no haemorrhage into the tumour, the CT scan usually shows a region of low attenuation (due to cerebral oedema) with imprecise boundaries and some mass effect, causing effacement of sulci or ventricular compression. If there is a breakdown of the blood–brain barrier, as commonly occurs in patients with cerebral tumours, then intravenous injection of iodinated contrast material leaks into the tumour and is seen on CT as an area of diffuse or peripheral enhancement (Fig. 3.38). A similar, but nevertheless distinctive, appearance of enhancement of the gyri, which may be seen within 1–2 weeks of a recent cerebral infarct, is also due to breakdown of the blood–brain barrier (Fig. 3.39). This can make interpretation difficult, particularly if only a contrast-enhanced scan is performed. If the clinical examination and CT are ambiguous, patients should be followed up clinically (because with a tumour they usually deteriorate), and a follow-up CT or MRI performed within a few weeks to a few months (depending on the patient’s progress) to see if the lesion has resolved or, if it is a tumour, continued to grow. Aneurysms and arteriovenous malformations Intracranial aneurysms and cerebral arteriovenous malformations (sections 8.2.3, 8.2.4 and 9.1.1), and particularly cavernous malformations (section 8.2.5), can cause transient focal neurological deficits mimicking a TIA or ischaemic stroke, and even transient monocular blindness.128,129 In the Scottish Intravascular Vascular Malformation Study, focal neurological deficits accounted for about 7% of presentations of symptomatic arteriovenous malformations (Rustam Al-Shahi Salman, personal communication). Possible explanations are embolization of thrombus from within an aneurysm (section 7.6), a partial seizure, a small intraparenchymal haemorrhage (e.g. from a cavernous malformation) and venous hypertension. Vascular steal around an arteriovenous malformation is not a certain mechanism. In many cavernoma cases it is impossible to identify the cause of some

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 90

Chapter 3 Is it a vascular event and where is the lesion?

(b) (a)

(c)

Fig. 3.35 (a) A plain CT brain scan showing acute right-sided subdural haematoma as an area of high density in the subural space (arrows). (b) A plain CT brain scan showing a subacute left frontal subdural haematoma (arrows) which is isodense, illustrating the difficulty of identifying subdural haematomas on CT at this stage of their evolution. (c) A plain CT brain scan showing chronic bilateral subdural haematomas as areas of low density in the subural space (arrows).

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 91

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset

Fig. 3.36 A post-contrast CT brain scan showing a meningioma involving the posterior part of the right frontal lobe (arrow) in a patient who presented with a history suggesting transient ischaemic attacks; over 3 months, this elderly lady had eight attacks of sudden onset of weakness affecting the left arm, each lasting a few minutes. Between attacks the neurological examination was normal.

(a)

(b) Fig. 3.37 A plain CT brain scan showing multiple areas of high density (arrows) due to spontaneous intracerebral haemorrhage in what turned out at postmortem to be metastases from choriocarcinoma.

Fig. 3.38 A plain CT brain scan of a brain tumour showing (a) a region of low attenuation (due to cerebral oedema) with imprecise boundaries (arrows) and some mass effect, causing effacement of sulci. (b) After intravenous contrast injection there is enhancement (arrow) of the tumour.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 92

Chapter 3 Is it a vascular event and where is the lesion?

haemorrhagic telangiectasia, a past history of subarachnoid haemorrhage, or, reputedly, a cranial bruit.128

3.4.5–Metabolic and toxic disorders Encephalopathies due to metabolic or toxic disturbances generally present with epileptic seizures or a subacute alteration in consciousness, and few if any focal neurological signs (perhaps only generalized hyperreflexia, with or without extensor plantar responses). However, occasionally the presentation may be acute, with focal neurological symptoms and signs, which can therefore mimic a stroke or TIA.44,45 The causes of the metabolic encephalopathy include hypoglycaemia, hyperglycaemia, hyponatraemia, hypercapnoea and hypoxia. Hypoglycaemia

Fig. 3.39 A post-contrast CT brain scan 8 days after onset of ischaemic stroke showing a region of low attenuation consistent with infarction (white arrowheads) in the territory of the left middle cerebral artery, and a serpiginous gyral pattern of high attenuation due to breakdown of the blood–brain barrier (black arrows). There is considerable mass effect, with displacement of midline structures (open arrows), effacement of the lateral ventricle and obliteration of sulci.

transient focal neurological attacks which do not sound epileptic, and where early imaging shows no fresh blood or is done too late (or is not done at all in known cases); a psychogenic cause would not be surprising given the stress of knowing there is an intracranial lesion which might bleed. Although aneurysms and arteriovenous malformations are one of the more common causes of intracerebral haemorrhage (sections 8.2.3 and 8.2.4), they may also cause focal neurological symptoms and signs as a direct consequence of their mass effect and compression of surrounding structures, e.g. a third cranial nerve palsy from a posterior communicating artery aneurysm. Brain CT, with and without contrast, MRI or intraarterial cerebral angiography, may be required to make the diagnosis. Just occasionally, arteriovenous malformations can be suspected clinically, e.g. if there is cutaneous evidence of the Sturge–Weber syndrome or hereditary

Hypoglycaemia may cause both transient and permanent focal neurological symptoms and signs, which can occur without any of the characteristic adrenergic symptoms41,103,130,131 (section 7.16). The patient is almost always being treated with hypoglycaemic agents, but other possibilities include factitious hypoglycaemia, insulinoma, Addison’s disease, hypopituitarism, hypothyroidism, sepsis, terminal malignancy, liver failure, starvation, or drugs whose adverse effects include hypoglycaemia. The symptoms tend to be stereotyped in an individual, and are most likely to occur before meals (i.e. before breakfast or during the night, after fasting for some time), after exercise, or 2–3·h after the ingestion of sugars and starch; they are relieved by glucose administration. The blood glucose is usually less than 2.5·mmol/L at the onset of an attack, but it may have normalized spontaneously or with glucose administration by the time the patient is seen. Particularly if a diabetic patient presents with a suspected stroke early in the morning, it is imperative that hypoglycaemia be considered and appropriate treatment given rapidly, although a blood glucose estimation should be mandatory for all patients with suspected stroke. A 63-year-old man was admitted to a surgical ward because of abdominal pain, diarrhoea and signs of intestinal obstruction, 2 months after a partial gastrectomy for a gastric ulcer. He was treated with intravenous fluids and received nothing by mouth, whilst awaiting investigations. Five days after admission, he became restless and confused. Two hours later, left-sided weakness was noted. The neurology resident found a left hemiplegia and hypoaesthesia, including the face, with completely normal consciousness. The provisional diagnosis was ischaemic stroke, but the next morning the nurses found him unresponsive. On repeat examination,

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 93

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset

there was no eye opening to pain, extension of the arm was the best motor response, and there was no verbal response. The pupils were normal, but no oculocephalic responses could be elicited. The glucose level was 1.5·mmol/L. After administration of 100·mL of 50% glucose, he quickly recovered. The blood glucose concentration should be assessed in all patients with suspected stroke. Hyperglycaemia

Wernicke’s encephalopathy

The hyperosmolarity of hyperglycaemia can itself cause regional reduction of cerebral blood flow, focal neurological deficits, focal epilepsy, stroke-like syndromes and cerebral infarction132 (section 7.16). These deficits usually resolve as the blood glucose returns to normal.

Wernicke’s encephalopathy can sometimes be mistaken for a stroke because of an unusually sudden onset of diplopia (due to abducens and conjugate gaze palsies and nystagmus), ataxia and mental confusion, either singly or, more often, in various combinations.134 It is due to thiamine deficiency and is seen mainly, although not exclusively, in alcoholics and the malnourished elderly. The diagnosis can be difficult because the history of symptom onset may be unclear due to recent alcohol intoxication or Korsakoff’s psychosis, in which retentive memory is impaired out of all proportion to other cognitive functions in an otherwise alert and responsive patient. Pointers to the diagnosis on examination are signs of peripheral neuropathy (present in more than 80% of patients), postural hypotension (autonomic neuropathy), disordered cardiovascular function (tachycardia, exertional dyspnoea, minor ECG abnormalities), and impaired capacity to discriminate between odours (in the chronic stage of the disease due to a lesion of the medial dorsal nucleus of the thalamus). The diagnosis is supported by a marked reduction in red cell transketolase activity (one of the enzymes of the hexose monophosphate shunt that requires thiamine pyrophosphate as a cofactor) and striking improvement in the oculomotor disorder (but not other disorders such as amnesia, polyneuropathy and blindness) within hours of the administration of thiamine; completely normal values of transketolase are usually attained within 24·h. Failure of the ocular palsies to respond to thiamine within a few days should raise doubts about the diagnosis. The medial thalamic and periaqueductal lesions may be demonstrated on brain MRI.

Hyponatraemia The most frequent neurological manifestation of hyponatraemia is reduced level of attention or consciousness. Long tract signs (6%), tremor (1%), hallucinations (0.5%), myoclonus and seizures (3%) may also occur. The reason for the appearance of focal symptoms, such as long tract signs, is unclear. The deficits usually respond to correction of the hyponatraemia, but this should be done cautiously to minimize the risk of developing central pontine myelinolysis. Of course, a stroke and in particular subarachnoid haemorrhage, may be the cause of hyponatraemia (section 11.18.2) and many cases are iatrogenic (e.g. use of intravenous fluids, diuretics and carbamazepine). Given that many patients with stroke have been smokers, a bronchial carcinoma should be considered if the pattern of serum and urine osmolalities suggests the syndrome of inappropriate ADH secretion. Hypercalcaemia The usual neurological manifestations of hypercalcaemia are either psychiatric symptoms or an encephalopathy, often accompanied by headache and sometimes seizures. Occasionally, cerebral infarction occurs. A proposed mechanism is vasospasm but this has not been established. Hepatic encephalopathy It is most unlikely that an acute or subacute portosystemic encephalopathy would be mistaken for a stroke or TIA. One case report in the literature describes a woman who developed a left hemiparesis following a general anaesthetic for manipulation of her left shoulder and at the time of transfer to the rehabilitation hospital

was found to have a mild hepatic encephalopathy.133 Acquired (non-Wilsonian) hepatocerebral degeneration can present with dysarthria, ataxia and intention tremor as well as upper motor neurone signs in the limbs, although there is usually a history of progressive deterioration. The catch is that routine liver function tests may be normal but the serum ammonia is raised and other tests of liver function may be abnormal (e.g. prothrombin time).

If Wernicke’s encephalopathy is a diagnostic possibility, take blood (for red cell transketolase, thiamine and glucose levels) and then treat immediately with thiamine (100 mg IV or IM daily for 3–5 days, followed by 100 mg orally, daily) and a glucose infusion (since hypoglycaemia can precipitate Wernicke’s disease); do not waste time waiting for the blood results to come back.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 94

Chapter 3 Is it a vascular event and where is the lesion?

Hypertensive encephalopathy Although encephalopathy caused by malignant or accelerated phase hypertension is now rare, it occurs most commonly in patients with an established history of hypertension, although only when the diastolic blood pressure exceeds about 150·mmHg. However, it is also seen in previously normotensive patients whose cerebral autoregulation is normal and so easily exceeded by a rapid rise in blood pressure, sometimes with a diastolic blood pressure of no more than 100·mmHg (i.e. associated with conditions such as pre-eclampsia, acute nephritis, phaeochromocytoma, renin-secreting tumour, ingestion of sympathomimetic drugs and tricyclic antidepressants, ingestion of tyramine in conjunction with a monoamine oxidase inhibitor, head injury, autonomic hyperactivity in patients with the Guillain–Barré syndrome or spinal cord disorders, and with baroreceptor reflex failure after bilateral carotid endarterectomy). The cause is thought to be widespread cerebral oedema resulting from a breakdown of cerebral blood flow autoregulation. The pathological hallmark is fibrinoid necrosis in resistance vessels in the retina and kidneys. The clinical picture is usually dominated by the subacute onset of headache, nausea, vomiting, confusion, declining conscious state, blurred vision, seizures and focal or generalized weakness. There may be focal neurological signs and hypertensive retinopathy (including papilloedema) (Fig. 3.30). Hypertensive encephalopathy can sometimes be mistaken for intracranial venous thrombosis (particularly in pregnancy) or an arterial ischaemic stroke, especially if there is doubt about the onset of symptoms (i.e. if the patient is confused or obtunded) and if the blood pressure is only moderately elevated. Even a severely elevated blood pressure can be a consequence as well as a cause of stroke, but in these cases there is unlikely to be any evidence of end-organ damage such as retinopathy (section 11.7.1). The aim of treatment should be a smooth reduction in blood pressure over hours rather than minutes; indeed precipitous falls in blood pressure can cause ischaemic stroke (section 4.2.4). Posterior reversible encephalopathy syndrome Posterior reversible encephalopathy syndrome (PRES), also called the reversible posterior leukoencephalopathy syndrome, is a fairly recently described cliniconeuroradiological syndrome that has many clinical and imaging similarities to hypertensive encephalopathy, and eclampsia, except there is usually no evidence of hypertensive end-organ damage, and the prognosis is

generally very good.135 Patients present with cortical blindness, headache, altered mental function (drowsiness, coma) and seizures, and brain MR usually shows very extensive white matter abnormalities, typical of oedema, in the posterior regions of both hemispheres (Fig. 3.40). This syndrome has been associated with a number of immunosuppressive therapies (e.g. cyclosporin, cisplatinum, tacrolimus, and intravenous immunoglobulin therapy), vasculitis and drug withdrawal (e.g. clonidine). The pathophysiology is uncertain, but endothelial dysfunction (sometimes induced by cytotoxic therapies), breakdown of the blood–brain barrier and vasogenic oedema are thought to play a role, as in hypertensive encephalopathy. Indeed, PRES and hypertensive encephalopathy are probably the same syndrome, but with different causes. There is usually, but not always, fairly rapid improvement of both the clinical and MR abnormalities with treatment that removes the cause (e.g. stopping cytotoxic drugs and lowering blood pressure) and controls the symptoms (antiepileptic drugs). However, the condition is not always reversible, confined to the posterior regions of the brain, or the white matter, and we agree with others that PRES is a misnomer.135 Wilson’s disease Wilson’s disease may very rarely present as an acute stroke-like problem.136

3.4.6 Central nervous system infections Encephalitis, brain abscess and subdural empyema If a patient with a focal neurological deficit has an altered conscious state and a fever, then focal or multifocal infection of the brain (meningoencephalitis, brain abscess, progressive multifocal leucoencephalopathy, or subdural empyema) or elsewhere in the body (e.g. pneumonia, urinary tract infection, venous thrombophlebitis, septicaemia) needs to be considered (and perhaps even treated empirically), particularly if the prevalence of infection (e.g. HIV) in the community is high.101,137 Indeed, among patients with a focal neurological deficit from a previous stroke, a subsequent systemic infection may lead to neurological decompensation and apparent worsening of the neurological deficit, prompting an incorrect diagnosis of recurrent stroke or ‘extension’ of the initial stroke. There may also be a history of subacute evolution of systemic upset (fever, malaise, lethargy) and focal neurological symptoms, as well as seizures, meningism or a predisposing

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 95

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset

(a)

(b)

Fig. 3.40 (a) A T2-weighted MR scan of a patient with posterior leucoencephalopathy syndrome, which occurred following rapid blood transfusion after placental abruption. On the day of onset (a), the patient suddenly became cortically blind in association with a severe headache. She also had several seizures. On T2 MR, there were areas of increased signal in

both occipital regions (long arrows). Note also the areas of increased cortical signal in the parietal regions (short arrows). All her symptoms resolved in 7 days. (b) Three months later, there was no residual visual field defect and the brain scan was normal.

condition such as sinusitis, mastoiditis, otitis, pneumonia, or congenital heart disease. Of course, it is conceivable that the diagnosis of stroke is correct, and the cause of the stroke is an infection (section 7.11). The EEG, brain CT, MR scan and cerebrospinal fluid (CSF) are usually characteristically abnormal. For example, the parasitic infection cysticercosis may present like a stroke, but the patient will probably have lived in an endemic area and the CT scan often shows an area of calcification within a cyst, as well as scattered foci of parenchymal calcification. In unilateral subdural empyema, the EEG shows extensive unilateral depression of cortical activity and focal delta waves lasting up to 2·s, and the CT shows a non-hom*ogeneous lenticular or semilunar extracerebral lesion with mass effect. However, CT can be normal early on, and MRI may be required to identify the subdural empyema. Likewise, brain CT or MRI of a cerebral abscess usually shows a low-density lesion that is not in a specific vascular territory and has peripheral ring enhancement following intravenous contrast; but sometimes the presentation

is acute and the typical CT findings of an abscess can be delayed for several weeks. It can also be difficult to distinguish radiologically between herpes simplex encephalitis of the frontotemporal lobes and a middle cerebral artery territory infarct (Fig. 3.41). If the patient has focal neurological signs, is systemically unwell with fever and has what looks like a normal CT scan, then an abscess, subdural empyema, and meningoencephalitis need to be excluded by MR scan, EEG and – if safe to do so – CSF examination. Creutzfeldt–Jakob disease Sporadic Creutzfeldt–Jakob disease (CJD) typically presents with a combination of rapidly progressive (over weeks) dementia and myoclonus, which may be accompanied by symptoms and signs of visual, pyramidal and cerebellar dysfunction. However, CJD may occasionally present acutely with a stroke-like syndrome.138,139

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 96

Chapter 3 Is it a vascular event and where is the lesion?

cerebellar artery territory, followed by the anterior inferior cerebellar artery territory. Patients with infarcts in the territory of the superior cerebellar artery or multiple cerebellar arteries do not tend to have isolated spontaneous prolonged vertigo.144 Ménière’s disease Ménière’s disease is characterized by repeated crises of quite severe rotatory or whirling vertigo, sometimes causing falling to one side, which can be sudden in onset and last from several minutes to a few days.143 Varying degrees of nausea and vomiting, unilateral (initially) low-pitched tinnitus, sensorineural deafness and a feeling of fullness or pressure in the ear are almost always present as well. It usually begins in middle age. The diagnostic difficulty may come with the first attack, when the auditory symptoms may be mild or non-existent and caloric tests normal. In these cases, cerebellar infarction is a possible differential diagnosis (section 3.3.7). Benign paroxysmal positional vertigo

Fig. 3.41 A plain CT brain scan showing a region of low intensity in the right temporal lobe (arrow) due to herpes simplex encephalitis.

If a ‘stroke’ patient deteriorates and develops myoclonus or dementia, think of sporadic Creutzfeldt–Jakob disease and order one or more EEGs (and MRIs).

3.4.7–Labyrinthine disorders Vestibular neuronitis (labyrinthitis) This is probably the commonest cause of severe acute vertigo.140–143 There is associated nausea, vomiting, nystagmus and ataxia, but no deafness or tinnitus. The acute symptoms usually last for several days and may be followed by positional vertigo for some weeks or months. There is scant evidence for viral involvement of the superior part of the trunk of the vestibular nerve. A very similar, if not identical, clinical syndrome can clearly be caused by minor cerebellar strokes and multiple sclerosis, and the diagnosis is only revealed by MR imaging144 (section 3.3.7). The arterial territory most commonly involved in cerebellar infarction causing vertigo is the medial branch of the posterior inferior

Benign paroxysmal positional vertigo is characterized by recurrent episodes of vertigo and nystagmus that occur only after suddenly changing the position of the head – for example, when looking up, rolling over in bed and turning the head toward the affected ear, lying down, bending over and straightening up.141 The vertigo is usually severe but very brief in duration, certainly less than 1·min, and usually less than 15 s. Hearing is normal. There may be a history of recent head trauma, viral illness, stapes surgery or chronic middle ear disease, but many cases are idiopathic. The cause in most cases is canalolithiasis – particles shed by the otolith membrane float freely in the endolymph of the posterior semicircular canal. A much less common cause is cupulolithiasis in which detritus gets caught on the cupula and causes positional vertigo due to excessive loading of the cupula. The diagnosis is established from the history and using the Dix–Hallpike manoeuvre. Benign recurrent vertigo Attacks of spontaneous vertigo not accompanied by cochlear or neurological symptoms in young or middleaged adults, lasting from 20·min to a few hours, have been called ‘benign recurrent vertigo’. The demographics and precipitants of this condition are very similar to those of migraine, and the attacks may respond to standard migraine prophylactic drugs, suggesting that it may be a migraine variant – but the exact pathophysiology remains uncertain.145

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 97

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset

3.4.8–Psychological disorders Psychological factors may give rise to subjective experiences (i.e. symptoms), such as hemiparaesthesiae during a panic attack, that commonly need to be differentiated from symptoms that arise from disease (e.g. TIA or stroke), physiological factors (e.g. physiological tremor), behaviours (e.g. excessive rest), and cultural or external factors (e.g. compensation and the welfare state). When there is no disease pathology it becomes tempting to suggest that the symptom must be ‘not real’ and that it is psychogenic and due to a conversion disorder. Conversion disorder is a psychoanalytic concept that describes the occurrence of motor or sensory neurological symptoms other than pain and fatigue that cause distress, are not explained by disease, not malingered (simulated for clear financial or material gain) and are thought to relate to psychological factors.74 Whatever the cause of the symptoms, those unexplained by disease may be as or even more distressing than those caused by disease. The most common psychogenic symptoms mistaken for TIA and stroke are functional weakness and sensory loss.103 Functional weakness Patients with functional weakness are likely to show inconsistency on observation of their behaviour. For example, their gait on entering the consulting room or at the beginning of the examination may differ from their gait when they leave the consulting room, or at the end of the examination. Further, their weakness when they have to take their clothes on or off may be inconsistent with their weakness when they do another functional task, such as when they have to get something from their bag. If the patient complains of unilateral leg weakness, Hoover’s sign of functional weakness is the only test that has been found in controlled studies to have good sensitivity and specificity.146 It relies on the principle that we extend our hip when flexing our contralateral hip against resistance. It can be performed in two ways: Hip extension: look for a discrepancy between voluntary hip extension (which is often weak) and involuntary hip extension (which should be normal) when the opposite hip is being flexed against resistance. It is important when testing involuntary hip extension to ask the patient to concentrate hard on their good leg. Hip flexion: test hip flexion in the ‘weak’ leg while keeping your hand under the good heel. Feel for the absence of downward pressure in the good leg.

A similar principle can be used to examine weakness of hip abduction which may initially be weak but then come back to normal if tested simultaneously with the ‘good side’. These tests, although useful, should be interpreted cautiously because: • pain in the affected hip may produce greater weakness on direct, compared with indirect, testing as a result of attentional phenomena (related to pain rather than weakness); • cortical neglect can cause a positive Hoover’s sign; • the test may be mildly positive in normal individuals because of a splinting effect; • none of the studies testing its utility were blinded and none mention the problem of neglect. ‘Collapsing’ and ‘give-way’ weakness are commonly present in patients with functional weakness: one moment the limb is strong, the next it is not. This should be not be described as ‘intermittency of effort’ since it is not possible to directly assess someone’s effort. In this situation, normal power can often be achieved transiently with encouragement, for example by saying to the patient, ‘At the count of three, stop me from pushing down . . . ’ Alternatively, gradually increase the force applied to the limb, starting gently and building imperceptibly up to normal force. Inability to understand the instruction, pain in the relevant joint, being generally unwell, and a misguided eagerness of some patients to ‘convince the doctor’ may be problematic. These concerns have been vindicated in the small number of validity studies which have found that this sign is a rather poor discriminator between functional and disease-related symptoms.74 However, it is our impression in daily practice that it is quite a good discriminating test, and more reliable than Hoover’s sign. Functional sensory disturbance Functional sensory disturbance may be reported as a symptom, or detected first by the examiner. While a number of functional sensory signs have been described, none appear to be specific and they should not therefore be used carelessly to make a diagnosis. • Patients may describe sensory loss that ends where the leg or arm ends, at the shoulder or groin. • The hemisensory syndrome is a disturbance which is usually considered by the patient as ‘something is not right down one side’ or that they feel ‘cut in half’. It is usually more patchy than complete, and of variable intensity. There are often accompanying symptoms of intermittent blurring of vision in the ipsilateral eye (asthenopia) and sometimes ipsilateral hearing problems.

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 98

Chapter 3 Is it a vascular event and where is the lesion?

• Exact splitting of sensation in the midline is said to be a functional sign because cutaneous branches of the intercostal nerves overlap from the contralateral side, so organic sensory loss should be 1 or 2 cm from the midline. However, midline splitting can also occur in thalamic stroke. Similarly, patients with disease should not report a difference in the sensation of a tuning fork placed over the left compared to the right side of the sternum or frontal bone, as these bones are a single unit and must vibrate as one. But these signs seem to be as common in patients with disease and so cannot be recommended.74,147 • Asking patients to ‘Say yes when they feel touch and no when they don’t’, in order to see whether they say ‘no’ when the affected area is touched, is difficult to interpret because the patient may be saying ‘no’ to mean ‘not as much’. • Although often considered left-sided, a systematic review found only a slight left-sided preponderance for functional motor and sensory symptoms.148 • ‘La belle indifférence’, an apparent lack of concern about the nature or implications of symptoms or disability, has no value in discriminating whether the patient is making an effort to appear cheerful in a conscious attempt to avoid being labelled as depressed, or factitious because he or she is deliberately making up the symptom and is not concerned about it. The diagnosis of functional motor and sensory symptoms depends on demonstrating positive functional signs as well as the absence of signs of disease. Most of these signs relate to inconsistency, either internal (for example, Hoover’s sign reveals discrepancies in leg power) or external (for example, tubular field defect is inconsistent with the laws of optics). However, although inconsistency may be evidence that the signs are functional, it does not indicate whether they are consciously or unconsciously produced, and a positive functional sign does not exclude the possibility that the patient also has disease. Dizziness Panic attacks can present somatically with dizziness, a fear of embarrassment, and an inability to escape from situations in which they are likely to occur, such as supermarkets. However, anxiety and phobic avoidance of situations, or head positions, that bring on dizziness does not necessarily indicate a psychogenic aetiology. For example, physiological vestibular sensitivity to certain visual stimuli such as patterned lines or bright lights (sometimes called visual vertigo) may cause symptoms

that also come on in crowded places. Asking the patient to hyperventilate to see if that reproduces the symptoms might appear straightforward, but this is often falsely positive in patients with dizziness caused by disease. Depersonalization and derealization also may be described by the patient as ‘dizziness’. If this sensation is there all the time, the patient may have depersonalization disorder (a chronic form of dissociation). Speech and swallowing symptoms Dysphonia is the most common functional speech complaint. Often the clinical presentation is of whispering or hoarse speech that is initially thought to be laryngitis by the patient, but then persists for months or years. The possibility of spasmodic adductor or abductor dysphonia must always be considered. Functional dysarthria typically resembles a stutter or is extremely slow with long hesitations that are hard to interrupt. The speech may be telegrammatic, consisting only of the main verbs and nouns in a sentence. In its extreme form the patient may become mute. However, these types of speech disturbance can also be seen in patients with disease. Word-finding difficulty is a common symptom in anyone with significant fatigue or concentration problems and may compound any functional dysarthria. True dysphasia as a more severe functional symptom, however, is rare. Globus pharyngis or functional dysphagia is also common. The patient normally complains of a sensation of a ‘ball in the throat’ and investigations do not reveal a cause. Visual symptoms Intermittent blurring of vision that returns to normal if the patient screws up their eyes tight and then relaxes them again is commonly reported. Some of these patients have convergence or accommodation spasm, with a tendency for the convergence reflex to be transiently overactive, either unilaterally or bilaterally. In this situation lateral gaze restriction can sometimes mimic a sixth nerve palsy, but miosis reveals the diagnosis. Voluntary nystagmus appears to be a ‘talent’ possessed by around 10% of the population. Tests for functional visual acuity problems are described in detail elsewhere.149 Simple bedside tests for a patient complaining of complete blindness are to ask them to sign their name or bring their fingers together in front of their eyes (which they should be able to do). They may have a normal response to menace and optokinetic nystagmus with a rotating drum. Decreased acuity in one eye can be assessed with a ‘fogging test’ in which plus lenses of increasing power are placed in front

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 99

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset

of the ‘good’ eye until the patient can only be using their ‘bad’ eye to see. Spiral or tubular fields are common, and are often asymptomatic. Remember to test the fields at two distances when looking for a tubular field. Patients with functional hemianopia have been described who have hom*onymous hemianopia with both eyes open and then, inconsistent with this, a monocular hemianopia in one eye with a full field in the other eye. Monocular diplopia or polyopia may be functional but can be caused by ocular pathology.

3.4.9 Head injury Head injury may cause a stroke, and a stroke may cause head injury. Without a detailed history of the onset from relevant people as well as the patient (witnesses, family members, ambulance officers, family doctor) it can be difficult to be sure what has happened. For example, head injury may cause intracranial haemorrhage, which can be mistaken for a primary stroke if the patient is amnesic for the injury and has no external scalp evidence of injury. Head injury can also cause ischaemic stroke as a result of arterial dissection (section 7.2). On the other hand, stroke can precipitate a fall, causing head injury and, if the CT scan shows intracranial haemorrhage, the primary stroke event may be missed. Brain imaging may shed some light on the cause; intracranial haemorrhage from head injury tends to be more common in the frontal and anterior temporal regions, superficial and multiple, and there may be accompanying extension into the subarachnoid space and associated skull fractures (Fig. 3.42) (section 9.1.4). One can often see soft tissue swelling of the scalp and brain contusions on the CT scan.

3.4.10 Multiple sclerosis Patients with multiple sclerosis (MS) usually develop focal neurological symptoms in their third or fourth decade (as opposed to the seventh and eighth decades for stroke). The onset is typically subacute, with the symptoms coming on over days or even weeks. The diagnosis is seldom difficult because these patients are usually young without any vascular diseases or risk factors, the symptoms are as often positive as negative, there are usually more neurological signs than symptoms (compared with TIA and stroke, in which there are often more symptoms than signs), and some may have evidence of disease in other parts of the central nervous system that is asymptomatic but readily detected by clinical examination or brain/spinal cord MRI. Also, there may be a history of previous episodes that are typical of multiple

Fig. 3.42 A plain CT brain scan showing blood in the subarachnoid space in a patient who presented with confusion following a head injury. There is blood in the anterior interhemispheric fissure (straight arrow) in a pattern suggesting spontaneous subarachnoid haemorrhage. However, the presentation apparently followed a head injury and the suggestion of frontal haemorrhagic contusion (curved arrows) clouded the interpretation of the CT scan and delayed recognition of the true pathology (anterior communicating artery aneurysmal haemorrhage).

sclerosis, such as optic neuritis or transverse myelitis. However, sometimes symptoms do seem to start abruptly and so can mimic a TIA or stroke.150 The location and shape of the lesions on brain imaging are usually fairly characteristic (Fig. 3.43) (i.e. discrete round or oval lesions in the white matter of the cerebral or cerebellar hemispheres and brainstem, in the corpus callosum, and adjacent to the temporal horns of the lateral ventricles, not corresponding to territories supplied by specific cerebral arteries). In addition, the cerebrospinal fluid usually shows raised immunoglobulin (IgG) and oligoclonal bands, which are not present in the serum. However, none of these features is specific; for example, oligoclonal bands may be found in patients with acute stroke, particularly if due to vasculopathies associated with Behçet’s disease, systemic lupus erythematosus and sarcoidosis, although in many of these cases oligoclonal bands are also present in the serum (section 7.3).

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 100

100

Chapter 3 Is it a vascular event and where is the lesion?

Fig. 3.43 A fluid-attenuated inversion recovery (FLAIR) MR brain scan showing typical ‘flame-shaped’ periventricular white matter lesions (white areas) in a patient with multiple sclerosis.

3.4.11 Neuromuscular disorders Mononeuropathy and radiculopathy It is not uncommon for TIA and stroke affecting just the hand or arm to have to be distinguished from a median neuropathy at the wrist (carpal tunnel syndrome), an ulnar neuropathy at the elbow, a radial nerve palsy, or cervical radiculopathy.44,151 This is because peripheral nerve and nerve root lesions may occasionally cause sudden onset (or sudden awareness, perhaps waking up from sleep) of persisting focal sensory or motor symptoms, which can be confused with ‘pseudoradicular’ strokes (those due to small lesions in the contralateral precentral or postcentral gyrus, corona radiata, or thalamus) (section 3.3.5). Similarly, acute tetraparesis and cranial polyneuropathy associated with acute polyradiculoneuropathy (e.g. Guillain-Barré syndrome and its Miller Fisher variant) can sometimes be mistaken for a brainstem stroke.152–154 However, the physical signs (i.e. lower motor neurone signs and/or sensory loss to pain in a dermatomal or nerve distribution) are different from those of an intracranial cortical/subcortical lesion which tend to be associated with upper motor neurone signs and/or loss of discriminative/‘cortical’ sensations, such as joint position sense and twopoint discrimination ability. Nevertheless, there are patients with the ‘cortical hand syndrome’ (section 3.3.4) in whom such distinguishing features are not present. Nerve conduction studies will usually identify a peripheral mononeuropathy unless the problem is very acute, but they are less helpful for more proximal radicular problems unless there are clearly absent F waves.

Myasthenia gravis Although the symptoms of myasthenia gravis usually come on gradually, there are instances of fairly rapid development, sometimes precipitated by infection (usually respiratory), drugs, or emotional upset.155,156 If the muscles of the eyes (levator palpebrae and extraocular muscles) and less often the face, jaw, throat and neck are the first to be affected (causing ptosis, diplopia, dysarthria or dysphagia), this can be mistaken for a brainstem vascular event, but the weakness tends to persist (if left untreated) and to fluctuate, increasing as the day wears on. Excess fatigability can be demonstrated by asking the patient to sustain the activity of the symptomatically involved muscles. Conversely, muscle power improves after a brief rest, or in response to 10 mg of edrophonium (tensilon) intravenously, or neostigmine (1.5·mg) intramuscularly. A preceding dose of atropine 0.5 mg IV not only counteracts parasympathetic overstimulation but also serves as a placebo control for the motor effects. However, because false positive (e.g. motor neurone disease) and false negative (about 10% of cases) results are fairly common, and adverse reactions can occur rarely, the tensilon test is not used unless neurophysiological tests are unavailable or a rapid diagnosis is required. The presence of anti-acetylcholine receptor antibodies confirms the diagnosis in most patients but in those presenting acutely with a suspected ‘stroke’ the test result takes too long to obtain to be useful. Motor neurone disease It is surprising how often patients with bulbar motor neurone disease (MND) are said to have had a brainstem

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 101

3.4 Differential diagnosis of focal cerebral symptoms of sudden onset

stroke. This is probably because MND symptoms sometimes seem to start surprisingly quickly and, early on, there may be little more than subtle dysarthria without any typical features of MND. What then tends to happen is that a brain CT or MRI is performed, and shows leukoaraiosis which is linked with ‘stroke’, despite this finding being quite common in the MND age group. The more useful investigation is electromyography which may reveal subclinical evidence of denervation which could not be caused by a stroke. Also, the MND patient continues to deteriorate, quite unlike a stroke patient.

3.4.12 Important non-focal disorders Syncope Syncope is probably the most important non-focal syndrome of impaired consciousness to distinguish from TIA and stroke, accounting for about one-fifth of patients presenting to the emergency room with a suspected brain attack in a recent study.45 In another study, in which at least two neurologists aimed to validate the diagnosis of TIA made by general practitioners (GPs) and by hospital emergency physicians, the most frequent conditions misdiagnosed as TIAs were transient disturbances of consciousness, mental status and balance.157 Compared with neurologists, GPs considered ‘confusion’ and ‘unexplained fall’ more often compatible with TIA and ‘lower facial palsy’ and ‘monocular blindness’ less often compatible with TIA. Syncope is defined as a loss of consciousness and postural tone due to a sudden fall in blood flow to the brain.158–161 Sometimes this may occur abruptly, without any warning (e.g. due to aortic stenosis, complete heart block), but more commonly there is a preceding feeling of light-headedness, faintness or ‘dizziness’ (not rotational vertigo), bilateral dimming or loss of vision (not to be confused with lone bilateral blindness), sounds seeming to be distant, generalized weakness, and symptoms of adrenergic activity such as nausea, hot and cold feelings and sweating.52 During the attack, the patient is pale, sweaty, clammy and floppy, rather than cyanosed and rigid as in an epileptic seizure. The pulse is absent or difficult to feel (but this cannot be relied on) and the patient may be incontinent of urine. If the patient is lying flat and not held upright (by someone or by an obstacle), then consciousness is regained within seconds, and there is very little mental confusion or difficulty in recalling the warning symptoms (unless there has been head trauma). Additional features during unconsciousness may include multifocal, arrhythmic, myoclonic jerks (particularly in patients who are held upright), head turns, oral automatisms, righting movements (sustained

head-raising or sitting up), eye movements (upward or lateral deviation of the eyes) and visual and auditory hallucinations, all of which are likely to lead to an erroneous diagnosis of epilepsy. There are usually no focal neurological symptoms unless the drop in blood pressure occurs in the presence of severe occlusive arterial disease in the neck, or impaired cerebral autoregulation due to a previous stroke. The key to the diagnosis is a sound history from a witness as well as the patient – use the telephone if necessary.159–161 Making a correct diagnosis is important, because some causes are quite serious (e.g. Stokes–Adams attacks) and if misdiagnosed as a transient ischaemic attack may lead to the patient being denied an effective and possibly life-saving treatment (e.g. a pacemaker). Drop attacks ‘Drop attacks’ are episodes of sudden loss of postural tone which cause the patient – usually a middle aged or elderly woman – to fall to the ground without apparent loss of consciousness, vertigo or other sensation. The attack occurs without warning and is not induced by a change of posture or movement of the head. The patient may be unable to rise immediately after the fall, despite being uninjured, perhaps because of the surprise. The most common differential diagnosis is vertebrobasilar ischaemia, but here there is usually some warning that the patient is going to fall, there are brainstem symptoms, such as vertigo or diplopia, and the limb weakness is persistent.53,157 Not a single patient in the New England Medical Center Posterior Circulation Registry had a drop attack as the only symptom of posterior circulation ischaemia.54 Drop attacks have also been attributed to syncope, epilepsy, tumours in the region of the foramen magnum or third ventricle, vestibular disease, myxoedema, old age and even subconscious guilt. In the vast majority of cases no cause is found, and although the attacks may persist, the patients do not seem to have a high risk of stroke or other vascular events and so the episodes are called ‘cryptogenic drop attacks’. Rarely the patient may have parasagittal motor cortex/subcortex ischaemia in the territory of both anterior cerebral arteries.162 Cataplexy Cataplexy is characterized by transient episodes of sudden bilateral loss of muscle tone, weakness and areflexia provoked by emotions, and is highly specific to the narcolepsy syndrome.163–165 Although excessive daytime sleepiness usually begins several months before the onset of cataplexy, up to 10% of patients have

101

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 102

102

Chapter 3 Is it a vascular event and where is the lesion?

cataplexy first. Laughter is the most common precipitant, but other forms of emotion and athletic activities can induce cataplexy. The three typical situations that trigger cataplexy, and differentiate it from other types of muscle weakness, are ‘when hearing and telling a joke’, ‘while laughing’, and ‘when angry’. Severe attacks cause complete paralysis except for the respiratory muscles, whereas the more common partial episodes cause patients to drop objects or to sit down or stop walking. Less than 1% of attacks cause unilateral weakness and so might be confused with a TIA. Momentary attacks are the usual pattern, and they generally last less than a minute. Prolonged episodes may be associated with hallucinations. Rarely, cataplexy can be almost continuous (‘status cataplecticus’). Despite the recent discovery of hypocretin deficiency in the lateral hypothalamus as a cause of narcolepsy, the cause of cataplexy remains unclear. Serotonin syndrome The serotonin syndrome is a dose-related range of toxic symptoms caused by excessive stimulation of the central and peripheral nervous system serotonergic (5-HT1A and 5-HT2) receptors. It is a result of increased serotonin concentrations associated with serotinergic drug use (e.g. serotonin precursors, serotonin agonists, serotonin releasers, serotonin reuptake inhibitors, monoamine oxidase inhibitors, some herbal medicines, and antimigraine medications such as sumatriptan and dihydroergotamine). The clinical picture includes mental-status changes, neuromuscular abnormalities and autonomic hyperactivity166 which may be mistaken for stroke (with or without a complication such as infection) because the onset is quite rapid and patients have hyperreflexia and clonus (often greater in the legs than arms).

3.4.13 Neuroimaging in the diagnosis of focal neurological symptoms of sudden onset Transient ischaemic attacks The main purpose of brain imaging in patients with a suspected transient ischaemic attack (TIA) is to identify a relevant focal ischaemic lesion on MR DWI in cases in which the diagnosis of TIA is uncertain, and to exclude any underlying structural intracranial lesion and, rarely, an intracerebral haemorrhage which may present like a TIA.13,167–169 Because the yield of CT for detecting structural lesions is only about 1% in patients with suspected TIA, the routine examination of every patient who has had a single TIA with CT is controversial (section 6.8.3).

We do not image the brain of patients with transient monocular blindness but do image patients with (potentially transient) ‘brain attacks’ with MR DWI. This is because DWI may not only help confirm the diagnosis,169 but also helps localize the ischaemia to one or more arterial territories and so differentiate carotid from vertebrobasilar ischaemia, and sometimes cardiac embolism in clinically uncertain cases. This is particularly important when carotid surgery is being considered. However, there are limited data on the cost-effectiveness of brain imaging in TIA patients.167,170 Stroke The diagnosis of stroke remains primarily clinical, and the main role of brain imaging is to exclude non-vascular structural pathology as the cause of the symptoms, and establish the underlying vascular pathology (Chapter 5) and aetiology (Chapters 6, 7 and 8). The choice of CT or MRI will depend on the question being asked of the test; local availability; how ill, confused or restless the patient is; and cost and effectiveness. If CT is available, the imaging strategy producing the highest number of quality adjusted life years (QALYs) at lowest cost is to ‘scan all patients immediately’ with CT, because CT is practical, quick (a few minutes to scan the brain), widely available, easy to use in ill patients, affordable (£44–130), it accurately identifies intracranial haemorrhage as soon as it occurs (section 5.4.1) and it is essential to image suspected SAH (section 3.7.3).13,167 Strategies that delay CT scanning reduce QALYs and increase cost. However, CT has limitations. Intracerebral haemorrhage will be misinterpreted as ischaemic stroke if CT is not done within 10–14 days after stroke;167 delays in seeking medical attention or poor access to CT for stroke will result in failure to identify up to three-quarters of cases of intracerebral haemorrhage, and may lead to inappropriate treatment (e.g. antiplatelet drugs or carotid revascularization) for many. Although CT shows positive features of ischaemic stroke in many patients with moderate and severe stroke scanned 2–7 days after the event, early signs of ischaemia, within 3–6 h, are difficult to recognize on CT (section 5.4.2). Furthermore, many patients with mild stroke never develop a visible infarct on CT, irrespective of when they are scanned.167 The Acute Cerebral CT Evaluation of Stroke Study (ACCESS), in which as many doctors and radiologists as possible worldwide interpret typical CT scans of stroke over the Internet, aims to improve recognition of early signs of infarction (http:// www.neuroimage.co.uk/access/).171 The advantage of MRI DWI is that it shows ischaemic changes as early as a few minutes after stroke as a bright

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 103

3.5 Differential diagnosis of transient monocular blindness

white lesion (‘light bulb’), and so shows more ischaemic strokes than CT or conventional MRI.13,14,169,172,173 It also shows more ischaemic strokes several weeks later, and therefore is particularly useful for positive identification of an ischaemic stroke in patients presenting up to 8 weeks after stroke.18 For patients presenting too late for CT to show intracerebral haemorrhage, gradient echo MR sequences reveal previous intracerebral haemorrhage indefinitely as a low signal (black) ring or dot, and can therefore distinguish previous haemorrhage from infarction. However, a limitation of MRI is that it may not identify hyperacute intracerebral haemorrhage correctly, and it can fail to detect subarachnoid haemorrhage. Further, it is difficult to use routinely in acute, particularly severe, stroke; it is less often available than CT; it requires more cooperation by the patient for a longer time; it is very noisy and it upsets confused patients. About one-fifth of patients cannot undergo MRI because they are too ill or confused, or have an intraocular or intracerebral metallic foreign body or pacemaker.

3.4.14 Electroencephalography in the diagnosis of focal neurological symptoms of sudden onset With the increasing availability of neuroimaging, the indications for an EEG in patients with suspected TIA or stroke are now very limited. An EEG may be helpful when the clinical diagnosis of TIA is in doubt and partial (focal or localization-related) seizures are a possibility. However, while about one-third of all patients with clinically definite epilepsy consistently have epileptiform discharges on the waking interictal EEG, and about one-half do so on some occasions with repeated sleep-deprived recordings,174 we know very little about the sensitivity and specificity of the test when the diagnosis is in doubt. Further difficulty arises as a result of the poor specificity of EEG abnormalities (all too frequently, patients with non-epileptic events such as TIAs are reported to have an abnormal EEG). The role of the EEG in patients with persisting neurological signs is to help exclude a number of conditions that may mimic stroke, particularly when brain imaging is normal. For example, non-convulsive status epilepticus can present with a sudden confusional state, and in both Creutzfeldt–Jakob disease and herpes simplex encephalitis, where there can be clinical deterioration with new focal neurological signs, the EEG can have characteristic abnormalities (although not in all the patients all the time). The usual abnormal EEG findings in acute stroke, at least in fairly large hemispherical cortical and subcortical stroke, are a localized reduction of normal cortical rhythms and presence of localized slow-wave abnormality.

Focal EEG slowing, however, is not specific, and only indicates the presence and side of the lesion. Although it has been suggested that the EEG may help distinguish between small deep (lacunar) and cortical infarction, the clinical features and brain CT or MRI scan are far more effective tools. A normal EEG may help to confirm the diagnosis of ‘locked-in syndrome’ caused by a lesion in the ventral pons (section 3.3.2), and can also lend some support to a diagnosis of a somatization disorder, at least if the clinical deficit is extensive.

3.5 Differential diagnosis of transient monocular blindness

Transient monocular blindness (TMB) and amaurosis fugax are two different terms for describing exactly the same symptom – blurring or loss of vision in the whole or part of the visual field of one eye. It is commonly caused by ischaemia of the retina (usually as a result of embolism from atherothrombosis of an artery between the heart and the eye, or from embolism form the heart), and sometimes by ischaemia of the anterior optic nerve (usually due to disease of the posterior ciliary artery, section 3.5.2).175,176 However, there are other causes of TMB which are important to differentiate because the prognosis and treatments differ (Table 3.30). Many are uncommon and frequently go unrecognized if the patient is not examined during the acute episode (which of course is very difficult to achieve). Patients with TMB should be considered for a competent ophthalmological examination to exclude primary disorders of the eye before concluding that the explanation is necessarily vascular disease.

3.5.1 Retinal disorders Retinal migraine or ‘vasospasm’ Migraine with aura (classical migraine) is usually ushered in by ‘positive’ binocular visual symptoms. However, transient monocular visual symptoms followed by pulsatile headache have occasionally been described in patients with known migraine, and classified as ‘retinal migraine’.177–183 TMB as a result of thromboembolism is distinguished from retinal migraine on the basis of the symptoms; the former is characterized by the abrupt onset of ‘negative’ monocular visual phenomena (blindness), which is painless and usually lasts only a few minutes, while

103

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 104

104

Chapter 3 Is it a vascular event and where is the lesion? Table 3.30 Causes of transient monocular blindness. Retinal disorders (section 3.5.1) Vascular Atherothromboembolism or other arterial disorders (e.g. dissection affecting the proximal internal carotid artery, and giant-cell arteritis affecting the posterior ciliary artery) Embolism from the heart (section 6.5) Low retinal artery perfusion pressure Retinal migraine High resistance to retinal perfusion Intracranial vascular malformation Central or branch retinal vein thrombosis Raised intraocular pressure (glaucoma) Raised intracranial pressure Increased blood viscosity (section 7.9) Malignant arterial hypertension (section 3.4.5) Retinal haemorrhage Retinal detachment Paraneoplastic retinopathy Phosphenes Lightning streaks of Moore Chorioretinitis Optic nerve disorders (section 3.5.2) Anterior ischaemic optic neuropathy Malignant arterial hypertension (section 3.4.5) Papilloedema Optic neuritis and Uhthoff’s symptom Dysplastic coloboma Eye/orbital disorders (section 3.5.3) Vitreous haemorrhage Reversible diabetic cataract Lens subluxation Orbital tumour (e.g. optic nerve-sheath meningioma)

the latter is characterized by the gradual build-up of transient monocular visual impairment (i.e. scotoma or blindness), which is usually incomplete and may be associated with ‘positive’ visual symptoms (e.g. scintillations) lasting for up to an hour, as well as a pulsatile headache or orbital pain. Sometimes, however, it can be very difficult to distinguish retinal ischaemia from retinal migraine, particularly in older patients without any headache. It has been suggested that non-invasive investigations, such as carotid ultrasound, may help distinguish between the two – i.e. tight stenosis of the internal carotid artery on the symptomatic side suggests retinal ischaemia (due to artery-to-artery embolism), and the absence of carotid disease favours retinal migraine. However, this is only circ*mstantial evidence. Patients have been described with frequent (1– 30 episodes per day), stereotyped episodes of brief (lasting less than 3·min) unilateral visual loss caused by presumed vasospasm. Fundus examination during the episodes shows constriction of the retinal arteries and segmentation in a thin and slowly moving column of

blood. The calibre of the retinal vessels is restored with the return of vision. Occasional patients have responded to a calcium channel blocker. However, as there is really no proof of vasospasm, we have to be very cautious in the interpretation of these observations in what appear to be very rare patients. Patients with TMB have a better prognosis than patients with TIA of the brain, despite similar degrees of carotid stenosis184 (section 16.11.8). The reason remains uncertain, but is unlikely to be that some of the cases of TMB were really cases of retinal migraine. The question therefore arises, how do we define retinal migraine? Is it a diagnosis based on the clinical symptoms and signs, the results of investigations, the response to treatment, or the prognosis? It has been traditionally diagnosed on the basis of the clinical history, but this may be non-specific, and it is seldom possible to examine the patient during the episode to see if ‘vasospasm’ is present. Any carotid disease may be coincidental, and of course the response to treatment and the prognosis are hardly helpful at the time the diagnosis has to be made. We cannot, therefore, provide any definitive answer. Arteriovenous malformation Anterior and middle fossa dural arteriovenous malformations very rarely cause TMB. When they do it is probably because of transient lowering of retinal arterial pressure associated with shunting of blood away from the ophthalmic artery. Central or branch retinal vein thrombosis Thrombosis of the central retinal vein, or branch retinal vein, sometimes presents with attacks of TMB.175 The visual loss tends to be patchy rather than complete. The fundoscopic appearance is characteristic: engorged retinal veins and multiple retinal haemorrhages (Fig. 3.44). Angle-closure glaucoma This typically presents in people over the age of 50 years who are hypermetropic (long-sighted). Apposition of the peripheral iris to the trabecular meshwork decreases the outflow of the aqueous humour which in turn increases the intraocular pressure and reduces perfusion pressure to the choroid, retina and disc. Transient monocular visual disturbance may occur, particularly in poor light when the pupil is dilated. The onset is usually subacute. Vision may be decreased, blurred, foggy, or smoke-like, and the patient may see haloes around lights. Some patients complain of light sensitivity during the attack and most, but not all, have eye pain, which may radiate

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 105

3.5 Differential diagnosis of transient monocular blindness

attenuated retinal arterioles, electroretinography demonstrates abnormal cone- and rod-mediated responses, and antiretinal antibodies may be identified in the serum. Over the subsequent months, progressive visual loss occurs, during which time a small-cell carcinoma of the lung often declares itself. Phosphenes

Fig. 3.44 An ocular fundus photograph showing engorged retinal veins and multiple retinal haemorrhages due to central retinal vein thrombosis. Also reproduced in colour in the plate section.

to the side of the head. The symptoms may last from a few minutes to hours. Eye pain and the stereotyped recurrence of attacks under certain lighting conditions are useful clues to the diagnosis, as are red eye, cloudy cornea and oval pupil (Fig. 3.27). The intraocular pressure should be checked in most, if not all, patients with TMB, but in those who have glaucoma it is not always raised between attacks. Retinal and other intraocular haemorrhages A small retinal haemorrhage may cause sudden reduced vision in one eye that resolves within hours. The diagnosis should be evident on ophthalmoscopy, particularly if the pupil is dilated. Similarly, vitreous and anterior chamber haemorrhage may cause TMB. The cause may be suspected from the history; for example, preretinal haemorrhage may occur during physical exertion, sexual activity, or Valsalva manoeuvre.

Phosphenes are flashes of light and coloured spots that are induced by eye movement in a dark environment and occur in the absence of luminous stimuli. They may occur with disease of the visual system at many different sites, such as optic neuritis in the recovery phase, perhaps as a result of the mechanical effects of movement of the optic nerve. However, mechanical pressure from a sharp tap on the normal eyeball may also induce a phosphene, as every child discovers, by stimulating the retina. Phosphenes may also occur in a healthy darkadapted closed eye after a saccade (flick phosphene). Lightning streaks of Moore In a dark environment, elderly people frequently experience recurrent, brief, stereotypic, vertical flashes of light in the temporal visual field of one eye, which are elicited by eye movement. These are known as Moore’s lightning streaks and are benign. It is believed that, with advanced age, the posterior vitreous may collapse and detach from the retina, leading to persistent vitreoretinal adhesions. The mechanical forces associated with eye movement exert traction on the macula and retina, and induce the photopsias (subjective sensations of sparks or flashes of light). Chorioretinitis Macular disease due to chorioretinitis or retinal pigmentary degeneration can sometimes lead to loss of vision in bright light.

Paraneoplastic retinopathy Transient episodes, lasting seconds to minutes, of painless monocular dimming of the central field of vision and overwhelming visual glare and photosensitivity when exposed to bright light suggests not only photoreceptor dysfunction due to transient retinal ischaemia, but also paraneoplastic retinopathy. Patients may also experience transient bizarre entoptic symptoms (alterations in normal light perception resulting from intraoptic phenomena, akin to the subjective perception of light resulting from mechanical compression of the eyeball – phosphenes). Ophthalmoscopy usually reveals

Central serous retinopathy This typically affects young men. Over a period of hours to days, central vision becomes blurred associated with various degrees of metamorphopsia (defective, distorted vision), micropsia (objects appear smaller than their actual size due to segregation of the photoreceptors), chromatopsia (objects appear unnaturally coloured), central scotoma, and increasing hyperopia (far-sightedness). It lasts for days to weeks, usually resolving within 4–8 weeks. Visual acuity in the acute stage ranges from 6/6 to 6/60 and averages 6/9. On fundoscopy, there is an accumulation

105

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 106

106

Chapter 3 Is it a vascular event and where is the lesion?

of transparent fluid between the retinal pigment epithelium and the outer segments in the macular area, elevating the retina and causing a circ*mscribed area of retinal detachment at the posterior pole. The diagnosis is confirmed by fluorescein angiography. Typically the fluorescein enters into the blister and stains its contents, identifying one or more leakage points.

3.5.2 Optic nerve disorders Anterior ischaemic optic neuropathy Anterior ischaemic optic neuropathy (AION) is due to ischaemia in the territory of supply of the posterior ciliary arteries, which are branches of the ophthalmic artery, and which supply the anterior part of the optic nerve, the choroid and outer retina. Because the optic disc is close to an arterial border zone between the territories of the two major posterior ciliary arteries, it is particularly liable to ischaemia when the systemic blood pressure falls, when the intraocular pressure rises, or when there is occlusive disease of local small arteries. Less often, AION is caused by embolism from the heart or proximal arteries to, or in situ occlusion of, the posterior ciliary arteries or the arterioles feeding the anterior part of the optic nerve, e.g. by giant-cell arteritis (Fig. 3.45) and other types of vasculitis, such as polyarteritis nodosa (section 7.3). Atherosclerosis of the posterior ciliary artery is another presumed cause, given that patients with AION have an increased prevalence of hypertension and diabetes and an increased risk of subsequent cerebrovascular and cardiovascular events, but we are not aware of any histological proof of atheromatous occlusion of these arteries. Ischaemia of the optic disc is characterized clinically by the sudden onset of painless visual loss in one eye.185 This may involve the whole field, but it tends to be more severe in the lower half because the upper segment of the optic disc is more vulnerable to ischaemia. In the early stages, the circulation may be so precariously balanced that minor postural change may have profound effects on the degree of visual loss. The visual loss tends to be severe, non-progressive and prolonged, but it can be brief (and present as transient monocular blindness), or it can progress over several hours or days. In AION due to giantcell arteritis, the visual loss may develop sequentially in both eyes within a few days and, rarely, almost simultaneously. Normal visual acuity does not exclude AION, and almost any part of the visual field can be affected. The common patterns of visual loss are an altitudinal hemifield defect (loss of either the upper or, more frequently, the lower half of the field in one eye), an inferior nasal segmental loss, and a central scotoma. The disc may appear normal at first, but within a few days

Fig. 3.45 A photomicrograph of anterior ischaemic optic neuropathy caused by giant-cell arteritis. The arrow indicates the infarcted optic nerve head (courtesy of Dr J.F. Cullen, Western General Hospital, Edinburgh). Also reproduced in colour in the plate section.

it becomes pale and swollen, often with small flameshaped haemorrhages radiating from the disc margin (ischaemic papillopathy), distended veins and, occasionally, cotton-wool spots due to ischaemic change in the surrounding retina. The swelling may involve only one segment of the disc, or may be more marked in one segment than in another. The disc swelling is attributed partly to leakage of plasma from damaged blood vessels and partly to arrest of axoplasmic transport along damaged nerve fibres. The disc swelling itself may be indistinguishable from that seen with raised intracranial pressure, but the vision is usually normal in the latter. In about half of the eyes with AION due to giantcell arteritis, the disc swelling has a chalky white appearance. Later, the swelling subsides, to be succeeded by

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 107

3.5 Differential diagnosis of transient monocular blindness

optic atrophy with attenuation of the small blood vessels on the disc surface. The prognosis for recovery of vision is variable. Many patients recover good central vision, but are left with arcuate and sectorial visual field defects corresponding to loss of bundles of nerve fibres. In other patients, visual loss may be complete and permanent. Malignant arterial hypertension Some patients with malignant hypertension experience transient monocular blindness due to ischaemia of the optic nerve head. Associated headache, seizures, encephalopathy, renal impairment, high blood pressure and the characteristic ophthalmoscopic features of hypertension point to the diagnosis (Fig. 3.30). Papilloedema Patients with papilloedema (Fig. 3.46) from any cause may experience transient visual blurring or obscurations, with or without photopsias. The visual loss in chronic papilloedema is often postural, occurring as patients get up from a chair (or bend over), and may involve either eye alone, or both eyes together. The visual loss is typically ‘grey’ rather than ‘black’ and lasts for seconds rather than minutes. The explanation may be transient optic nerve ischaemia secondary to a relative decrease in orbital blood flow, as a result of raised cerebrospinal fluid pressure in the subarachnoid space around the optic nerve, and so increased pressure in the veins draining the optic nerve head. Episodes of visual blurring or blindness in

someone with papilloedema should lead to urgent investigation and appropriate action because permanent visual loss will eventually follow, gradually or suddenly. Optic neuritis and Uhthoff’s symptom Patients with acute and chronic optic nerve demyelination due to multiple sclerosis may experience transiently decreased vision in one or both eyes during exercise (Uhthoff’s symptom), or in association with other causes of increased temperature, emotional stress, increased illumination, eating, drinking, smoking and menstruation.185,186 The pathophysiology is unknown, although reversible conduction block in demyelinated nerve fibres secondary to an increase in body temperature or to changes in blood electrolyte levels or pH are believed to play a role. Ophthalmoscopy may be normal, but if the optic nerve head is inflamed, the optic disc will be swollen and will look similar to papilloedema. Optic disc anomalies Transient monocular visual obscurations are occasionally associated with an elevated optic disc without increased intracranial pressure. Examples include congenital anomalies of the disc, such as drusen or posterior staphyloma.

3.5.3 Orbital disorders Transient changes in the ocular media or intraocular pressure, such as vitreous floaters, vitreous haemorrhage, anterior chamber haemorrhage, lens subluxation, reversible cataract (in a diabetic) and glaucoma may cause transient monocular visual disturbance. Most of these conditions can be excluded by a competent ophthalmological examination. An intraorbital mass, such as an optic nerve sheath meningioma, may produce gaze-evoked transient monocular blindness; the blindness is limited to the duration of gaze in the affected direction (usually abduction of the affected eye) and the visual acuity usually returns to normal about 30·s after the eye moves back to the primary position. The loss of vision is possibly caused by a reduction in flow to the blood vessels surrounding the optic nerve itself.

3.5.4 Corneal and eyelid disorders Fuch’s corneal endothelial dysfunction (corneal guttatae) Fig. 3.46 An ocular fundus photograph showing papilloedema. Note the congested, swollen disc, with loss of the physiological cup, a blurred disc margin and congested retinal veins. Also reproduced in colour in the plate section.

In the seventh and eighth decade, the corneal endothelial cells may become deficient and malfunction due to abnormal excrescences of the basem*nt (Descemet’s)

107

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 108

108

Chapter 3 Is it a vascular event and where is the lesion?

membrane, called corneal guttatae. The corneal epithelium may then fail to pump fluid out of the cornea and patients complain of blurred vision (due to hydration of the cornea). It tends to occur in the morning and gradually clears as the day evolves, and as the cornea clears due to evaporation of tears. Similar symptoms can occur due to endothelial cell loss as a consequence of intraocular surgery. Ptosis Intermittent ptosis (e.g. myasthenia gravis) can cause transient monocular visual loss.

3.6 Improving the reliability of the clinical diagnosis

For every patient who presents to their family doctor with a definite transient ischaemic attack (TIA), there are many more who present with transient neurological symptoms due to other disorders. For example, in the Oxfordshire Community Stroke Project (OCSP), 512 patients were referred by their family doctor or a hospital doctor with a diagnosis of ‘possible TIA’, of whom 317 (62%) were considered by the OCSP neurologists not to have had a TIA (Table 3.22).6 This problem is not unique to family doctors and junior hospital doctors. The interobserver agreement for the diagnosis of TIA of the brain amongst eight senior and interested neurologists from the same department who interviewed 56 patients in alternating pairs revealed that both neurologists agreed that 36 patients had a TIA and 12 had not, but they disagreed about eight (kappa =·0.65; for perfect agreement kappa would be 1.0).187 Even experienced neurologists with an interest in cerebrovascular disease show considerable interobserver variability in the diagnosis of transient ischaemic attack. This does not imply lack of skill, but rather it is inherent in the clinical assessment of symptoms and signs. Sometimes the available information does not allow one to come to a ‘right answer’ as to whether the event was a transient ischaemic attack or not – in which case one ends up working on the basis of probability. The inter-observer agreement for the diagnosis of stroke (vs not stroke) is made with moderate to good reliability (kappa = 0.77).37

The clinical features associated with high interobserver agreement for the diagnosis of stroke or TIA versus no vascular event are a sudden change in speech, visual loss, diplopia, numbness or tingling, paralysis or weakness and nonorthostatic dizziness (kappa = 0.60).49 Several studies have shown that clinicians can differ in the interpretation of even isolated elements of the history, such as loss of power, loss of sensation, ‘blurred or foggy vision’, and headache.37,188 When assessing patients with suspected stroke, the interobserver agreement for most items of the clinical history is moderate to good (kappa statistics for vascular risk factors range from 0.44 to 0.69) but there is much less inter-observer agreement for various features of the neurological examination (Table 3.31).37,188,189

3.6.1 Reasons for clinical disagreement There are several factors that can increase the likelihood of clinical disagreement. For patients with suspected TIA, the main problem is eliciting and interpreting the history of an event that has resolved by the time the history is taken. For example, since the symptoms of most TIAs resolve within about 15–60·min, the diagnosis is almost always based entirely on the clinical history, which for a number of reasons may not be very clear: the patient may have forgotten the symptoms (because of poor memory or delay in presenting to medical attention); the symptoms may have been remembered, but are difficult to describe (e.g. transient hom*onymous hemianopia); or the patient may have been so frightened by the attack that he or she was more preoccupied with the immediate outcome than the exact nature of the deficit. All of these problems are more common in the elderly. For patients with suspected stroke, the inter-observer reliability of the clinical assessment is affected by a variety of factors, such as the time since the onset of symptoms; there is a trend for worse inter-observer reliability among patients assessed very early and very late after symptom onset.37 Another problem is that the generally accepted definitions of stroke and TIA lack specific detail about which ‘focal’ symptoms are not acceptable (e.g. ·isolated vertigo or not?). In the context of TIAs, what is an acceptable lower time limit for the duration of symptoms? Is a sudden focal neurological deficit, particularly a sensory deficit, of less than 5·s duration a TIA? Diagnostic conformity can only be achieved if precise criteria are available that are valid, reliable and generally accepted. In the particular case of TIA, it is difficult to determine the validity of any single diagnostic criterion, because there is no ‘gold standard’ against which to

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 109

3.6 Improving the reliability of the clinical diagnosis Table 3.31 Inter-observer agreement for neurological signs in stroke patients.

Kappa value Signs examined

Lindley et al.189

Shinar et al.188

Hand et al.37

Conscious level Confusion Dementia Weakness of arm Weakness of hand Weakness of leg Weakness of face Sensory loss on hand Sensory loss on arm Aphasia/language problem Dysarthria Visuospatial dysfunction Hemianopia Cerebellar signs/ataxia Cranial nerve palsy Extraocular movement disorder

0.60 0.21 NS 0.77 0.68 0.64 0.63 0.19 0.15 0.70 0.51 0.44 0.39 0.46 0.34 0.30

0.38 NS 0.34 NS 0.58 (R) 0.49 (L) NS 0.51 (R) 0.66 (L) 0.50 (R) 0.32 (L) NS 0.54 0.53 NS 0.40 0.45 NS 0.77

0.70 0.45 NS 0.65 0.72 0.57 0.50 0.49 0.49 0.66 0.41 0.41 0.46 NS NS NS

NS, not stated.

judge it, other than perhaps the prognosis. For example, patients with lone bilateral blindness were followed up and found to have a similar prognosis to patients with TIA, implying that lone bilateral blindness is also a TIA.48 Whether the prognosis of patients with isolated dysarthria or vertigo is similar is an important research question. More widespread and consistent consideration, discussion and application of the diagnostic criteria could enhance diagnostic accuracy and inter-observer agreement, and might improve patient management and care.188 However, as the diagnostic criteria become more specific, sensitivity is sacrificed and so an increasing number of genuine TIAs may be discarded and left untreated. Conversely, if the criteria become less specific, there may be a tendency to overdiagnose TIA, but this too can have adverse consequences such as the loss of a job, driver’s or pilot’s licence, money and self-esteem, as well as resulting in inappropriate investigation and the prescription of numerous unnecessary drugs. The inter-observer reliability of the clinical assessment of stroke is also improved by the experience and confidence of the assessor.37

3.6.2 Strategies to reduce inter-observer variation The clinical skills required to obtain an accurate and useful history are listed in Table 3.32, and strategies

Table 3.32 Clinical skills required to obtain an accurate and useful history. The ability to: Establish understanding Establish information Interview logically Listen Interrupt only when necessary Observe non-verbal cues Establish a good relationship Interpret the interview Tell the story in plain language Tell the story in chronological order Make the story ‘human’ Like this: This 85-year-old widow was standing at the kitchen table peeling potatoes at 7 p.m. on 26 November 2006, with her daughter, when she suddenly stopped talking, dropped the potato peeler she was holding in her right hand and fell to the floor. She was unable to get up and has not been able to speak or move her right arm or leg since. Not like this: This lady developed sudden dysphasia and right hemiparesis.

for preventing or minimizing clinical disagreement in Table 3.33. If these principles are applied with a knowledge of the diagnostic criteria (Table 3.5), then diagnostic inconsistency should be minimized. If there is still uncertainty about the diagnosis at the end of the

109

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 110

110

Chapter 3 Is it a vascular event and where is the lesion? Table 3.33 Strategies for preventing or minimizing clinical disagreement. Assess the patient in a suitable consulting environment, i.e. quiet room, minimal interruptions if possible Necessary equipment available: ophthalmoscope, sphygmomanometer, etc., telephone on desk to call witnesses (to clarify history) or colleagues (to obtain advice) Clarify and confirm key points Repeat key elements of the history or examination Corroborate important findings with witnesses, documents and, if necessary, appropriate tests Ask ‘blinded’ colleagues to see the patient also (e.g. at ward teaching sessions) Record evidence as well as inference, making a clear distinction between the two, reporting exactly what the patient said and then your interpretation (e.g. ‘the patient complained of heaviness of the right arm and leg’ and not ‘the patient complained of right-sided weakness’ or ‘the patient complained of right hemiparesis’) Apply the art and social sciences of medicine, as well as the biological sciences of medicine Make sure you have enough time for the entire consultation

history, then the general examination and special investigations aimed at detecting vascular diseases and risk factors may provide useful circ*mstantial evidence (Table 3.34).44,45,49,190 The odds of an event being due to cerebrovascular disease are likely to be substantially less if the patient is young and has no vascular risk factors, compared with an elderly patient who has several vascular risk factors, clinical evidence of established vascular disease (e.g. carotid or femoral bruits, absent peripheral pulses) or symptomatic vascular disease elsewhere (e.g. angina, intermittent claudication)190 – i.e. make sure that you use all the clinical evidence available to you. When coming to a view about whether an event was a transient ischaemic attack or stroke, make use of all the clinical evidence, both general and neurological, that is available after a detailed history and examination. If the history is elicited independently by a second physician, a subsequent comparison of the symptoms as well as their interpretation may lead to a greatly increased reliability of the diagnosis. Teaching hospitals in particular are in a privileged position to apply this powerful but expensive diagnostic ‘instrument’, but even so, it is still unrealistic except for the occasional very difficult case. What is often available to those in

Table 3.34 Clinical features (i.e. the milieu) influencing the probability that the event was a stroke or transient ischaemic attack. Very likely to be vascular, almost definite Atrial fibrillation and rheumatic heart disease Frequent carotid-distribution transient ischaemic attacks and focal, long, loud bruit over the carotid bifurcation on the symptomatic side History and physical signs suggestive of infective endocarditis (i.e. fever, splinter haemorrhages, cardiac murmur) Recent myocardial infarction (in last 3–4 weeks) Likely to be vascular, but less definite Atrial fibrillation and non-rheumatic valvular heart disease (but a few fibrillating stroke patients have primary intracerebral haemorrhage as the cause of the stroke) Arterial bruits anywhere (e.g. carotid, orbital, aortic, femoral) Prosthetic heart valve, taking anticoagulants (but some strokes are haemorrhagic, and some transient ischaemic attacks are related to coexistent carotid artery disease) Unlikely to be vascular (particularly if neurological symptoms are transient) Less than 40 years of age, no symptomatic vascular disease, no vascular risk factors, no family history of thrombosis or premature vascular disease, normal heart

hospitals, however, is an account by the referring physician who saw the patient at an earlier time after the onset of symptoms – be sure to read it carefully and go over any symptoms described that were not reported to yourself by the patient. Implicit within this is the responsibility for all clinicians who take a history to document it in the patient’s or witness’s own words and not simply record their own interpretation. The use of checklists written in simple language improves inter-observer reliability, and is likely to be useful for computer-aided diagnosis and further research studies.191 However, although checklists may encourage more thorough history taking, the symptoms still have to be interpreted correctly. Very short checklists have been used successfully by paramedical personnel for diagnosing cerebrovascular events hyper-acutely (section 3.3, table 3.6).39,40,42,43,49 The relatively poor inter-observer agreement for physical signs (Table 3.31) is not peculiar to patients with stroke. The advent of MRI technologies such as DWI, which is becoming more widespread and can show what were hitherto unseen and yet recent relevant ischaemic lesions, has also improved the sensitivity, specificity and reliability of the diagnosis of TIA and minor ischaemic stroke169 (section 3.2.1).

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 111

3.7 Is it a subarachnoid haemorrhage?

Subarachnoid haemorrhage (SAH) refers to the spontaneous extravasation of blood into the subarachnoid space when a blood vessel near the surface of the brain ruptures. It is a condition, not a disease, which has many causes (section 9.1).29,30 Although, as we will see (section 5.3.1), the clinical distinction between ischaemic stroke and intracerebral haemorrhage is unreliable and we have to rely on imaging, the clinical features of SAH are reasonably distinct; at least this type of stroke can be diagnosed clinically with reasonable confidence. However, confirmatory investigations are needed in almost all cases.

3.7.1 Clinical features Blood in the subarachnoid space is a meningeal irritant and incites a typical clinical response, regardless of the source. Patients usually complain of headache, photophobia, stiff neck and nausea, and they may also vomit. Confusion, restlessness and impaired consciousness are also frequent (Table 3.35). Precipitating factors More often than not, there is no obvious precipitating factor.192 Among the 33 patients with SAH in the OxfordTable 3.35 Diagnosis of subarachnoid haemorrhage. Principal symptoms headache – usually sudden, maximal in seconds, severe, and occipital or retro-orbital – duration: hours (possibly minutes, we do not know) to weeks nausea vomiting neck stiffness photophobia loss of consciousness Neurological signs none (very often) meningism (after several hours) focal neurological signs: third nerve palsy (mostly posterior communicating artery aneurysm), dysphasia, hemiparesis (arteriovenous malformation, intracerebral haemorrhage) subhyaloid haemorrhages in optic fundi fever raised blood pressure altered consciousness

shire Community Stroke Project, six (18%) occurred while resting (none occurred while asleep), 13 (39%) during moderate activity, and six (18%) during strenuous activity, such as weight-lifting and sexual intercourse; the activity at onset was not known in the other eight patients.193 In a more recent series, higher proportions (up to 50%) of aneurysmal SAHs occurred during physical exertion.194 Sexual activity may precipitate not only SAH,195 but also – and more often – relatively harmless headaches, migrainous or not (section 3.7.2).196 Headache Headache is the cardinal clinical feature of SAH. It is the only symptom in about one-third of patients,194,197 but a symptom at some stage in almost every patient.198,199 Conversely, in a prospective series of patients with sudden headache in general practice, SAH was the cause in one out of four patients, and only in one out of ten patients in whom sudden headache was the only symptom.197 One out of every four patients with sudden, severe headache has a ruptured cerebral aneurysm, or one out of ten if sudden headache is the only symptom. In SAH, the headache is generally diffuse and poorly localized but tends to spread over minutes to hours to the back of the head, neck and down the back as blood tracks into the spinal subarachnoid space. Sometimes the headache is maximal behind the eyes. The headache is often described by patients as the most severe headache they have ever had, but it can occasionally be milder. It is the suddenness of onset which is most characteristic. Speed of onset and disappearance The headache arises suddenly, classically in a split second, ‘like a blow on the head’ or ‘an explosion inside the head’, reaching a maximum within seconds. A potential pitfall is that patients may sometimes use the word ‘sudden’ to describe an episode of headache that came on over half an hour or longer, perhaps depending on the interval after which the history is taken. And even if the headache really comes on within seconds or minutes, this is not specific for ruptured aneurysms, or even for SAH in general. The reason is that in general practice, exceptional forms of common headaches outnumber common forms of a rare disease, in this case a ruptured aneurysm. This is one of the many examples of the risk paradox (section 18.5).200 Another, more striking example is that most children with Down

111

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 112

112

Chapter 3 Is it a vascular event and where is the lesion?

syndrome are born of mothers under 30 years old, despite older mothers being at higher risk. In the same way, most patients with ischaemic stroke are not severely hypertensive, etc. Given that the incidence of aneurysmal haemorrhage is about 10 per 100 000 population per year, a family physician with a practice of 2000 people will, on average, see only one such patient every five years. Patients with rare manifestations of common headache syndromes (i.e. sudden onset of migraine) probably outnumber those with common manifestations of rare headache syndromes (i.e. sudden headache in subarachnoid haemorrhage). Patients with headache may present not only to general practitioners but also to an accident and emergency department where they make up around 1% of all attendances. The proportion with a serious neurological condition ranges from 16% for all patients with any headache, to 75% for those with sudden headache specifically referred to a neurologist.201 The exact speed of onset of ‘sudden’ headache, seconds or minutes, in patients without any other deficits, is of little help for the hospital physician in distinguishing aneurysmal haemorrhage from innocuous headaches, or from non-aneurysmal perimesencephalic haemorrhage (section 9.1.2). The predictive value of the speed of onset (seconds vs 1–5 min) can be calculated from two data sets. First, SAH from ruptured aneurysms is nine times as common as non-aneurysmal perimesencephalic haemorrhage203 and in hospital series these two forms of SAH together are twice as common as innocuous headaches.202 Second, headache develops almost instantaneously in 50% of patients with aneurysmal SAH, in 35% of patients with non-aneurysmal perimesencephalic haemorrhage, and in 68% of patients with benign ‘thunderclap headaches’; for an onset within 1–5 min these proportions are 19%, 35% and 19%, respectively.194 If, for the sake of simplicity, we ignore patients with sudden headache from other serious (nonhaemorrhagic) brain disease, such as intracranial venous thrombosis,204 a simple calculation leads to the disappointing conclusion that an onset within seconds correctly predicts aneurysmal haemorrhage in only 55%, and that headache onset in 5–10 min correctly predicts innocuous headache in only 30%. The headache of SAH usually lasts 1–2 weeks, sometimes longer. We do not know exactly how short in duration a headache may be and still be due to SAH. However, we have not come across anyone with a headache due to proven SAH that resolved within 1 h. However, since there is no prospective study that has

addressed this question it is still conceivable that this may occur, so it is perhaps best to consider SAH in anyone with a sudden unusually severe headache that lasts longer than – shall we say – 1 h? In brief, there are no single or combined features of the headache that distinguish reliably, and at an early stage, between SAH and innocuous types of sudden headache. The discomfort and cost of referring most patients for a brief consultation in hospital (which should include CT scanning and a delayed lumbar puncture if the scan is negative) is probably outweighed by avoidance of the potential disaster of missing a ruptured aneurysm and the patient later being admitted with rebleeding, or another secondary complication. There is no feature of the headache that distinguishes reliably, and at an early stage, between subarachnoid haemorrhage and innocuous types of sudden headache. Therefore, although most people with a sudden severe headache have not had a subarachnoid haemorrhage, they must all be investigated to exclude this diagnosis. A biphasic headache may occur in patients whose SAH is due to dissection of a vertebral artery (section 9.1.3): first, a severe occipital headache radiating from the back of the neck, followed after an interval of hours or days by sudden exacerbation of the headache but of a more diffuse type. A history of one or more previous episodes of suddenonset headache (‘sentinel headaches’) is generally believed to be common in patients with aneurysmal SAH and these are often attributed to a ‘warning leak’ of the finally rupturing aneurysm. However, the notion of ‘minor leaks’ does not really hold up (section 9.2.4). It is doubtful if patients with aneurysmal subarachnoid haemorrhage often have preceding and unrecognized ‘warning leaks’. Whatever the case, doctors must be educated to consider subarachnoid haemorrhage in any patient who reports a sudden severe headache. Vomiting Vomiting (and nausea) is common at the outset, in contrast to other differential diagnoses, such as migraine, in which vomiting more often comes after the headache starts. Neck stiffness Meningism refers to painful resistance to passive or voluntary neck flexion, mainly because of irritation of

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 113

3.7 Is it a subarachnoid haemorrhage?

the cervical meninges by subarachnoid blood, or by inflammation. This sign can be elicited in the supine patient by placing both hands behind the patient’s head and, as one attempts to lift the head up off the pillow the patient does not allow the neck to be flexed and so the examiner lifts the patient’s head, neck and shoulders off the bed, as if the patient were like a board. In contrast, passive rotation of the neck is achieved with ease. Neck stiffness is a common symptom and sign of blood in the subarachnoid space, but it does not occur immediately; it takes some 3–12 hours to appear and may not develop at all in deeply unconscious patients, or in patients with minor SAH.205 Therefore, its absence cannot exclude the diagnosis of SAH in a patient with sudden headache. Brudzinski’s sign (flexion at the hip and knee in response to forward flexion of the neck) is also associated with blood in the subarachnoid space but its sensitivity and specificity in this context are unknown. Pain and stiffness in the back and legs may follow SAH after some hours or days because blood irritates the lumbosacral nerve roots. We know of one patient who was admitted during a weekend because of a sudden headache; on the following Monday, when seen by a new resident, he complained of pain in the back of his legs, but his notes could not be found (they were discovered only later, buried under a pile of X-rays). He was worked up for a back problem, but he suddenly died from what turned out to be rebleeding from a ruptured aneurysm. Photophobia Patients are often photophobic and irritable for several days after SAH. Loss of consciousness Loss of consciousness occurred in 50% of a large group of patients with presumed aneurysmal SAH who were well enough to be entered into a clinical trial of medical treatment.206 As these figures did not include the 10% or so of patients with SAH who die at home or during transportation to hospital,207 or the 20% who reach hospital and die within the first 24 h,208 it is likely that at least 60% of all patients with SAH lose consciousness at, or soon after, onset. The patient may regain alertness and orientation or may remain with various degrees of lethargy, confusion, agitation or obtundation. An acute confusional state can occur and be misinterpreted as psychological in origin (grimacing, spitting, making sucking or kissing sounds, spluttering, singing, whistling, yelling and screaming).209–211

The impaired consciousness may be caused by the large amount of blood in the subarachnoid space, the reduced cerebral blood flow caused by the sudden increase in cerebrospinal fluid pressure, or by a complication of SAH such as brain displacement by haematoma or hydrocephalus, or a fall in systemic blood pressure or arterial oxygen concentration. Epileptic seizures Epileptic seizures (partial or generalized) may occasionally occur at onset or subsequently, as a result of irritation or damage to the cerebral cortex by the subarachnoid and any intracerebral blood. In the Oxfordshire Community Stroke Project, two of the 33 patients (6%) with SAH had an epileptic seizure at onset, but neither had later seizures.97 Data from other series indicate that about 10% of patients with SAH develop epileptic seizures, most occurring on the first day of the SAH, but one-third not having their first seizure until 6 months later and one-third even more than a year later.212–215 The only independent predictors of epilepsy after SAH are a large amount of cisternal blood on brain CT, and rebleeding.213 Subhyaloid haemorrhage Intraocular haemorrhage develops in approximately 20% of patients with a ruptured aneurysm and may also complicate non-aneurysmal SAH, or intracranial haemorrhage in general. The haemorrhage is caused by a sustained increase in the pressure of the cerebrospinal fluid, with obstruction of the central retinal vein as it traverses the optic nerve sheath, in turn leading to congestion of the retinal veins.216 Mostly, the haemorrhages appear at the time of aneurysmal rupture, but exceptionally later without evidence of aneurysmal rebleeding. Linear streaks of blood or flame-shaped haemorrhages appear in the pre-retinal layer (subhyaloid), usually near the optic disc (Fig. 3.47); one-third lie at the periphery.217 If large, the pre-retinal haemorrhage may extend into the vitreous body (Terson syndrome; section 14.10.1). Patients may complain of large brown blobs obscuring their vision. Focal neurological signs Focal neurological signs may occur when an aneurysm has compressed a cranial nerve or has bled into the brain substance, causing an intracerebral haematoma (section 9.3.6). Sometimes, therefore, the clinical manifestations of a ruptured aneurysm may be indistinguishable from a stroke syndrome due to intracerebral haemorrhage or

113

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 114

114

Chapter 3 Is it a vascular event and where is the lesion?

returns to normal within a few days. The blood pressure changes probably serve to counteract the decrease in cerebral perfusion resulting from increased cerebrospinal fluid pressure and later ischaemia (section 14.3.1). Cardiac arrhythmias and ECG abnormalities are common after SAH (section 14.9.3).221 The mechanism is unexplained but is thought to be sustained sympathetic stimulation, perhaps caused by dysfunction of the insular cortex, which results in reversible structural neurogenic damage to the myocardium, such as contraction bands, focal myocardial necrosis and subendocardial ischaemia.

3.7.2 Differential diagnosis

Fig. 3.47 Ocular fundus of a patient with subhyaloid haemorrhage, appearing as sharply demarcated linear streaks of brick red-coloured blood or flame-shaped haemorrhage in the preretinal layer, adjacent to the optic disc and spreading out from the optic disc. Also reproduced in colour in the plate section.

cerebral infarction, particularly if little or no blood has entered the subarachnoid space. Systemic features Fever, hypertension, albuminuria, glycosuria and electrocardiographic (ECG) changes may be present in the acute phase. Pyrexia rarely exceeds 38.5°C during the first 2–3 days, but thereafter it may rise to over 39°C, presumably due to accumulation of breakdown products of blood in the subarachnoid space; the pulse rate is not raised concomitantly.218 The important distinguishing feature between pyrexia due to blood in the subarachnoid space and pyrexia due to intercurrent infection is the pulse rate; it remains disproportionately low in the former and rises with the latter. Hypertension is a well-established risk factor for SAH;219 a quarter to one-third of patients have a history of hypertension.206,220 On admission, about 50% of patients with aneurysmal SAH have a markedly raised blood pressure. Unfortunately, high blood pressure is found in many patients presenting to an accident and emergency department for any reason, which makes the predictive value of blood pressure readings in that situation abysmally low. In many patients this raised blood pressure is a reactive phenomenon, and not a marker of long-standing hypertension; often the blood pressure

The abrupt onset of a severe headache may not only be caused by subarachnoid haemorrhage (SAH), but also by several other conditions, such as meningitis or encephalitis, intracerebral haemorrhage (particularly posterior fossa haemorrhage), obstruction of the cerebral ventricles, intracranial venous thrombosis, a rapid rise in blood pressure and, finally, a variety of alarming but innocuous conditions.222 Associated features such as female sex, the presence of seizures, loss of consciousness or focal symptoms, vomiting, or exertion at onset increases the probability of aneurysmal SAH,194 but the predictive value of these characteristics is far from helpful. Only a sudden headache during sleep should suggest hypnic headache (see below). Acute painful neck conditions to be distinguished from meningism Meningism may be a feature of SAH, meningitis, a posterior fossa mass and cerebellar tonsillar coning. However, it characteristically disappears as coma deepens. Other causes of a painful or stiff neck include bony lesions (i.e. trauma or arthritis) and ligamentous strain in the neck, extrapyramidal rigidity, systemic infections such as pneumonia, cervical lymphadenitis, parotitis, tonsillitis and upper lobe pneumonia. However, it is usually quite easy to distinguish meningism from these other acute painful neck conditions. For example, pain arising from the cervical spine may not only be felt in the neck and back of the head but also in the shoulder and arm, it is often evoked or exacerbated by certain movements or positions of the neck other than flexion, and there is usually tenderness to palpation over segments of the cervical spine. Meningitis Meningitis is an acute febrile illness that usually presents subacutely over 1 or 2 days with generalized

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 115

3.7 Is it a subarachnoid haemorrhage?

headache, meningism, photophobia and fever. However, it can be difficult to distinguish from SAH if the patient is found confused or comatose, with marked neck stiffness and no available history. Clues to the diagnosis of meningitis include a high fever, tachycardia and a purpuric skin rash (meningococcal meningitis). If the patient is fully conscious and has no focal neurological signs and if meningitis is suspected, a lumbar puncture should be done immediately. But if the patient is very ill, antibiotics and steroids should be given at once, i.e. after venepuncture for blood cultures but before proceeding to brain CT and then, if no blood or intracranial mass is seen, a cerebrospinal fluid examination (section 3.7.3). Cerebellar stroke Cerebellar stroke often gives rise to sudden severe headache, particularly if it is haemorrhagic, and also to nausea and vomiting, but is usually accompanied by neurological symptoms and signs, such as vertigo, dysarthria and unsteadiness, which help distinguish it from SAH. However, if the lesion is large, the patient may present in coma due to direct brainstem compression or obstruction of cerebrospinal fluid flow from the fourth ventricle, causing hydrocephalus and raised intracranial pressure, or there may be meningism without signs of definite brainstem dysfunction. In a consecutive series of 100 patients with an initial diagnosis of subarachnoid haemorrhage, eight had a cerebellar haematoma (and another seven had supratentorial intracerebral haemorrhage).223 Urgent brain CT is required to confirm the diagnosis of cerebellar haematoma, and lumbar puncture should certainly not be done; indeed, lumbar puncture should almost always be preceded by brain CT in unconscious patients, even if there are no focal signs of a mass lesion and no clinical evidence of raised intracranial pressure (such as papilloedema). Intracerebral haemorrhage More than 50% of patients with spontaneous intracerebral haemorrhage have headache at onset, particularly those with superficial lobar haemorrhages, but the headache is generally not as strikingly sudden in onset as SAH.89 Furthermore, focal neurological deficits are almost always present, but they can also occur in about 20–30% of aneurysmal SAHs where there is intraparenchymal extension of the haemorrhage.224 Conversely, some intracerebral haemorrhages, particularly those which are deep, have less prominent focal neurological signs and can easily be mistaken for SAH. CT brain scanning is always required.

Intraventricular haemorrhage Intraventricular haemorrhage, which may arise within the ventricles or from immediately beneath the ependymal lining, or by extension from an intracerebral haemorrhage (i.e. from a caudate haemorrhage or ruptured subependymal vascular malformation into the intraventricular system), may mimic SAH (section 9.1.5). Also, intraventricular haemorrhage may occur together with SAH, usually from a ruptured aneurysm, most frequently at the anterior communicating artery complex. Patients present with sudden severe headache, confusion, vomiting or collapse with loss of consciousness.225 Again, brain CT or MR scan is required for diagnosis during life. Carotid or vertebral artery dissection Dissection of the wall of the internal carotid artery may cause a fairly distinctive headache syndrome, which is ipsilateral, involving the forehead, periorbital region, face, teeth or neck, and has a burning or throbbing quality (section 7.2.1.). The headache may be associated with an ipsilateral Horner’s syndrome or monocular blindness, and with contralateral focal neurological symptoms or signs. Dissection of the wall of the vertebral artery generally causes pain in the upper posterior neck and occiput, usually on one side, and may be associated with symptoms and signs of posterior circulation ischaemia, such as the lateral medullary syndrome. Transmural dissection of an intracranial artery can cause SAH (section 9.1.3). Acute obstructive hydrocephalus Any acute obstruction of the flow of cerebrospinal fluid causes headache through a rapid increase in intracranial pressure. The headache is commonly bilateral and exacerbated by coughing, sneezing, straining or head movement. Intermittent obstructive hydrocephalus may therefore cause severe paroxysmal headaches. A colloid cyst of the third ventricle is the classical cause of this syndrome; sometimes more than one family member is affected.226 A decreased level of consciousness may follow the headache.227 The outcome can be lethal if a diagnosis is not made.228 Brain CT or MRI usually identifies the offending lesion (Fig. 3.48). Migraine Migraine headache can sometimes arise suddenly (‘crash’ migraine), be severe and prostrating, unilateral or generalized, and associated with photophobia, irritability,

115

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 116

116

Chapter 3 Is it a vascular event and where is the lesion?

‘Exploding head’ syndrome Clusters of attacks characterized by a sensation of sudden noise in the head and terror, rather than pain, can strike individuals over the age of 50 years, particularly during the twilight of sleep.232,233 The cause is uncertain. Cough headache Idiopathic cough headache is defined as head pain brought on by coughing or other Valsalva manoeuvres, but not by prolonged physical exercise, in the absence of any intracranial disorder.234,235 It is probably caused by distension of venous structures in the brain. It is a sudden-onset headache and lasts from 1 s to 30 min, tends to be bilateral and posterior, does not begin earlier than the fifth decade of life, is more frequent in men, is not accompanied by other neurological manifestations and often responds to indomethacin. These clinical characteristics allow its differential diagnosis from posterior fossa lesions, especially herniation of the cerebellar tonsils (Chiari-I malformation), even though a craniocervical magnetic resonance imaging study is useful to rule this out. Fig. 3.48 CT brain scan showing a tumour in the third ventricle (arrow) at the foramen of Monro, causing obstructive hydrocephalus.

mild confusion, anorexia, mild fever, extraocular muscle palsy (ophthalmoplegic migraine) or symptoms of brainstem disturbance (basilar migraine) and thus be mistaken for SAH. However, migraineurs generally have a past or, less diagnostically helpful, a family history of migraine and the headache is commonly unilateral and throbbing, not so rapid in onset and of shorter duration than the headache of SAH.86 Vomiting tends to start well into the migraine attack, in contrast to SAH, in which it commonly occurs at or soon after onset of the headache.

Hypnic headache This type of headache has only been recognized recently.236,237 It wakes patients up in the middle of the night, 2–6 h after sleep onset; it is usually diffuse and bilateral, or in the neck. It occurs almost exclusively in middle age, twice as often in women as in men. Patients usually get out of bed and walk around until the pain subsides, within 10 min to 3 h. Typically the attacks are recurrent, from once a week to 6 times per night. There are no known precipitating factors, there is no photophobia, phonophobia or nausea and there are no concomitant autonomic features, such as in cluster headache or chronic paroxysmal hemicrania. Post-traumatic headache

Idiopathic stabbing headache Three specific varieties of sudden sharp stabbing headache have been described: ice-pick-like pains, ‘jabs and jolts syndrome’ and ophthalmodynia.229,230 The pains are mostly at the temples or orbits but on occasion are elsewhere in the head.231 Migraineurs are particularly susceptible.231 Precipitants – which are seldom present – may be postural change, physical exercise, or head motion. As these pains are transient and lancinating, they are unlikely to be confused with the headache of SAH. The mechanism is unknown.

Immediately after a head injury there is often headache; this may be pulsating and made worse by head movement, jolting, coughing, sneezing and straining.229 Normally the headache gradually disappears as the soft and any bony tissue damage resolves. In a series of 200 patients admitted to hospital with head injury only 83 still had a headache by a day or so afterwards; only 22 of those 83 (11% of all patients) complained spontaneously, and only three required an analgesic.238 The diagnosis of post-traumatic headache should not be confused with SAH if there is a history of head injury,

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 117

3.7 Is it a subarachnoid haemorrhage?

but the patient may be amnesic and there may be no witness, in which case acute head injury with secondary bleeding into the subarachnoid space can be confused with spontaneous SAH. If the circ*mstances of a traumatic head injury are unclear, and there is a reasonable chance that spontaneous intracranial haemorrhage was the cause of the accident and so the head injury, then brain CT should be done as soon as the patient’s condition allows, regardless of the severity of the head injury. Benign org*smic cephalalgia and benign exertional headache Acute, severe, explosive occipital or generalized headache, usually occurring at the moment of sexual org*sm or during strenuous exercise (benign org*smic cephalalgia and benign exertional headache, respectively) may mimic SAH.196,239,240 The history of onset during sexual intercourse (or masturbation) may not be forthcoming without specific and sensitive enquiry. Points in favour of the diagnosis are a history of similar previous sexual or exertional headaches, no alteration in consciousness, short duration of the headache (minutes to hours) and no signs of meningeal irritation such as neck stiffness, or low back pain, and nor sciatica in the ambulant patient. These headaches can occur at any time in life and do not necessarily occur every time the patient experiences org*sm, or exercises strenuously. If patients present soon after their first-ever sudden org*smic headache, it is not possible to exclude SAH without brain CT and lumbar puncture. If a patient presents after recurrent attacks, and the history is characteristic, investigation is seldom necessary. Reaction to monoamine oxidase inhibitor drugs People taking classic monoamine oxidase inhibitors (MAOIs), of which phenelzine and tranylcypromine are the most commonly used, may experience sudden severe headache after ingesting sympathomimetic agents, red wine or foods with a high tyramine content, such as mature cheese, pickled herrings, game and yeast extract. This is because MAOIs irreversibly block the ability of both MAO isoforms (A and B) to metabolize dietary tyramine in the liver (A) and gut wall (B). The combination of a classic MAOI and oral tyramine can provoke dangerous hypertension. The headache tends to be over the occipital region of the head and is associated with a rapid rise in blood pressure. It can be relieved by the alpha noradrenergic blocking agent, phentolamine.

Phaeochromocytoma Patients with a phaeochromocytoma experience acute pressor reactions; in about 80% of attacks they complain of headache241,242 often in combination with palpitations or sweating.243 The headache is usually of sudden onset, bilateral, severe and throbbing. It appears to be related to a rapid increase in blood pressure and lasts less than an hour in about 75% of patients, but it may last from a few minutes to a few hours. Some patients may collapse with loss of consciousness or develop focal neurological signs during the episode, as a result of cerebral oedema or sometimes haemorrhage. Attacks may be provoked by exertion, straining, emotional upset, worry or excitement.242 The diagnosis depends on clinical suspicion being aroused when the history is first taken (which can be difficult because the condition is so rare) and is confirmed by finding increased excretion of catecholamines (metanephrine and vanillylmandelic acid) in three 24-h specimens of urine, or raised plasma free metanephrine or normetanephrine levels during the attack.244 The blood sugar is usually raised at the time of the attack, a useful distinction from hypoglycaemic attacks, which may simulate phaeochromocytoma because of secondary release of adrenaline in response to low blood sugar. The tumour may arise at any point along the line of development of the sympathetic chain from the neck to the pelvis and scrotum. Headache may also reflect idiopathic surges of hypertension, not associated with phaechromocytoma or any other identifiable condition.245 Occipital neuralgia Occipital neuralgia is characterized by an aching or paroxysmal jabbing pain in the posterior neck and occipital region in the distribution of the greater or lesser occipital nerves (Fig. 3.49). It may rarely present quite dramatically, like SAH,246 but there is usually diminished sensation or dysaesthesiae of the affected area (C2 distribution), focal tenderness over the point where the greater occipital nerve trunk crosses the superior nuchal line, and a therapeutic response to infiltration of local anaesthetic near the tender area on the nerve trunk. Benign ‘thunderclap headache’ ‘Thunderclap headache’ is not a true disease entity, but a convenient term to describe unclassifiable varieties of sudden-onset, severe, generalized pain in the head, sometimes with vomiting.247 It may last up to a day or so. Clinically, the syndrome cannot be reliably distinguished

117

9781405127660_4_003.qxd 10/13/07 11:11 AM Page 118

118

Chapter 3 Is it a vascular event and where is the lesion?

Greater occipital nerve

Lesser occipital nerve

from SAH, but the chances of aneurysmal SAH are increased in the presence of female gender, epileptic seizures, loss of consciousness at onset, focal neurological symptoms (e.g. diplopia), vomiting or exertion preceding the onset of headache.194 The diagnosis is therefore made by exclusion, mainly of SAH. About 50% of these patients have a history of typical migraine or tension-type headache (with gradual onset). The prognosis is benign; a 3-year follow-up of 71 patients seen in hospital found identical recurrences in 12, again without evidence of SAH, whereas nearly 50% developed episodes of more obvious migraine or tension headache.248 Of 93 such patients identified in general practice and followed up for a median of 5 years, again none suffered SAH; recurrent attacks of ‘thunderclap headache’ occurred in eight patients, and 13 developed new tension headache or migraine.249 Headache is common in clinical practice, but ‘thunderclap’ headache is not. No physical sign can definitely exclude subarachnoid haemorrhage if a sudden onset headache persists for a few hours.

3.7.3 Investigations to confirm the diagnosis of subarachnoid haemorrhage Investigations are essential in making the diagnosis of subarachnoid haemorrhage (SAH), given the clinical features are relatively non-specific.250 Brain CT scan All patients presenting with a suspected recent SAH (i.e. within the last few days) should initially have an urgent brain CT to determine: • whether there is blood in the subarachnoid space;

Fig. 3.49 Diagram showing the anatomical distribution of the sensory innervation of the greater and lesser occipital nerves.

• the site of any SAH or intracerebral or intraventricular haemorrhage and therefore the likely cause (section 9.4.1 and Fig. 3.50); • the presence of any complications, such as hydrocephalus; • any contraindications to lumbar puncture such as cerebral oedema or haematoma with brain shift, or a large cerebellar infarct, when there is no blood evident on CT; • and whether there is any other intracranial abnormality that may account for the symptoms and signs. The sensitivity of CT in SAH depends on the amount of subarachnoid blood, the resolution of the scanner, the skills of the radiologist and the timing of the CT after symptom onset. The sensitivity is greatest in the first few days and falls thereafter, as blood in the subarachnoid space is resorbed (Fig. 9.19). In fact, the term ‘resorption’ may not always be appropriate to describe this process, diffusion and sedimentation being other explanations. CT evidence of subarachnoid blood can disappear very rapidly. If brain CT is done within 1–2 days after SAH onset, extravasated blood will be demonstrated in more than 95% of patients.202 But the chance of finding subarachnoid blood on brain CT then decreases sharply, to 50% on day 7, 20% on day 9, and almost nil after 10 days.251,252 If brain CT is done within 1–2 days after subarachnoid haemorrhage onset, extravasated blood will be demonstrated in more than 95% of patients. Of course, minute amounts of subarachnoid blood may be overlooked by the uninitiated (Fig. 3.51). Depending on the amount of blood in the cisterns and the delay before scanning, an ‘absent’ or ‘missing’ (isodense)

9781405127660_4_003.qxd 10/13/07 11:12 AM Page 119

3.7 Is it a subarachnoid haemorrhage?

Fig. 3.50 Brain CT scan (a) showing a 2- or 3-week-old intracerebral haemorrhage in the frontal lobes (arrow) of a patient who presented with an acute behavioural disorder and was misdiagnosed as being a hysterical alcoholic. Subsequent

catheter angiography (b) revealed an anterior communicating artery aneurysm (arrow) which had ruptured and bled into the frontal lobes.

cistern, or absent cortical sulci, may be the only clue to the presence of subarachnoid blood (Fig. 3.52). If subarachnoid haemorrhage is suspected and yet the brain CT scan appears normal, look carefully at the interpeduncular cistern, ambient cisterns, quadrigeminal cistern, the region of the anterior communicating artery and posterior inferior cerebellar artery, the posterior horns of the lateral ventricles and the cortical sulci. If blood is present in these sites, it may be isodense or slightly hyperdense, and hence the normally hypodense cisterns and sulci may be difficult to see and seem ‘absent’.

Fig. 3.51 CT brain scan showing a subtle amount of subarachnoid blood in the anterior part of the interhemispheric fissure (arrow), from a ruptured aneurysm of the anterior communicating artery.

Haemorrhage from an intracranial aneurysm can result not only in SAH but also in an intracerebral haemorrhage, which is easily seen on plain CT and generally persists for longer than subarachnoid blood, because the resorption of intraparenchymal blood, as seen on CT, occurs over several days to weeks rather than a few days.251 However, small intracerebral haemorrhages can also resolve very quickly, within days (section 5.4.1). About one-quarter of unselected patients who reach hospital alive after aneurysmal haemorrhage have an intracerebral haemorrhage on CT.224,251

119

9781405127660_4_003.qxd 10/13/07 11:12 AM Page 120

120

Chapter 3 Is it a vascular event and where is the lesion?

Fig. 3.52 Brain CT scan of a patient with subarachnoid haemorrhage showing isodense blood in the cortical sulci giving the appearance of ‘absent’ sulci. In particular, the sylvian fissures are not seen because they are filled with just enough blood to raise the density of the cerebrospinal fluid to that of brain parenchyma.

False positive evidence of SAH on CT scans A false positive diagnosis of SAH may be made on the CT brain scan, for example in patients who are comatose and brain dead (i.e. have no cerebral blood flow at the time of the scan). The CT scan not only shows cerebral oedema, but hyperdense material in the subarachnoid space which represents blood in congested subarachnoid blood vessels253–255(Fig. 3.53). Increased density of the tentorium and basal cisterns as a false positive sign of SAH on an unenhanced CT scan has also been described in polycythaemia,256 purulent meningitis,257 subdural haematoma,258,259 gliomatosis cerebri260 and spontaneous intracranial hypotension.261 Lumbar puncture According to a consecutive series from the Netherlands, the negative predictive value of CT scanning performed within 12 h of onset of sudden headache is 97%.202 In other words, there is an important small minority (of about 3%) with sudden headache and normal CT within 12 h who do have xanthochromia in the cerebrospinal fluid, and in whom angiography subsequently confirms a ruptured aneurysm. Therefore a lumbar puncture is necessary in any patient with sudden headache and a normal CT scan, even though the results will be normal in most patients.

Fig. 3.53 Brain CT scan of a patient who is brain dead from general hypoxia. There is generalized oedema, with effacement of the cisternal cerebrospinal fluid spaces. These spaces appear hyperdense, through venous stasis, which may falsely suggest subarachnoid haemorrhage.

Always do a lumbar puncture if the history is suggestive of subarachnoid haemorrhage and the CT scan (performed early, within a few days) is normal. Frequently, the cerebrospinal fluid will be normal too, but occasionally in this setting an abnormal cerebrospinal fluid will provide the only evidence of subarachnoid haemorrhage. Lumbar puncture without prior brain CT is potentially dangerous in patients with an intracerebral haemorrhage.262 Brain herniation may occur even in patients without focal signs or a decreased level of consciousness.263 Once the decision has been taken to do a lumbar puncture, the next requirement is to do it well. This is more difficult than it seems. Before drawing cerebrospinal fluid, the first rule is to wait until at least 6 and preferably 12 h have elapsed after the headache onset. This delay is absolutely essential because if cerebrospinal fluid obtained earlier turns out to be blood-stained, it is absolutely and irrevocably impossible to distinguish between

9781405127660_4_003.qxd 10/13/07 11:12 AM Page 121

3.7 Is it a subarachnoid haemorrhage?

blood that was there before (genuine SAH) and blood that was introduced by the needle (a bloody tap). With pre-existing blood in the cerebrospinal fluid, bilirubin will have been formed in the interval from SAH onset, from the breakdown of erythrocytes in the cerebrospinal fluid (see below), but of course not with a traumatic tap. A false positive diagnosis of SAH can be almost as damaging as a missed one, because insurance companies are bound to remain wary, despite negative investigations for an aneurysm. Never believe a colleague, however senior, who tells you that ‘the tap went so smoothly, it is impossible that the blood was traumatic’. Even the smoothest puncture can hit a vein. Also the ‘three-tube test’ (a decrease in red blood cells in consecutive tubes) is notoriously unreliable.264 Immediately proceeding with CT or MR angiography in all patients with bloodstained cerebrospinal fluid is not a good idea either, despite some people advocating this way of circumventing the ‘bloody tap’ problem: a small ( 1.5 cm diameter on CT/MR Tests At least one cardiac source of embolism should be identified (from high-risk group for ‘probable’ or medium-risk group for ‘possible’ cardiac embolism) Potential large artery atherosclerosis sources of thrombosis or embolism should be excluded Small artery disease Clinical Patient should have one of the traditional clinical lacunar syndromes and no evidence of cerebral cortical dysfunction, supported by history of diabetes or hypertension Imaging Normal CT/MRI or relevant brainstem or subcortical hemispheric lesion < 1.5 cm diameter Tests Potential large artery atherosclerosis sources of thrombosis or embolism, and cardiac sources of embolism should be excluded Other determined aetiologies Clinical Any symptoms compatible with an acute stroke Imaging CT/MRI findings of acute ischaemic stroke regardless of the size or location Tests Includes patients with non-atherosclerotic vasculopathies, hypercoaguable states or haematological disorders Potential large artery atherosclerosis sources of thrombosis or embolism, and cardiac sources of embolism should be excluded Undetermined aetiology Includes No aetiology determined despite extensive evaluation No aetiology determined but cursory evaluation Patients with two or more potential causes of stroke

As with the OCSP syndromes, a further problem is that of trying to classify the mechanism of stroke when patients are assessed in the hyperacute phase. In a study using the detailed TOAST protocol, the initial ( 1.2 for the hyperattenuated MCA : normal MCA, improved identification of true hyperattenuated MCA signs74 but needs to be confirmed in other studies. A hyperattenuated MCA mimic has been reported in herpes simplex encephalitis, where the low attenuation of the temporal lobe adjacent to the MCA made the MCA look whiter than it actually was.75 Increased haematocrit can also give the appearance of a hyperattenuated MCA, although usually bilateral, not unilateral.76 This sign was always related to occlusion of the MCA in patients who had angiography, giving a specificity of 100%.68 But the hyperattenuated MCA sign was present in only about one-third of patients in the Prolyse in Acute Cerebral Thromboembolism (PROACT) trial on CT scanning within 6 h of stroke – as all had MCA main stem or major branch occlusion on angiography within 6 h of stroke as a trial inclusion criterion, providing a sensitivity of around 33%.77 The hyperattenuated artery sign is more frequent in patients with more severe strokes who are scanned early, e.g. in 50% of 36 patients scanned within 4 h of MCA territory stroke;70 18% of 620 patients scanned within 6 h of stroke;78 and 41% of 620 patients scanned within 12 h of MCA territory stroke;68 but in only 5% of a more general population of acute stroke patients.79 In the first 500 patients randomized in the Third International Stroke Trial (IST3) of tissue plasminogen activator vs control, 41% had a hyperattenuated artery sign on the baseline scan within 6 h of stroke and this was associated with having a total anterior cerebral infarction, i.e. severe stroke. Of those with a hyperattenuated MCA sign on admission, the hyperattenuated artery was still visible on 24-h follow-up CT in 60% (personal data). Patients with the sign on baseline CT more frequently had more extensive tissue hypoattenuation and swelling, and a significantly worse ASPECTS score (5.6 vs 8.3). The hyperattenuated artery sign on CT is a reasonably reliable indicator of an occluded cerebral artery when the hyperattenuation is visible at a distance from the carotid siphon. An absent sign is certainly not a reliable indicator of a patent artery.

Evolution of the CT appearances of infarction Evolution of the CT appearances of infarction is illustrated in Fig. 5.16. Initially, the lesion has ill-defined margins and slight swelling and is slightly hypoattenuated (dark) compared with normal brain. The infarct becomes more clearly demarcated and hypoattenuated during the first few days.80,81 The swelling is usually maximal around the third to fifth days and gradually subsides during the second and third week.82 Infarct swelling can occur very rapidly – within the first 24 h – to cause brain herniation, but generally only with large infarcts. In general, swelling is most apparent in large infarcts but presumably also occurs in small ones, although it is difficult to see and is probably clinically less important. Extensive infarct swelling can compress adjacent normal brain and cause brain herniation (sections 11.5 and 12.1.5). It is an impression that younger patients are more prone to severe infarct swelling than older patients for a similar size of infarct, but perhaps older people – because of brain atrophy – simply have more space inside the cranial cavity to accommodate the swollen brain. The presence of recent haemorrhage in the infarct produces areas of increased density relative to both normal brain and the infarcted tissue, and further contributes to any swelling. The amount of infarct swelling, and the rate at which it appears, varies between patients for reasons which are still not well understood. Limited data from observational studies83 and thrombolysis trials84 indicate that swelling is associated with large infarcts and is more severe in patients whose occluded artery fails to recanalize (spontaneously or pharmacologically) (Fig. 5.20). This notion is supported by very limited data that suggest that treatment with thrombolysis up to 6 h after stroke may reduce the risk of massive infarct oedema (Fig. 5.21). During the second week, the infarct gradually increases in attenuation, sometimes becoming isointense and so indistinguishable from normal brain. This is the socalled ‘fogging effect’ and without close inspection even quite sizeable infarcts may be overlooked (Figs 5.16 and 5.22). Indeed, the ‘fogging effect’ may make the infarct impossible to see on CT scans performed at this time. This may lead to serious underestimation of infarct size. It does not occur in all infarcts and varies from 54% of patients scanned 10 days after onset85 to all patients at some time examined with six consecutive CT scans within 42 days of stroke.86 The ‘fogging effect’ may last up to 2 weeks and then the infarct becomes progressively more hypoattenuated (black). Eventually, a sharply demarcated, atrophic, hypoattenuated (similar to cerebrospinal fluid) defect remains (Fig. 5.16). Although old infarcts usually have sharply

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 201

5.4 CT scanning

Fig. 5.20 An unenhanced CT brain scan from a patient at 4 h after sudden right hemiparesis showing (a) a hyperattenuated left middle cerebral artery (MCA) (arrow) and (b) early hypoattenuation of the left basal ganglia insular and temporal cortex. Twenty-four hours later, (c) the left MCA is still hyperattenuated (arrow) and (d) there is massive swelling of the infarct with midline shift to the right, obstruction and dilatation of the right lateral ventricle.

(a)

(b)

(c)

(d)

Study or subcategory

OR (fixed) 95% Cl

ECASS 1995 NINDS 1995 ECASS II 1998 ATLANTIS B 1999

1.54 0.68 0.44 0.82

Total (95% Cl) Total events: 208 (Treatment), 239 (Control) Test for heterogeneity: Chi2 = 6.38, df = 3 (P = 0.09), F = 53.0% Test for overall effect: Z = 2.14 (P = 0.03) 0.1

0.2 0.5 Favours treatment

Fig. 5.21 Forest plot showing the limited data from the randomized controlled trials in acute ischaemic stroke of the effect of thrombolysis on major infarct swelling. The vertical line shows an odds ratio (OR) of 1.0 (no treatment effect), the horizontal lines represent the 95% confidence intervals around the point estimates of each trial, the boxes are the point estimates for each trial, their size being in proportion to the

5 2 Favours control

[0.79, 3.02] [0.49, 0.93] [0.19, 1.03] [0.07, 9.56]

0.74 [0.57, 0.98]

amount of information each contains, and the diamond represents the overall OR with its 95% confidence interval to the left of the vertical line of no effect, suggesting that thrombolysis reduces infarct swelling. Note the data for ATLANTIS B are only for the subgroup of patients randomized within 3 h of stroke onset (data for the whole trial are not available).

201

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 202

202

Chapter 5 What pathological type of stroke is it, cerebral ischaemia or haemorrhage?

(a)

(b)

Fig. 5.22 CT brain scans to demonstrate ‘fogging’. (a) Three days after onset of a right hemiparesis there is an obvious infarct (hypoattenuated area) in the left parietal cortex and adjacent white matter (thin arrows). (b) At 14 days after onset the recent infarct in the anterior left parietal region is almost invisible – it is now mainly iso-attenuated with normal brain and there is no mass effect – as a result of ‘fogging’.

(b)

Fig. 5.23 CT brain scan (a) without and (b) with intravenous contrast at 10 days after a right basal ganglia infarct in a 10year-old girl. There is marked serpiginous enhancement of the infarct (arrows) attributed to breakdown of the blood–brain barrier (sometimes also referred to previously as ‘luxury perfusion’).

demarcated borders, so making it possible to ‘age’ infarcts approximately, it is not always possible to tell with absolute certainty how old an infarct is, something that may be overlooked when ascribing particular clinical symptoms to lesions seen on CT. It is not always possible to tell the age of an infarct on CT, particularly in the case of small deep infarcts. Therefore, do not assume that a particular hypodense area on CT is necessarily relevant to a recent stroke – it could be old and irrelevant. Effect of intravenous contrast In the first week after stroke, intravenous X-ray contrast usually has little effect on the appearance of an infarct,

although some enhancement of the gyri may be seen.87,88 But, from the end of the first to the third week, more striking contrast enhancement occurs.53,89 The likely cause is a combination of blood–brain barrier breakdown, neovascularization and impaired autoregulation, and the resulting appearance on CT was previously referred to as ‘luxury perfusion’ (because we seldom use contrast at this stage after stroke, this feature is rarely seen).90 The enhancement can be very marked in children and young adults (Fig. 5.23). The tendency to enhance with contrast gradually resolves over the following few weeks. It was suggested some years ago that stroke patients might deteriorate as a result of intravenous contrast.91 It is certainly possible that extravasation of neurotoxic contrast agents could be harmful, but most patients

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 203

5.4 CT scanning

described had large infarcts with a poor prognosis in any case. In recent years, contrast has been used much more frequently in patients with acute ischaemic stroke, although usually in the first few hours, for perfusion CT, CT angiography (section 5.4.3) or during catheter angiography prior to and during intra-arterial thrombolysis. However, the contrast used now is ‘non-ionic’ and considered to be much less toxic (including neurotoxic) than contrast agents available in the 1970s and 1980s which were ‘ionic’. Contrast extravasation into subdermal tissues at the time of intravenous injection certainly causes a nasty local tissue reaction and so contrast extravasating locally into the brain through the ischaemic arteriolar endothelium might also cause a nasty local tissue reaction and worsen outcome, but we have no information on whether this is likely to be relevant in practice. Another potential problem is that all non-ionic contrast agents so far tested significantly slow both spontaneous and thrombolysis-induced thrombus lysis,92 prolonging the time to clot lysis threefold in a canine experimental model.93,94 With the increasing use of CT perfusion prior to intravenous thrombolysis, and of intra-arterial thrombolysis, this delay in lysis time may increase the risk of a poor outcome (in addition to the delay to starting treatment caused by the time taken to do the extra procedure) which may inadvertently be causing more harm than good. There is virtually no information on whether acute stroke patients given contrast (intravenously or intraarterially) have a worse prognosis than those not given contrast. Many of these patients have severe strokes and without a randomized trial comparing patients who received contrast with those who did not (all other factors being equal) it would be difficult to detect whether

contrast is harmful or not. Thus, until there is better evidence that the information provided by perfusion CT improves clinical outcome, or that intra-arterial thrombolysis is better than intravenous, the use of X-ray contrast in acute stroke should probably be avoided unless absolutely necessary. Intravenous contrast is rarely required to clarify the CT diagnosis in acute stroke, and should in general be avoided unless required for CT angiography or perfusion imaging. X-ray contrast agents interact with all thrombolytic drugs tested so far to significantly prolong the time to thrombus lysis, so should probably be avoided. Thus neither perfusion CT nor intra-arterial thrombolysis should be used routinely until there is good evidence to support this practice. How often does the appropriate infarct become visible on CT? In general, the more severe the stroke, the greater the volume of infarcted tissue, and so the more often the infarct is visible on CT scanning. Among 993 stroke patients scanned up to 99 days after stroke, 60% had a visible relevant infarct on CT within the first 24 h, rising to 70% by 72 h.95 A greater proportion of the total anterior circulation infarction (TACI) patients had a visible infarct than those with small cortical, lacunar and posterior circulation infracts, no matter how early or late they were scanned after the stroke. And among 13 000 patients randomized after CT scanning in the International Stroke Trial the highest proportion of patients with visible infarction was in the TACI group, the proportion increasing with time from onset to randomization in the first 48 h (Fig. 5.24).96 Small infarcts

Time (days) 0–3 4–6 7 – 12 13 – 24 25 – 48

Fig. 5.24 Data from the First International Stroke Trial showing the effect of time and stroke clinical syndrome on the visibility of infarcts on CT brain scanning. The visibility of increases with time over the first 2 days from the onset of the stroke, and the larger the infarct the more often it is visible (i.e. total anterior circulation infarcts, TACIs, show up more than lacunar infarcts, LACIs, etc.). PACI, partial anterior circulation infarct; POCI, posterior circulation infarct.

% Patients with visible infarction

70 60 50 40 30 20 10 0

TACI

PACI

POCI

LACI

203

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 204

204

Chapter 5 What pathological type of stroke is it, cerebral ischaemia or haemorrhage?

appear later than large ones, because there is less tissue to alter their density. Therefore, lacunar infarcts are less likely to show up in the first 24 h and sometimes do not do so at all (Fig. 5.24).95–100 Small infarcts in the brainstem and cerebellum are particularly difficult to visualize with CT because of artifacts arising from the petrous bones. In any series, the proportion of patients with a visible infarct also depends on local factors, such as case mix, time to first scan, whether or not the scan is repeated, generation of CT scanner, etc. Not surprisingly, therefore, the proportion of TACI patients with a visible infarct is lower in some studies than others.101–104 It is uncertain just how many patients with a clinically definite stroke never have an appropriate infarct visible on brain CT, but the proportion is probably quite high (up to 50%) depending on the timing of the scan, the age of the scanner, the thickness of the slices, the cooperation of the patient, the size and age of the infarct, the location (e.g. the brainstem is a difficult area to visualize infarcts on CT), the vigilance of the radiologist, and possibly some pathophysiological characteristic of the lesion itself. However, the main clinical reason for performing the scan is to exclude haemorrhage (or tumour or infection) as the cause of the symptoms and, if the scan is normal, the presumptive diagnosis is of an ischaemic event if the clinical picture is compatible with a stroke. If, in a particular patient’s case, it is absolutely critical to see positive evidence of an ischaemic stroke, then either the CT scan could be repeated at 3–4 days when the infarct is more likely to be visible, and certainly before 7–21 days when fogging is likely to obscure the infarct again. Brain CT only shows the appropriate infarct in about half of patients with an ischaemic stroke overall – those with symptoms of a more extensive stroke are more likely to have a visible relevant infarct than those with symptoms of a minor stroke. Therefore, absence of a visible infarct does not mean that the patient has not had a stroke; a patient with a clinical diagnosis of stroke, and an early CT scan which is normal, has still had an ischaemic stroke. How well do the clinical and CT diagnosis of the site of the stroke lesion correspond? The middle cerebral artery (MCA) territory is the most frequently affected on CT (60%), followed by the posterior cerebral artery (14%), the anterior cerebral artery (5%), the posterior fossa (5%) and the major artery territories combined or boundary zones (14%).105 There is reasonable agreement between the stroke syndrome

defined clinically using the Oxfordshire Community Stroke Project (OCSP) classification,106 with the site of the infarct as demonstrated by cross-sectional imaging (sections 4.3.2, 4.3.3, 4.3.4 and 4.3.5). Of course, in all the studies which assessed this agreement, a proportion of the patients had ‘normal’ imaging – that is, the imaging did not show an appropriate recent infarct – but it is unlikely that any imaging technique, no matter how sophisticated, will ever be able to demonstrate all infarcts. One relatively common discrepancy between the clinical and radiological diagnosis of the ischaemic stroke subtype is that about 20% of patients thought to have a mild cortical infarct clinically actually had a recent and lacunar infarct on imaging in a relevant site, and about 15% of patients thought to have a lacunar stroke clinically actually had a small and relevant cortical infarct radiologically.107,108 These patients behave epidemiologically more like the stroke type suggested by the imaging lesion than by the clinical syndrome, and so for them it is best to reclassify the stroke subtype from cortical to lacunar, and from lacunar to cortical on the basis of the imaging.107 The clinical diagnosis of lacunar stroke is particularly difficult in the first 12–24 h after onset.109,110 This implies that about one-fifth of all lacunar or mild cortical strokes are misclassified clinically and that studies of pathogenesis, risk factors and prognosis for lacunar or mild cortical infarction have probably been ‘clouded’ by inadvertent inclusion of some lacunar infarcts in the mild cortical group and vice versa. Because MR diffusionweighted imaging (DWI) is the most sensitive method of detecting infarcts underlying mild stroke syndromes (section 5.5.2), in future studies where it is necessary to distinguish lacunar from cortical ischaemic stroke this will be needed to establish the diagnosis with the best degree of reliability.111 In studies of stroke lesion site diagnosed clinically vs radiologically, which have used other more complex clinical classifications, the level of agreement between the clinician equipped with the traditional neurological tools of pin and tendon hammer, and the radiologist armed with a scanner, has been poorer.112 Even the National Institute of Neurological Diseases and Stroke (NINDS) classification,113 the Trial of ORG 10172 in Acute Stroke Therapy (TOAST) Trial classification,114 the Stroke Data Bank classification,115 all of which use risk-factor based classification and many investigations to apportion a cause of stroke, still find up to 40% of patients are unclassifiable. A further point to note is that classifications which use features other than simply the neurological symptoms and signs can introduce bias into studies of risk factors for stroke.116

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 205

5.4 CT scanning

Does infarct visibility on CT (early or established) have prognostic significance? Among 3468 patients from 15 studies examining the relationship between infarct signs present on CT scanning within the first 6 h after stroke and functional clinical outcome (assessed at least 1 month or more after stroke), the presence of any early infarct sign increased the risk of poor functional outcome on univariate analyses (odds ratio for any early infarct sign 3.11, 95% CI 2.77–3.49).59 This increased risk of poor functional outcome was present for all early infarct signs for which it was possible to extract data: any hypoattenuation vs no hypoattenuation (OR 2.73), more than 1/3 middle cerebral artery (MCA) territory affected vs less than 1/3 or normal (OR 6.44), less than 1/3 MCA vs normal CT (OR 2.66), and hyperattenuated artery vs no hyperattenuated artery (OR 2.09).59 Studies comparing the ASPECTS CT rating score and outcome indicate that an ASPECTS score < 7 is associated with poor outcome.61,117 There is less information on whether thrombolysis alters the relationship between CT infarct signs and poor outcome. Two studies which sought an ‘early infarct sign–thrombolysis’ interaction found no evidence that thrombolysis given up to 3 h118 or 6 h after stroke119 in the presence of early infarct signs worsened the outcome over that due to early signs alone. The ASPECTS score applied to the PROACT II trial data showed improved functional outcome for those treated with thrombolysis compared with control in patients with ASPECTS > 7 (OR 5) but not in those with ASPECTS of 7 or less (OR 1).120 When ASPECTS was applied to the NINDS, ECASS 2 and ATLANTIS B trials, there was also no association between an ASPECT score < 7 and worse outcome with thrombolysis,121–123 although ECASS 2 patients with a low ASPECTS were at increased risk of haemorrhagic transformation after thrombolysis.123 Early infarct signs are subtle. They are therefore difficult to recognize and may have poor inter-observer reliability59 (see below). Also CT technology has improved since the thrombolysis trials from which these data came (in the first half of the 1990s). Therefore, further data from current trials (e.g. the Third International Stroke Trial, www.ist3.com) are needed to determine whether thrombolysis does or does not modify the early infarct sign–poor prognosis relationship, and if so whether the treatment effect is different in patients with increasingly obvious early infarct signs. At present there is no evidence that patients with early CT signs of ischaemia are at increased risk of poor outcome if given thrombolysis.

Established visible infarction on CT has a clearer adverse independent relationship with prognosis than early infarct signs, it is easier to recognize, and there are more studies with larger sample sizes so the data are likely to be more robust95,96 (note the the odds ratios for early infarct signs are mostly from univariate analyses which may be why they appear larger than those for established infarct signs). In the International Stroke Trial (IST), visible infarction on the CT scan within 48 h of the stroke onset was independently associated with increased death within 14 days (odds ratio (OR) 1.17), and of death or dependency at 6 months (OR 1.42), an absolute increase of 13%, or 130 per 1000 more dead or dependent patients with visible infarction than without it.96 In our hospital stroke registry study, of 993 patients scanned mostly within the first week of stroke, the presence of a visible infarct was associated with an increased risk of poor functional outcome (OR 2.5; 95% CI 1.9–3.3) and death (OR 4.5; 95% CI 2.7–7.5) at 6 months, even after adjusting for time from stroke to scan, and important clinical prognostic variables.95 Others have also found that established visible infarction conferred an adverse prognosis.124–126 Therefore, it seems that a visible established infarct indicates a profound depth of ischaemia, which reflects a marked drop in blood flow to a largish area of brain, which suggests that the outcome will be poor. Visible infarction on CT, whether within the first 6 h or several days after the stroke, is independently associated with a poor functional outcome compared with no visible infarction. For patients scanned within the first 6 h, there is as yet no definite evidence that thrombolysis of those with early visible infarct signs either worsens their prognosis or is a clear contraindication to treatment. Inter-observer reliability in the analysis of CT scans of stroke patients very early after the stroke, and in established infarction Our systematic review of the inter-observer reliability of early infarct signs on CT scans up to 6 h after stroke (complete up to mid-2003) found 15 studies in which a median of 30 CT scans were rated by six observers.59 Few studies gave any information on the demographics of the patients from whom the scans came. Between one and five early CT signs were analysed including: hypoattenuation, swelling, hyperattenuated middle cerebral artery sign, loss of grey and white matter definition, loss of basal ganglia outline, loss of the insular ribbon, infarction of more or less than one-third of the middle cerebral artery (MCA) territory, and cortical sulcal effacement. The reference standard was mostly a neuroradiologist

205

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 206

206

Chapter 5 What pathological type of stroke is it, cerebral ischaemia or haemorrhage?

with access to a follow-up CT scan. In the 15 studies, the mean prevalence of any early infarct sign was 61%. For any early CT sign, the inter-observer agreement ranged from 0.14 to 0.78 (kappa). The mean sensitivity and specificity (and range) for infarct sign detection were 66% (20–87%) and 87% (56–100%) respectively. Inter-observer agreement was worst for hypoattenuation and best for the hyperattenuated artery sign. The one study of inter-observer reliability of the ASPECTS scoring system available at the time of the study, which suggested very good inter-rater reliability of the scoring system, has since been replicated in other datasets with similar results.127 The information, where given, suggested that experience improved detection, but knowledge of symptoms did not.59 These rather limited data suggest that some early infarct signs are detected better than others (e.g. mass effect better than hypoattenuation) so there may be ways in which the inter-observer reliability for early infarct signs could be improved by focusing the observer’s attention on specific features. Finding out specifically what makes more experienced observers more reliable would also help. Our ACCESS Study (Acute Cerebral CT Scanning for Stroke, www.neuroimage.co.uk) was a very large interobserver reliability study of early infarct signs on CT to see whether some signs were more reliably detected than others, whether scoring systems improved inter-rater reliability, whether ‘distracters’ on the scan (e.g. old infarcts) affected relability, and whether observer experience had any effect. We used a novel web-based method of showing digital scans to observers, recording some basic demographic information about the observers, and then their interpretation of each scan on a structured questionnaire via the same web page. The results to date, on 207 raters from 36 countries who read all 63 scans, and taking neuroradiologists as the reference standard, indicate that: • hypoattenuation and swelling are recognized to an equal degree; • hyperattenuated arteries are seen better than parenchymal hypoattenuation or swelling; • any scoring system (whether 1/3 MCA,63 ASPECTS61 or our own method62) improves inter-rater reliability; • distracters (e.g. old infarcts) worsen inter-rater reliability, particularly among less-experienced individuals. On CT scans up to 6 h after stroke, signs of hypoattenuation and swelling are detected equally reliably, using any scoring system helps improve interrater reliability for detection of early infarct signs, and any distracter (e.g. old infarcts, leukoaraiosis and atrophy) reduces early infarct detection. The emphasis on detecting subtle early signs of ischaemia may have deflected attention away from

looking for other important causes of the stroke which themselves may have only subtle CT signs. Worryingly, in a study of observer reliability of early infarct and haemorrhage detection on CT, only 17% of admitting physicians and 40% of neurologists were able to recognize intracranial haemorrhage with complete reliability.49 In established infarction, high levels of agreement (substantial to perfect) have been demonstrated.128,129 However, where observers had free access to relevant clinical details, their interpretation of the scans may have been biased,129 although there is little information about what effect this has on scan interpretation. When neurologists and radiologists reviewed the same two CT brain scans from patients with a lacunar stroke (camouflaged by an assortment of other scans), accompanied by misleading clinical information, the diagnosis of lacunar infarction did not appear to be biased by informing the observer that the patient was thought clinically to have had a stroke.100 This lack of bias, even with knowledge of the clinical details, may have been because the study was small and may not have been a true reflection of the difficulties encountered in routine practice when faced with a CT scan showing multiple ‘holes in the brain’ and generalized atrophy. The former makes the diagnosis of recent lacunar infarction difficult, because it is impossible to decide which ‘hole’ is the relevant one unless serial scans show a new ‘hole’ developing, and the latter makes the diagnosis of a small cortical infarct difficult (when is a large sulcus actually an infarct?). In established infarction, the agreement as to the site of infarction, the amount of swelling, and haemorrhagic transformation was excellent between experienced neuroradiologists and good between trainee radiologists when using a simple stroke CT classification system.62 Online training in CT intepretation As, in the interests of speed, most stroke CT scans in the very acute phase are likely to be read by the admitting neurologist or stroke physician (who are both probably less experienced than a dedicated neuroradiologist), methods of improving interpretation of CT scans from patients presenting with acute stroke are required. At the time of writing, there are two sources of scan examples for trainees available over the web, and more in development. At www.neuroimage.co.uk the reader can register to read CT scans from acute stroke patients over the web in their own time, and then obtain feedback comparing their performance with experts and observers from other specialties, as well as seeing what the follow-up scan showed. At www.ist3.com click on ‘see BASP CT Training Series’ for a demonstration of a variety of conditions presenting as acute stroke.

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 207

5.4 CT scanning

Appearances of the infarct: can we infer anything about its cause? A typical established large artery infarct is wedge-shaped, of decreased attenuation compared with normal brain, sharply demarcated, and occupies a recognizable vascular territory (Fig. 6.20c).105,130 It is usually due to embolism from a proximal arterial source or the heart. Lacunar infarcts, thought to arise from occlusion of a single perforating artery (section 6.4), are arbitrarily less than 1.5 or 2 cm in diameter (section 4.3.2), usually rounded in shape, and sited in the deep white matter, basal ganglia and pons (Fig. 6.20d).97,98 Boundary zone infarcts lie in areas of brain at the edge of the large artery vascular territories, i.e. in the parietooccipital region for the middle–posterior cerebral arteries and over the vertex for the anterior–middle cerebral arteries boundary zone, or in the internal boundary zone in the centrum semiovale at the junction of the deep and superficial arterial supply areas (section 4.2.4).130–132 However, the boundary zone areas are potentially quite extensive, and vary between and within individuals (section 4.2.4 and Fig. 6.21),133 so that in any one individual patient it can be very difficult to decide on the

100

basis of brain imaging whether an infarct has arisen from occlusion of a cortical branch of the middle cerebral artery (MCA), which might be embolic, or from poor perfusion in the boundary zone as a result of established internal carotid artery occlusion (section 6.7.5).132,134 Striatocapsular infarcts are larger than lacunes and occur in the deep white matter and basal ganglia, with preservation of the overlying cortex (section 4.2.2). They are thought to arise from transient occlusion of the MCA mainstem, prolonged occlusion of the MCA mainstem with good cortical collaterals, or occlusion of multiple lenticulostriate artery origins from atheroma of the MCA.135–137 This pattern is illustrated in Fig. 6.20e. In our hospital-based stroke registry, we attempted to assign a cause of stroke in all patients admitted during one sample year.138 Among 479 patients, those with large cortical infarcts had the greatest proportion of arterial (59%) or cardiac (29%) embolic sources, or both (15%), followed by small cortical infarcts (45%, 18% and 5%, respectively); lacunar infarcts (33%, 8% and 4%, respectively); and posterior circulation infarcts (32%, 9% and 4%, respectively) (Fig. 5.25). Thus, while patients with a cardiac source of embolism are at greater risk of having

n = 34

n = 100

n = 72

n = 56

n = 119

n = 32

n = 20

Total anterior circulation infarction

Partial anterior circulation infarction

Lacunar infarction

Posterior circulation infarction

Cerebral transient ischaemic attack

Amaurosis fugax

Retinal infarction

% of patients

Fig. 5.25 The frequency of various sources of emboli in patients with different ischaemic stroke subtypes, ocular and hemispheric transient ischaemic attacks (prepared by Dr Stephanie Lewis, Edinburgh).138

Source of embolus Cardiac Both cardiac and arterial Arterial Neither

207

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 208

208

Chapter 5 What pathological type of stroke is it, cerebral ischaemia or haemorrhage?

a large cortical than any other type of infarct (if they have a stroke), 74% of the patients who do have a large cortical infarct have an arterial source of embolism. Patients with several recent infarcts in more than one vascular territory are more likely to have a proximal embolic source (e.g. myocardial mural thrombus, aortic atheroma), but in other circ*mstances, the use of a complex classification or multiple imaging modalities does not solve the problem of deciding what caused the stroke in patients with more than one possible cause. It is not valid to assume that a particular size, shape and position of an infarct on imaging was definitely caused by a particular embolic source or in situ thrombosis (and of course finding a potential source of embolism does not necessarily mean that it was the cause of the stroke). In up to 15% of patients (probably a greater proportion the harder one looks) there is more than one potential cause of the stroke. As indicated above, about one-fifth of patients presenting with a cortical stroke syndrome (i.e. partial anterior circulation infarction) are found on brain imaging to have a recent lacunar infarct in the correct hemisphere without any sign of the expected cortical infarct. Equally, between one-sixth and one-fifth of patients with a lacunar infarct syndrome may have a recent cortical infarct on brain imaging with no hint of any subcortical lesion. This suggests that the syndrome (and so anatomical location of the brain lesion) attributed to an ischaemic stroke following clinical examination could and indeed should be modified by the position of any recent, and likely to be relevant, infarct on brain imaging, as this better reflects the underlying vascular lesion and so improves management and resource use.

(a)

(b)

The anatomical location attributed to an ischaemic stroke following clinical examination may need to be modified if brain imaging shows a recent and probably relevant infarct in a different territory to that expected clinically. For example, a patient with a clinical lacunar syndrome, but whose CT scan shows a recent cortical infarct in the relevant hemisphere, should be regarded as being similar to a patient with a cortical syndrome (i.e. at high risk of early recurrent stroke, high probability of ipsilateral carotid stenosis or cardiac source of embolism).

Distinction of arterial infarcts from other conditions on CT While many infarcts as a result of arterial occlusion are easy to diagnose on imaging from their site, shape, density and appropriate clinical features, other lesions occasionally produce very similar appearances, which can be confusing. Viral encephalitis, if relatively focal in distribution, can appear exactly like an infarct, although this is rare. The typical appearance of herpes simplex encephalitis with involvement of the medial temporal lobes should not cause confusion, but we have seen patients with lowdensity areas in the temporoparietal region associated with rising viral titres, which resolved following treatment with aciclovir (Fig. 5.26). Encephalitis can also give the impression of a hyperattenuated middle cerebral artery,75 possibly because the hypoattenuated temporal cortex (due to encephalitis) makes the adjacent middle cerebral artery look more hyperattenuated than it really is. As ever, it is vital that the imaging is reviewed with all the clinical information and, if there is any doubt (e.g. fever, subacute onset), other diagnostic tests must be

Fig. 5.26 Unenhanced CT brain scan (a) in a 70-year-old woman with a right hemiparesis, confusion and drowsiness for 6 h, showing an extensive area of low attenuation in the left temporoparietal region including the basal ganglia. There is loss of the outline of the normal basal ganglia (thick arrows) and loss of the overlying cortical sulci indicating slight swelling (compare with the easily seen right parietal sulci; thin arrows). The initial clinical diagnosis was of a stroke (infarct), but 1 week later (b) the patient deteriorated and the CT scan showed extensive haemorrhage in the left temporoparietal region. Subsequent postmortem confirmed the diagnosis of herpes simplex encephalitis.

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 209

5.4 CT scanning

Fig. 5.27 CT brain scan showing a tumour which initially looked like an infarct in a 60-year-old man who had presented with ‘a stroke’. (a) The unenhanced scan shows a left temporal low attenuation area (the ‘infarct’) with a high attenuation dot in the sylvian fissure (the ‘hyperattenuated artery’; arrow). This was misinterpreted as an acute left middle cerebral artery infarct. (b) Several weeks later a CT scan with intravenous contrast shows that the lesion had grown, with enhancement, consistent with a tumour. It was a secondary deposit from a colon primary.

(a)

carried out, including an electroencephalogram, MR scan and cerebrospinal fluid examination. Purulent cerebritis can look like an infarct, although the lesion is usually not wedge-shaped, it involves more white matter than cortex, and the clinical picture should allow the distinction to be made. Tumours: Occasionally, a peripheral metastasis with a large amount of white matter oedema can mimic an infarct on CT, although the density is usually lower than expected for an infarct, and administration of contrast may show up a cortical nodule (Fig. 5.27). If there is still doubt, either a repeat CT a few weeks later (because infarcts and tumours evolve differently, Fig. 5.4) or an MR examination (pinpoint metastases or a central tumour nodule may be detected) will usually determine the cause of the abnormality. Venous infarction is not uncommonly mistaken for an arterial infarct. The difference between them is discussed in section 5.8. Time is a useful diagnostic tool, ‘a chronogram’. If in doubt whether a lesion is a tumour or infarct on CT, repeat the scan in a few weeks. Infarcts get smaller (usually) due to ex vacuo effect, whereas tumours stay the same or get bigger.

5.4.3 CT perfusion imaging and CT angiography Cerebral perfusion imaging with CT is a technique which has come to prominence in the last few years as the technology has improved, enabling rapid ‘cine’ scanning. At the time of writing, spiral and multislice scanners were able to provide sufficiently rapid data acquisition at four slices in the brain to enable the data to be processed to produce cerebral blood flow, cerebral blood volume and mean transit time maps.139 This technique requires an intravenous bolus of about 50 mL of iodinated contrast

(b)

by pump injector and so exposes the patient to the lysis-retarding effects of the contrast agent described in section 5.4.2. The raw data then need to be processed, sometimes on a separate workstation, to produce the actual perfusion maps. This may take several minutes. CT perfusion imaging is relatively more accessible than MR perfusion imaging, and can be performed during the same examination as the plain CT brain scan. Hence it provides limited data on cerebral perfusion to many who do not have access to MR perfusion imaging or other techniques.139,140 There is considerable interest in whether any lesion seen on the different forms of CT perfusion imaging indicates ‘tissue at risk of infarction’ and so helps guide thrombolytic treatment. However, so far, relatively few observational case series of patients undergoing CT perfusion have been published. All suggest that CT perfusion may improve outcome prediction over plain CT alone,141 identify penumbral tissue142,143 or thresholds of tissue viability,144 but all used very early time points to assess ‘final infarct extent’ (2–7 days) and included a mixture of some thrombolysis-treated and some not thrombolysis treated patients. Larger studies141–143 suggest that relative mean transit time (rMTT) maps predict final infarct extent in the absence of recanalization and that cerebral blood volume maps predict final infarct extent if there is early recanalization (i.e. infarct core),143 although a combination of CBF and CBV may have highest sensitivity and specificity for predicting infarct core.142 These results are preliminary and have yet to be tested in a prospective trial where patients are randomly allocated to thrombolysis vs control. Thus, more work is required to determine whether or not CT perfusion identifies a subset of patients who are more likely to benefit from thrombolysis than can be identified from plain CT alone, and if so which perfusion lesion is the best predictor, or whether the additional time taken

209

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 210

210

Chapter 5 What pathological type of stroke is it, cerebral ischaemia or haemorrhage?

to perform the CT perfusion and the thrombolysisretarding properties of CT contrast agents negate the beneficial effects of the additional information gained. CT angiography (CTA) has also become more practical and accurate with improvements in CT technology. It requires rapid scanning with a spiral or multislice scanner through the skull base and circle of Willis during an intravenous injection of about 50 mL of iodinated contrast, followed by reconstruction of the data into axial, coronal, sagittal or 3D ‘angiographic’ images of the arteries. This technique can identify major intracranial and branch artery occlusions, and so might improve selection of patients for thrombolysis if, for example, only patients with a visible arterial occlusion were found to benefit.145 However, as with CT perfusion, it takes extra time to acquire and process the CT angiographic images, the contrast agent may delay thrombolysis, and it is unclear whether the additional information really does improve patient selection because there have been no randomized studies of thrombolysis vs control based on CTA. Indeed, the available evidence (n = 151 patients) suggests that CTA probably does not improve outcome prediction beyond that of clinical stroke severity alone and therefore may not aid decision making about thrombolysis treatment either.146 It is useful to note that the NIHSS stroke severity score is closely related to the probability of having an arterial occlusion on angiography within the first 6 h of stroke – patients with proximal intracranial arterial occlusion have worse NIHSS scores.147 Therefore, if it turns out that only patients with a large intracranial artery occlusion should be treated with thrombolysis, it may make sense to use the NIHSS score within 6 h of stroke as a surrogate for angiography-detected intracranial arterial occlusion. This would enable faster treatment delivery, avoid the cost of angiography (whether MRA, CTA or catheter) and the risk of catheter angiography, and minimize exposure to X-ray contrast agents. Until further reliable data are available, CTA is a useful diagnostic tool in occasional patients (but not for routine stroke investigation) but further research is needed to sort out its clinical application.

5.5 Magnetic resonance imaging

Historical note and practical points regarding use of MR Magnetic resonance imaging (MRI) was first used as a clinical tool in the early 1980s. The equipment is more

expensive than for CT scanning, both to purchase and to run, and MR was initially little used in acute stroke. It is not a practical technique for many acutely ill patients for the following reasons. • The patient must be placed inside a tube-like structure, which creates difficulties for being able to see and communicate with the patient, and for access for monitoring and administration of anaesthetics. • The patient must lie still, usually for at least 5 min at a time, although recent fast-scanning techniques mean that diagnostic images may be acquired in seconds. • Because of the need to perform several sequences, the whole scan takes longer than a CT brain scan (i.e. longer to obtain a diagnosis), even with ultrafast scanners which are now becoming available.148 • Fifteen to twenty per cent of patients with acute stroke may not be suitable for MR on medical grounds because they are too ill, vomiting or unable to protect their airway, or because of a contraindication to MRI.149–151 • Many stroke patients are confused, restless and frightened by the noise and vibration of the scanner, and so move during the scan, which impairs the images. • It is inadvisable for patients who cannot protect their airway to be placed supine for any length of time after stroke,152 regardless of the value of the information so obtained. • About 20% of patients with acute stroke become hypoxic while in the MR scanner.150 • In the only study which actually recorded it, about one-third of acute stroke patients required some form of physical intervention while in the scanner (reassurance, assistance while vomiting, administering oxygen, etc.), and 20% that started scanning did not complete all the sequences.150 In view of these constraints, patients with mild stroke (lacunar or small cortical) are more able to cooperate with scanning and a much larger proportion can be scanned (>95%).153 Given all these very real practical issues with MR, and until it is clear that the information gained from MR is so much better than from CT that it is worth the extra time, cost, risk to the patient from hypoxia, etc. and organizational difficulties, the first-line investigation for most stroke patients will probably continue to be CT in the immediate future,154,155 reserving MR for ‘difficult cases’ with specific questions (see below), and for research.156 CT has enjoyed somewhat of a renaissance in the last few years with the advent of widely available multislice perfusion imaging140 and rapid CT angiography with very good image reconstruction.145 However, use of MR scanners may expand as their design evolves, the sequences become faster148,157 and access to patients within the scanner becomes easier (e.g. wider bore), and

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 211

5.5 Magnetic resonance imaging

therefore more user-friendly. The advantages of MR are not only its superior pathoanatomical cross-sectional imaging properties, but that the same machine can image vessels and assess cerebral blood flow non-invasively; provide diffusion imaging and spectroscopy to elucidate the pathogenesis of brain damage in ischaemia and how experimental treatments might modify it; and, by functional imaging, elucidate mechanisms of brain recovery. Many of the general principles of recognizing an infarct by its shape and site, the correlations between clinical findings and infarct site, the time course of infarct swelling, fogging and haemorrhagic transformation as have been described for CT are equally relevant to MR and will not be repeated here. Rather, this section highlights where MR imaging provides different information to CT. Routine MR brain imaging for suspected stroke A routine MR brain scan for stroke should include a midline sagittal localizing view of the brain, usually T1-weighted, followed by axial diffusion-weighted images (DWI), a gradient echo (or T2*) sequence, axial T2-weighted images and/or a fluid attenuated inversion recovery (FLAIR) sequence – all covering the whole brain (Fig. 5.28). MR dynamic susceptibility perfusion imaging following an intravenous injection of a bolus of contrast, MR angiography and MR spectroscopy are also commonly available on modern MR scanners. In simple terms, it is useful to think of: • DWI showing as bright white (hyperintense) areas where water movement is restricted (i.e. intracellular oedema); • FLAIR and T2-weighted images as showing brain water content (so areas of oedema will show up as hyperintense, i.e. white, with CSF also white on T2 and dark on FLAIR; • gradient echo as being particularly sensitive to anything with paramagnetic effects, such as haemoglobin breakdown products, therefore being very good for showing new and old haemorrhage; • T1-weighted images as showing brain structure. Most scanners now use a ‘fast spin echo’ sequence for T2 which takes less time to acquire than the older ‘spin echo’ sequences. However, it is important to recognize that spin echo and fast spin echo do not produce the same images, the main difference being that fast spin echo is less sensitive to haemosiderin, a haemoglobin breakdown product.158 A routine MR examination for stroke should include DWI, gradient echo and either T2 or FLAIR sequences.

Contraindications to MR All patients should be screened for MR compatibility before they enter the magnetic field, but the key contraindications are mentioned here to help avoid inappropriate referrals: pacemakers, intracranial aneurysm clips (virtually all types), definite metallic intraocular foreign bodies, and cochlear implants (see www.mrisafety.com for details on specific manufacturers’ devices and MRI compatibility). Intracranial aneurysm coils are not contraindications. Metallic prostheses, foreign bodies other than in the eyes or brain, some ventriculoperitoneal shunts, some artificial heart valves and first trimester of pregnancy are all relative contraindications which are useful for the MR department to know about well in advance, so that the individual patient’s circ*mstances can be checked (to see whether MR scanning will be possible without undue risk) and to tailor the scan to the patient. Patients who are dysphasic may need to have their orbits and chest X-rayed prior to MR to exclude any intraocular metallic foreign body, and pacemakers (feeling the chest wall is probably not a sufficiently reliable substitute for chest X-ray for excluding a pacemaker).

5.5.1 The appearance of haemorrhage on MRI The appearance of haemorrhage on MR is governed by the paramagnetic properties of haemoglobin breakdown products, and so it changes with time (Fig. 5.29).46 Freshly extravasated red blood cells contain oxyhaemoglobin which does not have any paramagnetic properties. As soon as the blood becomes deoxygenated, the signal properties change, and this can be detected on sequences that are particularly sensitive to the oxygen state of haemoglobin, such as the echoplanar sequences used for diffusion-weighted imaging, or T2*-weighted gradient echo sequences. Thus, depending on the state of the oxygen in the intracerebral haemorrhage (ICH), there may be little immediate signal change and, although a lesion may be visible, the differentiation from infarction or some other mass lesion may be difficult within the first few hours, even when specific blood-sensitive sequences are used (Fig. 5.30). In some patients with ICHs scanned within the first few hours, there may be serpiginous low signal (dark) bands at the margins of the lesion on gradient echo (T2*) sequences and sometimes on DWI. However we have seen cases where low signal bands were minimal, only visible on the T2*-weighted gradient echo sequence, and which could easily have been overlooked if only the DWI had been done (Fig. 5.30). Thus, while several publications are optimistic about the ability of MR to detect parenchymal haemorrhage reliably within the first few hours of acute stroke,159–162 we would

211

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 212

212

Chapter 5 What pathological type of stroke is it, cerebral ischaemia or haemorrhage?

(a) (b)

(d)

(c)

(e)

Fig. 5.28 Appearance of normal brain to compare various MR sequences. (a) T1-weighted midline sagittal – note the black CSF and little difference between grey and white matter. (b) T2-weighed fast spin echo axial – note CSF is very white (hyperintense), and grey matter is whiter (more intense) than white matter. (c) Fluid attenuated inversion recovery (FLAIR) axial – CSF is dark, grey and white matter are of similar intensity. (d) T2-weighted gradient echo axial – CSF is white, grey and white matter can be easily distinguished although the anatomical definition is not great – very sensitive to

haemosiderin and calcium if present (see also Figs 5.30–5.33). (e) Diffusion-weighted axial imaging – poor anatomical detail, dark CSF – acute ischaemic lesions are bright white. Note that the CSF appears white on T2 and gradient echo imaging and dark on the other sequences. T2 and T1 provide the best anatomical detail, FLAIR is useful for seeing abnormal tissue near CSF interfaces (see Fig. 5.38), gradient echo is very sensitive to blood, metal and calcium, and diffusion imaging is very sensitive to acute ischaemia.

continue to urge caution – acute haemorrhage on MR in general is just not as easy to diagnose as CT acute haemorrhage, especially for the inexperienced or unwary. However, increasing use of MR in hyperacute stroke has raised confidence in distinguishing lesions which are primarily haemorrhagic from those which are ischaemic. The time course of the change in appearance of an ICH on different sequences as it ages is now known to be much more variable than was originally thought.163 In general, the pattern is as follows, but it will vary with the

sequences used, and probably factors related to the patient and the ICH: • Once enough deoxyhaemoglobin has formed the ICH shows a centrally hypointense (dark) area on T1-weighted imaging and a markedly hypointense (dark) area on T2-weighted imaging. • As methaemoglobin is formed in the red cells, the lesion on the T1-weighted image becomes hyperintense (bright) while on the T2-weighted image it initially remains dark.

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 213

5.5 Magnetic resonance imaging

Oxy-Hb

Oedema Deoxy-Hb Deoxy-Hb Methaemoglobin

Oedema

Haemosiderin

Hemichromes + serous fluid

Methaemoglobin

= Normal brain signal intensity

Fig. 5.29 Diagram of the change in the appearance of parenchymal brain haemorrhage on MRI. The top row shows the typical appearance on T2-, and the lower row on T1-weighted imaging. The normal brain is represented by the stippled background. To a certain extent, the exact appearance and timing of the changes depend on the field strength of the

• As the methaemoglobin becomes extracellular and the ICH liquefies, the lesion on the T2-weighted image becomes as hyperintense (bright) as on the T1-weighted image. • After several weeks, lesions on T2-weighted images become bright in the centre with a very dark rim, and on T1-weighted images they are also bright centrally and moderately dark around the rim. • Eventually the dark rim (haemosiderin) around a bright ‘hole’ on T2 is the only remaining feature. The exact timing and degree of these signal changes depend on the strength of the magnet, which compartment of the brain the haemorrhage lies in, abnormal clotting, haematocrit, and the exact scan sequence used.46 Some sequences are very sensitive to the presence of haemoglobin breakdown products, e.g. T2* (gradient echo T2)164 and ‘susceptibility-weighted imaging’165 which show microhaemorrhages and petechial haemorrhage in infarcts, previously visible only to the pathologist (Fig. 5.31),166 while other sequences are much less blood-sensitive (see below). The clinical utility of being able to see tiny amounts of petechial haemorrhage or chronic microhaemorrhages remains uncertain (see below), but it is certainly important to make sure that the routine MR imaging protocol for stroke (acute or chronic) includes sequences which can reliably detect blood of any age.

Time (days)

magnet. The most important things to remember are that intracellular deoxyhaemoglobin (deoxy-Hb) is dark on T1 and T2, methaemoglobin is bright on T1 and T2, and haemosiderin is dark on T1 and T2. Haemosiderin persists in the margins of a haematoma for years after the original haemorrhage. Oxy-Hb, oxyhaemoglobin.

Does haemorrhage remain visible indefinitely on MR? The MR features of haemosiderin may persist for life (Fig. 5.32) because, in general, old ICHs are visible pathologically as haemosiderin deposits at postmortem.167 Haemosiderin is one of two major haemoglobin-derived pigments identifiable in tissue sections from around ICHs and represents aggregates of ferritin micelles within lysosomes of phagocytic cells. The other pigment is haematoidin, which is chemically identical to bilirubin, and forms locally as a result of haemoglobin breakdown in a milieu of reduced oxygen tension.167 Haemosiderin is ferromagnetic and therefore visible on MR, whereas haematoidin has no particular magnetic properties and is not visible on MR. We found that among 116 survivors of moderate to severe head injury of whom 78 had one or more parenchymal haemorrhages (total 106) at the time of injury, using a 1.0 Tesla MR scanner between 1 and 5 years after injury, 96 of the 106 (90%) haemorrhages were visible as haemosiderin on spin echo T2weighted imaging. Of the ten haemorrhages without haemosiderin on MR (11% of patients), seven were in patients where another haemorrhage with haemosiderin was still visible elsewhere in the brain. There were no features of the haemorrhage (site, size, density, sharpness of edges) in the acute stage which correlated

213

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 214

214

Chapter 5 What pathological type of stroke is it, cerebral ischaemia or haemorrhage?

(a)

(c)

(a)

(b)

(d)

Fig. 5.30 MR brain scan at 4 h after sudden right hemiparesis: (a) diffusionweighted; (b) T2-weighted; (c,d) gradient echo axial images. There is a large mass of increased signal on all sequences which could easily be mistaken for a tumour on T2 and an infarct on diffusion imaging. The clue that it was a haemorrhage is the hypointense (dark) wavy lines around the edges of the mass which represent blood products (thick arrows). Note that this patient has had an ICH previously (d) in the left frontal lobe (arrowhead) and has microhaemorrhages (c,d) (thin arrows). The CT scan from this patient is shown in Fig. 5.6.

(b)

Fig. 5.31 MR scan of a 5-day-old left striatocapsular infarct to show the extreme sensitivity of appropriate MR sequences to tiny areas of petechial haemorrhage. (a) T2-weighted spin-echo sequence shows the infarct (arrows), but no haemorrhage. (b) T2-weighted gradient-echo ‘haem’ sequence shows tiny low-signal areas within the infarct caused by tiny amounts of petechial haemorrhage (arrows) which would previously have only been visible to the pathologist. The clinical relevance of this is uncertain.

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 215

5.5 Magnetic resonance imaging

Fig. 5.32 MR scan showing an old ICH in the right basal ganglia. (a) T2 spin echo: the lesion has an obvious low signal (dark) area surrounding the high signal centre (arrow). (b) T2 fast spin echo: the lesion is more difficult to identify as an old haemorrhage as the low signal (dark) area is less obvious. (c) FLAIR: several small deep lesions shown (arrows), but no features to identify any of them as haemorrhage. (d) Gradient echo T2: the haemorrhage is obvious. Note the enlarged perivascular spaces (pinpoint high signal areas) best seen on T2 (b).

(a)

(b)

(c)

(d)

with the likelihood of haemosiderin being present at 1 year. Some quite large ICHs showed no trace of haemosiderin.168 Gradient echo (T2*-weighted) sequences are highly sensitive to deoxyhaemoglobin and haemosiderin (see above) but are still not routinely performed (in acute or chronic stroke) unless specifically requested, as they add an extra 2–5 min to the scanning time.158,163,164,166 Furthermore, areas of calcification give a similar appearance to haemosiderin on gradient echo sequences, and so may cause confusion.164 Fast spin echo T2-weighted and FLAIR sequences, particularly the latter, are too insensitive to deoxyhaemoglobin or haemosiderin to be reliable for identifying acute or old ICHs, or microhaemorrhages.158

In general, the characteristic signal changes of parenchymal haemorrhage on MRI persist indefinitely, but only if the appropriate imaging sequences are used; therefore, intracerebral haemorrhages can be identified even years after they have occurred. However, not all haemorrhages form haemosiderin during their resolution, therefore not all will be discernible as such on MR performed late after the event, even if the correct sequences are used. Microhaemorrhages More widespread use of MR in patients with stroke, dementia and normal older subjects has revealed that some patients, as well as occasional normal older people,

215

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 216

216

Chapter 5 What pathological type of stroke is it, cerebral ischaemia or haemorrhage?

(a)

(b)

have solitary or multiple small ( 1.2) and in the PWI lesion used (although most tested a mean transit time derivative). The available data showed that mismatch (vs no mismatch) without thrombolysis was

associated with a non-significant twofold increase in the odds of infarct expansion (OR 2.2), which did not change with thrombolysis (OR 2.0). Half the patients without mismatch also had infarct growth, and thrombolysis did not change the odds ratio (Fig. 5.51).310 These data suggest that mismatch may be associated with an increased risk of infarct growth but are very unreliable, particularly on the question of whether thrombolysis changes infarct growth. A trial testing the mismatch hypothesis is ongoing (EPITHET).327 About half of our own 46-patient cohort had mismatch on admission, and neither DWI/CBF nor DWI/MTT mismatch (both PWI parameters were semi-quantitative) predicted lesion growth; lesion growth was equally common in those with and without mismatch.311 Thus although ‘DWI/PWI mismatch’ is being used to identify patients for inclusion in trials of stroke treatments,326 our results imply that patients without ‘mismatch’ may also benefit and should not be denied that possibility by being excluded from acute stroke treatment trials. It is conceivable that the mismatch approach could be used to extend the time window for treatment. There is increasing evidence of viable but at risk tissue up to 24 h after stroke, which might be amenable to thrombolytic or other treatment.328 For example, thrombolysis is currently licensed for use within 3 h of stroke, but

233

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 234

234

Chapter 5 What pathological type of stroke is it, cerebral ischaemia or haemorrhage?

Fig. 5.51 An example of MR diffusion-perfusion mismatch. Top row: 4 h after onset of right hemiparesis shows T2, diffusion, ADC, relative CBF and mean transit time (relative MTT) maps. There is early ischaemic change on diffusion and ADC but not on T2; a perfusion deficit is visible on CBF and MTT, largest on MTT, but larger than the diffusion lesion on both CBF and MTT. The ‘DWI/CBF mismatch’ would be smaller

than the ‘DWI/MTT mismatch’. Bottom row: 7 days after stroke, the infarct is now visible on T2, diffusion and ADC and there is no longer any lesion on CBF or MTT which indicates reperfusion. The diffusion lesion does not seem to have grown particularly. This patient did not receive thrombolysis, so these perfusion changes were spontaneous. Also reproduced in colour in the plate section.

some clinicians use mismatch to decide whether to treat patients with thrombolysis between 3 and 6 h of stroke, and trials using DWI/PWI mismatch to select patients up to 9 h after stroke has recently finished (DIAS,326 DEDAS and DIAS 2).

circ*mstances such as suspected arterial dissection or venous disease (sections 5.8 and 7.21.3).

MR angiography Magnetic resonance angiography allows the acquisition of images of blood vessels, without injection of contrast, by using the signal characteristics of flowing blood.329–331 This technique can be used in acute ischaemic stroke, but the patient must keep very still for several minutes so it is not always of practical help (Fig. 5.52). It is used for the assessment of carotid stenosis in patients being considered for carotid endarterectomy (section 6.8.5)332 and is also being evaluated for the detection of intracranial aneurysms (section 9.4.3). We do not use it routinely in acute stroke as there is no good evidence that routinely identifying the presence of arterial occlusion influences management in the majority of patients. We use it in specific

5.5.3 Advanced magnetic resonance techniques Magnetic resonance (MR) permits other useful methods of examining the brain, including MR spectroscopy to show the effect of ischaemia on brain metabolites, and functional MR to identify areas of the brain which control particular body and mind functions. The details of these techniques are beyond the scope of this chapter, but there are excellent reviews describing each technique, as detailed below. Magnetic resonance spectroscopy Magnetic resonance spectroscopy (MRS) can demonstrate metabolic changes in ischaemic tissue in vivo, particularly hydrogen, phosphate, carbon, fluorine and sodium metabolism.329,333–335 N-acetyl aspartate (considered a marker of ‘normal neurones’), creatine and phosphocreatine, choline-containing compounds, lactate

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 235

5.5 Magnetic resonance imaging

(a)

Fig. 5.52 MR imaging with (a) T2, (b,c) proton density and (d) MR angiography of a left middle cerebral artery (MCA) branch occlusion. Note the high signal of recent thrombus is visible on proton density (c, arrow) and the infarct in the insular cortex (a,b, arrows). (a) On T2 only dark signal mimicking a flow void is visible in the MCA branch so the thrombus might be overlooked. (d) The MR angiogram confirms the MCA branch occlusion (arrow). See also Fig. 5.36c.

(d)

(c)

and pH have all been measured in stroke patients, but it is still too early to be certain of the significance of the changes.334,336–338 MRS can be performed by either the single voxel, in which a small cubic volume of brain – typically 8 cm3 – is sampled, or by chemical shift imaging techniques in which spectra from a whole slice of brain are obtained simultaneously (Fig. 5.53). Both scan techniques require the patient to keep still for up to 10 min and are not in routine use. Functional MR imaging (fMRI) When a part of the cerebral cortex becomes metabolically active its blood flow increases, thereby changing the MR signal obtained from this part of the brain. For example, if the subject is asked to undertake a task, such as wiggling a finger, while in the MR scanner, and the

(b)

MR prior to the movement is compared with the image during the task, the difference between the two demonstrates the area of cortex activated by the function. It is thought that areas of the brain which may perhaps be ‘taking over’ – or at least influencing in some way – the function of damaged areas, can be identified by their activation on fMRI, and the pattern of brain compensation to injury studied during recovery from ischaemic stroke or head injury, or in patients with brain tumours. However, the technique is difficult and great care is required to obtain meaningful results. The subject must keep his or her head very still and only perform (and think about) the requested task, because inadvertent movement of other parts of the body will appear on MR and be misleading. Nonetheless, the technique shows promise by allowing insights into normal brain function as well as into recovery from injury.339

235

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 236

236

Chapter 5 What pathological type of stroke is it, cerebral ischaemia or haemorrhage?

(a)

(b)

5.5.4 Comparative studies of CT and MRI in acute stroke MR is more sensitive for detecting minor degrees of haemorrhagic transformation of infarcts during followup, but CT is better than MR (even with gradient echo imaging) for detecting acute intracerebral haemorrhage. Although MR may identify low signal areas indicating haemoglobin breakdown products, these can be quite subtle in the acute stage and CT is simply much easier to interpret.340 Most studies of intracerebral haemorrhage on MR in hyperacute stroke report apparently high sensitivity for haemorrhage,159 but in our experience acute haemorrhage can have few distinguishing features from an ischaemic lesion or other pathology and in some cases we have ended up doing a CT scan to make the diagnosis. Most studies comparing the visibility of ischaemic lesions on brain CT and MRI in acute stroke performed MR after CT, making it likely that MR would be more often abnormal than CT. Some did not take account of the case mix of mild to severe strokes. Only one study apparently randomized the order of scanning,341 but others found this to be impractical.151 Several early studies suggested that in patients imaged within the first 6 h of stroke, DWI had a higher sensitivity and specificity and inter-rater reliability than CT.342–345 All found excellent agreement between CT and DWI. DWI was performed on average 2–3 h after CT in all studies. None mentioned the proportion of patients who could not be scanned with either technique or in whom the images were not diagnostic. A more recent and complete comparison of CT and MR in consecutive patients with moderate and severe

Fig. 5.53 MR spectroscopic chemical shift imaging at 1 day after stroke in a patient with a right middle cerebral artery infarct. The T2-weighted image is in the background and the spectroscopic colour maps of the distribution of (a) the normal neuronal metabolite N-acetyl aspartate (NAA) and (b) the marker of anaerobic metabolism lactate, are superimposed. Red indicates large amounts and blues and greens small amounts of the metabolites. You will need to see the colour plate section to appreciate the following. The area of increased T2 signal of the infarct corresponds with the area of blue on the NAA map (in keeping with neuronal loss) and red on the lactate map (indicating anaerobic metabolism taking place in the infarct). Also reproduced in colour in the plate section.

strokes admitted hyperacutely and considered for thrombolysis found that MR could not be obtained in about one-third (patient too unwell, uncooperative, scanner unavailable despite ‘24-hour’ availability) and when both modalities were interpreted using ASPECTS (a structured scoring system for scan interpretation) there was no difference between CT and DWI in the number of infarcts detected or in the inter-rater reliability.151 Thus for patients admitted and scanned within 6 h of symptom onset, who are likely to have moderate to severe strokes, CT with a structured scoring system can detect as many early infarcts as DWI, detects haemorrhage better than MR, and can image all patients whereas the practical limitations of MR prevent imaging in up to one-third of patients. The situation is different for patients with mild strokes, including lacunar syndromes, and who present later, as discussed above. In this case, MR with DWI and gradient echo imaging is much more likely to show the acute ischaemic lesion (or old haemorrhage) than CT and this patient group is easier to scan with a much higher success rate. In patients with moderate to severe strokes, CT with a structured scoring system can detect as many early infarcts within 6 h as DWI, detects haemorrhage better than MR and can image all patients whereas the practical limitations of MR prevent imaging in up to one-third of patients. On the other hand, MRI with DWI is superior for lacunar, posterior fossa and small cortical and subcortical infarcts. Whether MRI is ‘better’ than CT depends on the question being asked of imaging. MRI is certainly less practical and takes longer.

9781405127660_4_005.qxd 10/13/07 11:05 AM Page 237

5.6 Other ‘sophisticated’ methods of imaging cerebral ischaemia

areas in terms of symptoms and signs not fully explained by loss of the area of brain directly involved in the infarct.358,359 Positron emission tomography

Single-photon emission CT Single-photon emission CT (SPECT) is a technique in which cross-sectional images of the brain are obtained by using either a rotating gamma-camera, or a dedicated SPECT scanner, and computerized reconstruction of the data from the emitted radiation following intravenous administration of a radioactive isotope. Various isotopes are available, such as hexamethyl propylene amine oxime (HMPAO) (Ceretec) which images regional perfusion and therefore can identify areas of abnormal perfusion early, or xenon-133, which can measure regional cerebral blood flow quantitatively.346 However, the equipment for SPECT is not widely available, the scanning times are long (20 min or more) and so problems caused by movement artifacts can arise from restless stroke patients, and the isotopes are expensive. A potential source of error, which is particularly relevant to ischaemic stroke, is that blood–brain barrier breakdown alters the behaviour of the isotope and hence the accuracy and interpretation of the images.347 SPECT can demonstrate appropriate areas of low flow within a few hours of stroke onset, and even up to 60% of patients with a transient ischaemic attack have an appropriate abnormality within 24 h of onset.348 In acute ischaemic stroke, SPECT demonstrates an area of relative hypoperfusion corresponding with the ischaemic tissue in most patients studied, although very few if any lacunar stroke patients were included in the studies.348–353 Several studies have indicated that the size of the initial perfusion deficit on SPECT predicts the likely clinical outcome: the larger the deficit, the worse the outcome.84,350–353 Also, evidence of reperfusion on SPECT is associated with improved clinical outcome both spontaneously352,353 and in response to thrombolytic therapy.84,354,355 In the first week after the stroke, the relative blood flow in the infarct may be increased or decreased, corresponding with luxury perfusion or persistent hypoperfusion, respectively.356 Abnormalities may be visible not only in the infarct itself, but also in remote parts of the brain, such as the opposite cerebral hemisphere or the contralateral cerebellum in patients with cerebral cortical infarcts.357,358 This has given rise to the concept of ‘diaschisis’ or ‘shutdown’ of areas of the brain not directly involved in the infarct but whose function is in some way influenced, possibly through association fibres.359 There is some evidence of clinical correlates of ‘shutdown’ of these

Positron emission tomography (PET) can display a variety of physiological processes, from glucose metabolism to neuroreceptor density, thus allowing study of the normal workings of the brain, as well as the consequences of pathology.360–362 The data, derived from collection and analysis of complicated signals from positronemitting isotopes, may be shown as colour-coded tomographic pictures. PET has been available since the 1970s, but the imaging equipment and isotopes are extremely expensive, the scan times are prolonged and the patient must keep very still. Therefore PET is likely to remain a research tool with limited clinical applications, at least in stroke for the foreseeable future.363,364 PET studies in ischaemic stroke have provided evidence that the ischaemic penumbra really does exist in humans and that neural tissue in the penumbra may survive for up to 48 h.328,361,365–368 They have also suggested that neurones in areas of the brain with cerebral blood flow below the threshold traditionally regarded as producing irreversible damage ( 160/90 mmHg, at least twice pre-stroke) Angina and/or past myocardial infarction Current smoker Claudication and/or absent foot pulses Major cardiac embolic source Transient ischaemic attack Cervical arterial bruit Diabetes mellitus Any of the above

126

(a)

Endothelium Fatty streak

66 60

27 25

50 35 33 24 196

20 14 14 10 80

Lipid core (c)

extensive or severe atheroma when exposed to causal risk factors, such as hypertension, while the arterial anatomy determines where the atheroma occurs. There appear to be important racial differences in the distribution of atheroma, and race is an independent predictor of lesion location.42 White males tend to develop atheroma in the extracranial cerebral vessels, the aorta and coronary arteries while intracranial large vessel disease appears to be relatively more common in black, Hispanic and Asian populations43–45 and tends to affect younger patients43,45 and those with type 1 diabetes mellitus.43 Some sources report that women have more intracranial disease compared to men but this is disputed by others.44 However, carotid bifurcation stenosis does appear to be becoming more frequent in Oriental populations either because the pattern of atheroma is evolving as a result of lifestyle changes, or because previous studies were confounded by selection and other biases.44,46,47 There are very few data on the aetiology of stroke in Africa, but reports from stroke registers in South Africa show that atheroma is a less common cause of stroke in blacks than in whites.48 Individuals with atheroma affecting one artery almost always have atheroma affecting many other arteries, with or without clinical manifestations.

6.3.2 The nature, progression and clinical consequences of atheroma Atheroma begins as intimal fatty streaks in children, it is thought in response to endothelial injury18,19 (Fig. 6.3).

Fibrous cap

(b)

Platelet–fibrin thrombus

Haemorrhage

Necrosis

Calcification Inflammatory cell infiltrate

(d)

Embolus

Platelet–fibrin thrombus

Fibrosis Ulceration

Calcification

Fig. 6.3 The growth, progression and complications of atheromatous plaques: (a) early deposition of lipid in the artery wall as a fatty streak; (b) further build-up of fibrous and lipid material; (c) necrosis, inflammatory cell infiltrate, calcification and new vessel formation, leading to (d) plaque instability, ulceration and platelet–fibrin thrombus formation on the plaque surface.

Over many years, circulating monocyte-derived macrophages adhere to and invade the arterial wall. As a result, there is an inflammatory response with cytokine production and T-lymphocyte activation. Intra- and later extracellular cholesterol and other lipids are deposited, particularly in macrophages, which are then described as foam cells. Arterial smooth-muscle cells migrate into the lesion and proliferate, fibrosis occurs and so fibrolipid plaques are formed. These plaques, with their lipid core

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 265

6.3 Atheroma and large vessel disease

and fibrous cap, encroach upon the media and spread around and along the arterial wall. Some become necrotic, ulcerated and calcified with neovascularization and haemorrhage – so-called complicated plaques. The arterial wall thickens, the vessel dilates or the lumen narrows and the artery becomes stiffer and tortuous. Atherosclerotic lesions may cause ischaemic stroke through acute vessel occlusion, or by reducing vessel diameter and thus regional cerebral blood flow. Whether acute occlusion of a cerebral vessel, or reduction in cerebral blood flow, leads to infarction depends on not only for how long the blood flow is impaired, but also the availability and functional capability of the collateral circulation49–53 (section 4.2). Arterial occlusion secondary to atheroma occurs by three mechanisms. First, thrombi may form on lesions and cause local occlusion. Second, embolization of plaque debris or thrombus may block a more distal vessel; emboli are usually the cause of obstruction of the anterior circulation intracranial vessels29,30 at least in white males in whom intracranial atheroma is relatively rare. Third, small vessel origins may be occluded by growth of plaque in the parent vessel. This is seen particularly in the basilar artery and the subclavian artery around the vertebral origin. Low regional cerebral blood flow secondary to atheroma occurs when plaque growth causes severe reduction in the diameter of the vessel lumen and hypoperfusion of distal brain regions, particularly in the border zones where blood supply is poorest. This may lead to ‘borderzone infarction’ (section 6.7.5) of these regions following severe hypotension or hypoxia. The mechanisms of atherosclerosis-related stroke are discussed further below.

6.3.3 Atheromatous plaques complicated by thrombosis: atherothrombosis From an early stage, or perhaps even from the very first stage, atheromatous plaques promote platelet adhesion, activation and aggregation, which initiates blood coagulation and thus mural thrombosis.54–56 At first, any thrombus may be lysed by fibrinolytic mechanisms in the vessel wall, or incorporated into the plaque, which re-endothelializes and so ‘heals’. Gradually, the atheroand then atherothrombotic plaque grows, in part because of repeated episodes of mural thrombosis layering one on top of the other, and eventually the lumen may become obstructed. Such occlusive intraluminal thrombus may then propagate proximally or distally in the column of stagnant blood, but usually no further than the next arterial branching point. Thrombosis is opposed by the release of prostacyclin and nitric oxide, both vasodilators, from the vascular

endothelium and by endothelium-derived plasminogen activator. The balance of these pro- and antithrombotic factors may determine whether a thrombus complicating an atheromatous plaque or an occlusive embolus grows, is lysed or becomes incorporated into the arterial wall and so contributes to the gradually enlarging atherothrombotic plaque. Symptomatic in situ acute atherothrombotic occlusion – rather than artery-to-artery embolism – does not appear to be a very common cause for ischaemic stroke or transient ischaemic attacks (TIAs) in the carotid territory. Perhaps this is because atheroma affects the larger arteries (e.g. internal carotid artery (ICA) rather than middle cerebral artery) and it takes a very large plaque to occlude them, or because the potential for collateral blood flow is better distal to larger arteries.29,57 Indeed, once the ICA has occluded, the risk of ipsilateral ischaemic stroke appears to be less than for severe stenosis.58–60 On the other hand, symptomatic in situ atherothrombotic occlusion may be more common in the posterior circulation (e.g. of the basilar artery) but even here artery-to-artery embolism is well described.61–65

6.3.4 Embolism from atherothrombotic plaques: atherothromboembolism Atheroma and/or thrombus may embolize – in whole or in part – to obstruct a smaller distal artery, usually at a branching point – the same one or different ones on several occasions. Emboli consist of any combination of cholesterol crystals and other debris from the plaque, platelet aggregates, and fibrin which may be recently formed and relatively friable or old and well organized. Depending on their size, composition, consistency and age – and presumably the blood flow conditions at the site of impaction – emboli may be lysed, fragment and then be swept on into the microcirculation. Alternatively, they may permanently occlude the distal artery and promote local antero- and retrograde thrombosis, which is further encouraged by the release of thromboxane A2 from platelets, which is also a vasoconstrictor. Emboli are transmitted to the brain or eye via their normal arterial supply, which itself varies somewhat in distribution between individuals (section 4.2). An embolus from an atherosclerotic carotid bifurcation – usually a plaque at the origin of the internal carotid artery (ICA) but sometimes the distal common or proximal external carotid arteries – normally goes to the eye or the anterior two-thirds of the cerebral hemisphere. But, on occasion, it may go to the occipital cortex if blood is flowing from the ICA via the posterior communicating artery to the posterior cerebral artery, or if there is a fetal

265

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 266

266

Chapter 6 What caused this transient or persisting ischaemic event?

origin (section 4.2.3) of the posterior cerebral artery from the distal ICA (5–10% of individuals). However, if an artery is already occluded, then an embolus may travel via the collateral circulation and impact in an unexpected place. For example, with severe vertebral arterial disease, and therefore poor flow distally into the basilar artery, an embolus from the ICA origin may reach the basilar artery via the circle of Willis. With ICA occlusion, it is still possible to have an ipsilateral middle cerebral artery (MCA) distribution cerebral infarct as a result of: • an embolus travelling from the contralateral ICA origin via the anterior communicating artery; • an embolus from any blind stump of the occluded ICA, or from disease of the ipsilateral external carotid artery (ECA), via the ECA and orbital collaterals to the MCA; • an embolus from the tail of thrombus in the ICA distal to the occlusion; or • low flow distal to the ICA occlusion, perhaps within a boundary zone (section 6.7.5), particularly if the collateral blood supply is poor, cerebrovascular reactivity is impaired and the oxygen extraction ratio high.652,66–69 Curiously, emboli from the neck arteries (or from the heart) seldom seem to enter the small perforating arteries of the brain to cause lacunar infarction (section 6.4) perhaps as a consequence of the fact that the perforating vessels arise at a 90° angle from the parent vessel.

(a)

(b)

6.3.5 Atherothrombosis/embolism as an acute-on-chronic disorder: plaque instability Like the coronary arteries, atheromatous plaques in the cerebral circulation – particularly at the carotid bifurcation – become ‘active’ or ‘unstable’ from time to time with fissuring, cracking or rupture of the fibrous cap, or ulceration. The histological features of plaque instability are a thin fibrous cap, large lipid core, reduced smooth muscle content and high macrophage density31,70–72 (Fig. 6.4a). If the thrombogenic centre of the plaque is exposed to flowing blood, then complicating thrombosis occurs. Plaque instability may even be a ‘systemic’ tendency because irregularity on catheter angiography – and so presumed instability and ulceration – of symptomatic carotid stenosis is associated with irregularity of the asymptomatic contralateral carotid artery, and with coronary events assumed to be caused by plaque rupture.73 At other times the plaque is quiescent with a thick fibrous cap, or slowly growing, without causing any clinical symptoms30,31,74–77 (Fig. 6.4b). In other words, atherothromboembolism is an ‘acute-on-chronic disorder’. It is no surprise therefore that the clinical complications of atheroma reflect this: • transient ischaemic attacks (TIA) tend to cluster in time78 (section 3.2.1); • stroke tends to occur early after a TIA and affect the same arterial territory79 (section 16.2.1);

Fig. 6.4 Photomicrographs of transverse sections of the carotid sinus illustrating the pathological features of atherosclerotic plaque stability. (a) An unstable plaque characterized by a large necrotic core (✽), and a thin fibrous cap heavily infiltrated by macrophages (arrow). Elsewhere the plaque is ulcerated and the lumen occluded by thrombus, part of which had embolized to the ipsilateral middle cerebral artery causing a fatal ischaemic stroke. (b) A comparable stenotic but stable plaque, comprising largely fibrous tissue (F) with focal calcification (arrow). There is no significant necrotic core or inflammatory cell infiltrate. L lumen; A tunica adventitia; M tunica media. (Provided by Dr Alistair Lammie.)

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 267

6.4 Intracranial small vessel disease

• the risk of ischaemic stroke ipsilateral to severe carotid stenosis is highest soon after symptomatic presentation and then declines,73,80 even though the stenosis itself seldom regresses (Fig. 16.38); • presumed artery-to-artery embolic strokes tend to recur particularly early81 (section 16.2.3); • emboli are more often detected with transcranial Doppler sonography if carotid stenosis is severe or recently symptomatic;82–87 and • the rate of Doppler-detected emboli in the middle cerebral artery tends to decline with time after stroke.88 Increasing severity of symptomatic atherothrombotic stenosis, at least at the origin of the internal carotid artery, is undoubtedly a powerful predictor and cause of ischaemic stroke ipsilateral to the lesion.89 However, this cannot be the whole explanation because: • the risk of stroke distal to an asymptomatic stenosis is far less than distal to a recently symptomatic stenosis of the same severity (section 16.12.5); • by no means all patients with severe stenosis have a stroke (section 16.11.5); and • relatively mildly stenosed plaques can be complicated by acute carotid occlusion.31 Curiously, there is not such an obvious relationship between increasing coronary artery stenosis and coronary events. This may be because the coronary arteries are harder to image repeatedly and are more anatomically complicated, because coronary events are more often ‘silent’, or because the coronary arteries are smaller and more likely to be blocked if even a small plaque ruptures.90 Independently of the severity of stenosis, plaque irregularity on catheter angiography is associated with increased stroke risk, probably because irregularity represents plaque ulceration and instability with thrombosis and so likely complicating embolism71,85,91 (section 16.11.8). Atherothromboembolism is an acute-on-chronic disorder, both in its pathology and in its clinical manifestations. Although the formation of atherothrombotic plaques must be a long and gradual process over many years, the clinical manifestations usually occur acutely (e.g. an ischaemic stroke) and tend to cluster in time. For example, stroke tends to occur sooner rather than later after a transient ischaemic attack, perhaps as a result of the breakdown and ‘instability’ of an atherothrombotic plaque which later ‘heals’. Exactly why one plaque becomes unstable (and then perhaps ulcerates with complicating thrombosis) and another does not, is unknown. Histological comparisons

of recently symptomatic with asymptomatic carotid plaques – matched for stenosis severity – are not easy to perform. However, it seems as though intraplaque haemorrhage, calcification and the lipid core are similar in both but, crucially, the thickness of the fibrous cap has not generally been assessed.92–94 Similar problems arise with attempts to compare inflammatory processes and the expression of adhesion molecules,95,96 metalloproteinase expression,97 differences in plaque tissue factor98 and in plaque geometry and motion and so potential stresses on the fibrous cap.99,100 Whether infection causes plaque instability is very unclear (section 6.6.17). In all these cross-sectional studies there is always the possibility of reverse causality – that by becoming symptomatic a plaque is changed in its anatomy, motion, biochemistry, and so on.

6.3.6 Dolichoectasia This somewhat unusual pattern of arterial disease tends to affect the medium-sized arteries at the base of the brain, particularly the basilar artery, mostly in the elderly but occasionally in children101–105 (section 4.2.3). The arteries are widened, tortuous, elongated and are often enlarged enough to be seen as characteristic flow voids on MRI, and as tortuous channels – even without enhancement – or by virtue of calcification in their walls on brain CT106 (Fig. 6.5). When found in an individual, this arterial abnormality is not necessarily the cause of any ischaemic stroke (and very rarely of intracranial haemorrhage, despite the aneurysmal dilatation). However, these enlarged vessels can contain thrombus which embolizes, or occludes the origin of small branch arteries of the ectatic vessel, or even occludes the ectatic vessel itself. Cranial nerve and brainstem dysfunction due to direct compression or small vessel ischaemia, or hydrocephalus caused by cerebrospinal fluid pathway compression, are other occasional complications of basilar (or vertebral) ectasia.104,105,107–109 Atheroma is the most common cause.110 Other causes include various types of congenital defect in the vessel wall, Marfan syndrome, pseudoxanthoma elasticum and Fabry’s disease111 – both in hom*ozygotic males and heterozygotic females.

6.4 Intracranial small vessel disease

There are a number of pathologies which affect the small (40–400 µm diameter) arteries, arterioles and veins

267

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 268

268

Chapter 6 What caused this transient or persisting ischaemic event?

(a)

(b)

(c)

Fig. 6.5 Vertibrobasilar dolichoectasia. (a) On the brain CT scan there is a 1.5–2 cm rounded mass (arrows) which is of slightly higher density than adjacent brain and sits in the left cerebellopontine angle. The mass is indenting the brainstem. On adjacent sections the mass was obviously longitudinal and contiguous with the vertebral and top end of the basilar arteries. (b) MR examination of the same patient shows the mixed signal mass (arrows). The presence of increased signal indicates either very slowly flowing or partially clotted blood. (c) Catheter angiography shows a lateral projection of the vertebrobasilar circulation. Although the main area of expansion of the arterial system is in the lower basilar artery (arrows), the dolichoectasia actually extends from the upper right vertebral artery almost to the very tip of the basilar artery.

of the meninges and brain – for example, cerebral amyloid angiopathy (section 8.2.2); vasculitis (section 7.3); atheroma near the origin of the small perforating arteries (see below); and the angiopathy underlying cerebral autosomal dominant arteriopathy with subcortical infarcts and leucoenceophalopathy (CADASIL)

(section 7.20.1). Much more common, however, is what has been termed hyaline arteriosclerosis. This is an almost universal change in the small arteries and arterioles of the aged brain, particularly it is said in the presence of hypertension or diabetes, but sometimes in quite young patients without any of the classical

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 269

6.4 Intracranial small vessel disease

(a)

(b) Fig. 6.6 Photomicrograph of perforating lenticulostriate artery branches in the putamen, illustrating two distinctive patterns of vessel pathology. (a) Concentric hyaline wall thickening with a few remaining vascular smooth-muscle cell nuclei (arrow). The lumen remains patent. Such ‘simple’ small vessel disease is an almost invariable feature of elderly brains, most prominent in hypertensive and diabetic patients. (b) A complex, disorganized vessel segment showing an asymmetrical destructive process with focal fibrinoid material (✽) and mural foam cells (arrow). The lumen is visible cut in two planes of section. In this case the vascular lesion was adjacent to and, presumably, the cause of a right striatocapsular lacunar infarct. This ‘complex’ vessel lesion corresponds with what Miller Fisher termed ‘lipohyalinosis’. (Provided by Dr Alistair Lammie.)

vascular risk factors112 (Fig. 6.6). Important factors in its development appear to include not only hypertension and diabetes, but also breakdown of the blood–brain barrier with incorporation of plasma proteins into the vessel wall, possibly due to endothelial dysfunction.113 The smooth muscle cells in the wall are eventually replaced by collagen, which reduces vascular distensibility, and presumably reactivity, but not necessarily the size of the lumen.

Unfortunately, over the years, the pathological nomenclature for small vessel disease has been very confusing and, as a result, the ‘simple’ small vessel disease (SVD) described above has frequently been confused with a more aggressive-looking disorder of small arteries with disorganization of their walls and foam cell infiltration. It is this ‘complex’ SVD that Fisher first called ‘segmental arterial disorganization’ and then ‘lipohyalinosis’. The term fibrinoid vessel wall necrosis refers to a different, characteristic, but more acute change in the vessel wall seen, for example, as a consequence of accelerated hypertension and as a reactive phenomenon around acute intracerebral haematomas, spontaneous or traumatic. When healed, this probably takes on the appearance of ‘complex’ SVD114,115 (Fig. 6.6). However, it is by no means certain that ‘complex’ SVD and acute fibrinoid necrosis are more advanced forms of ‘simple’ SVD, and there is no evidence that ‘simple’ SVD is the ‘healed’ version of complex SVD. Nonetheless, both ‘simple’ and ‘complex’ SVD tend to affect the lenticulostriate perforating branches of the middle cerebral artery, the thalamoperforating branches of the proximal posterior cerebral artery, the perforating branches of the basilar artery to the brainstem, and the vessels in the periventricular white matter.116 Most of what is known about small vessel disease and lacunar stroke comes from a very small number of very careful clinicoanatomical observations by Fisher. He did not describe ‘simple’ SVD as the underlying vascular pathology of lacunes, but what we prefer to call ‘complex’ SVD in some cases, and atheroma affecting the origins of the small vessels where they come off the circle of Willis and major cerebral arteries in others.117–119 In theory, one can think of ‘mural atheroma’ affecting the parent arteries, ‘junctional atheroma’ affecting the origin of the small perforating arteries where they leave the parent artery, and ‘microatheroma’ affecting the proximal parts of these small arteries. In practice, these distinctions are more or less impossible to make at postmortem and the contribution of intracranial atheroma to lacunar infarction has still to be established. The current view is that both ‘complex’ SVD and atheroma at or near the origin of the small perforating vessels arising from the major cerebral arteries cause most, but not all, of the small deep infarcts responsible for lacunar ischaemic strokes (and, by implication, lacunar transient ischaemic attacks) which make up about one-quarter of symptomatic cerebral ischaemic events9,12,58,120–124 (sections 4.3.2 and 6.7.3). However, this hypothesis is not universally accepted125,126 since there is remarkably little direct postmortem evidence of occlusion of these vessels leading to lacunar infarcts,

269

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 270

Chapter 6 What caused this transient or persisting ischaemic event?

25 Lacunar (n = 222) Non-lacunar (n = 404)

20 Percentage of patients

270

0 –9

10–19 20–29 30–39 40 –49 50–59 60–69 70–79 80–89 90–99

Severity of stenosis (%) of symptomatic carotid artery

largely because the case fatality of lacunar stroke is so low and pathological material so scanty. The main supporting evidence for a specific small vessel lesion leading to lacunar infarction is indirect: • The relative lack of large vessel atheroma or embolic sources in the heart in the vast majority of lacunar patients, compared to those with cortical infarcts12,127–136 (Fig. 6.7). • Emboli are rarely if ever detected in the middle cerebral or common carotid artery by Doppler ultrasound in most studies of patients with lacunar infarction. However, it is not clear at what stage after the onset of stroke it is best to look – too early and an occluded middle cerebral artery might prevent the passage of emboli, too late and any proximal embolic source may have ‘healed’.88,137–140 • The low risk of early recurrence also argues against the concept of an active embolic source, either in the heart or an unstable atheromatous plaque.11,81,135,136 • The ‘capsular warning syndrome’ might suggest that a single perforating vessel is intermittently on the verge of occluding before it finally does so (section 6.7.3). • The impaired cerebrovascular reactivity in lacunar patients might suggest a specific problem, if not pathology, of the small intracerebral resistance vessels,141 such as endothelial dysfunction.113 It is conceivable that ‘complex’ SVD causes vessel rupture, leading to intracerebral haemorrhage as well as arterial occlusion, maybe at sites of so-called Charcot– Bouchard microaneurysms, although these may be artifacts of the pathological specimens. In fact, whether they are real or not is semantic, because what matters is not if the vessel wall bulges, but what weakens it in the first place122,142–146 (section 8.2.1). On the other hand, leukoaraiosis appears to be more associated with ‘simple’

100

Fig. 6.7 The relationship between the severity of symptomatic carotid stenosis and the likelihood of finding a nonlacunar (territorial) or lacunar infarct on the baseline brain CT in 626 patients in the European Carotid Surgery Trial. Those with severe stenosis are less likely to have lacunar infarcts. (With permission from Boiten et al. 1996.150)

SVD, and indeed this may be one underlying vascular cause, although it is clearly insufficient on its own, being so common in elderly brains. Interestingly, ‘cortical’ (presumed atherothrombotic) and ‘lacunar’ ischaemic stroke patients probably have a similar vascular risk factor profile (section 6.6), including hypertension.6,12,135,136,147–151 It is conceivable therefore that the same type of individual (i.e. hypertensive, diabetic, etc.) develops either small vessel disease (complex or atheroma) and so lacunar infarcts, or large vessel atherothromboembolism and so cortical infarction. The difference in the type of ‘degenerative’ vascular disease that occurs may perhaps reflect differing genetic susceptibilities. But, on the other hand, many individual patients have these different ischaemic stroke types at different times.144,152 On balance, we believe it likely that the lacunar hypothesis is correct and so, whatever the exact nature of the underlying small vessel lesion, lacunar infarction is seldom the result of embolization from proximal sites. About one-quarter of all ischaemic strokes and transient ischaemic attacks are ‘lacunar’ and many, if not most, lacunar infarcts are caused by disease of the small intracranial perforating arteries, either ‘complex’ small vessel disease, or atheroma affecting their proximal parts as they arise from their parent cerebral arteries. One question that has never been answered is whether there is a similar small vessel disease that affects the blood supply to the optic nerve and retina: are there ‘lacunar’ ocular syndromes equivalent to lacunar cerebral syndromes? What is clear, however, is that a high proportion of patients with ischaemic amaurosis fugax,

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 271

6.5 Embolism from the heart

retinal infarction and anterior ischaemic optic neuropathy do not have any detectable and likely proximal source of embolism (or evidence of low flow) in the arterial supply to the eye or the heart. Perhaps it is these patients who have small vessel disease like patients with ischaemic lacunar strokes.58,153,154

6.5 Embolism from the heart

That embolic material can pass from the heart to the brain, and from the venous system through the heart to the brain (paradoxical embolism), as well as to other organs, is undisputed. However, not all cardiac sources of embolism pose equal threats. For example, a mechanical prosthetic valve is much more likely to cause thromboembolism than mitral leaflet prolapse. In developed countries, embolism from the heart probably causes about one-fifth of ischaemic stroke and transient ischaemic attacks (TIAs), although a potential embolic source may be present in nearer one-third155–160. However, there are two very real and tiresome problems. As technology advances, more and more potential cardiac sources of embolism are being identified (Table 6.4), and patients may have two or more competing causes of cerebral ischaemia, such as carotid stenosis and atrial fibrillation. Therefore, it may well be unclear whether embolism from the heart is the cause in an individual patient, especially when the cardiac lesion is common in normal people.57,58,154,161–163 Not all emboli are of the same size, of the same age, or made of the same thing (fibrin, platelets, calcium, infected vegetations, tumour, etc.). Some are large and impact permanently in the mainstem of the middle cerebral artery to cause total anterior circulation infarction, others impact in a more distal branch of a cerebral artery to cause a partial anterior circulation infarct, others merely cause a transient ischaemic attack, and still others are asymptomatic.164–167 Emboli may also occlude the basilar artery and its branches, and even the internal carotid artery in the neck.61,165 Embolism from the heart causes about one-fifth of ischaemic strokes and transient ischaemic attacks. The most substantial embolic threats are non-rheumatic and rheumatic atrial fibrillation, infective endocarditis, prosthetic heart valve, recent myocardial infarction, dilated cardiomyopathy, intracardiac tumours and rheumatic mitral stenosis.

Table 6.4 Cardiac sources of embolism in anatomical sequence. Right-to-left shunt (paradoxical embolism from the venous system or right-atrial thrombus) (section 6.5.12) Patent foramen ovale Atrial septal defect Ventriculoseptal defect Pulmonary arteriovenous fistula Left atrium Thrombus atrial fibrillation* (section 6.5.1) sinoatrial disease (sick sinus syndrome) (section 6.5.14) atrial septal aneurysm (section 6.5.12) Myxoma and other cardiac tumours* (section 6.5.13) Mitral valve Rheumatic endocarditis (stenosis* or regurgitation) (section 6.5.4) Infective endocarditis* (section 6.5.9) Mitral annulus calcification (section 6.5.6) Non-bacterial thrombotic (marantic) endocarditis (section 6.5.10) Systemic lupus erythematosus/antiphospholipid syndrome (section 7.3.3 and 4) Prosthetic heart valve* (section 6.5.3) Papillary fibroelastoma (section 6.5.13) Mitral leaflet prolapse (uncertain) (section 6.5.7) Mitral valve strands (uncertain) (section 6.5.8) Left ventricle Mural thrombus acute myocardial infarction (previous few weeks)* (section 7.10) left-ventricular aneurysm or akinetic segment (section 7.10) dilated or restrictive cardiomyopathy* (section 6.5.11) mechanical ‘artificial’ heart* blunt chest injury (myocardial contusion) Myxoma and other cardiac tumours* (section 6.5.13) Hydatid cyst Primary oxalosis (uncertain, section 7.20.6) Aortic valve Rheumatic endocarditis (stenosis or regurgitation) (section 6.5.4) Infective endocarditis* (section 6.5.9) Syphilis Non-bacterial thrombotic (marantic) endocarditis (section 6.5.10) Systemic lupus erythematosus/antiphospholipid syndrome (section 7.3.3 and 4) Prosthetic heart valve* (section 6.5.3) Calcific stenosis/sclerosis/calcification (section 6.5.6) Aneurysm of the sinus of Valsalva Congenital heart disease (particularly with right-to-left shunt) Cardiac manipulation/surgery/catheterization/ valvuloplasty/angioplasty* (section 7.18.1) *Substantial risk of embolism.

271

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 272

272

Chapter 6 What caused this transient or persisting ischaemic event? Table 6.5 Prevalence of potential cardiac sources of embolism in 244 patients with a first-ever-in-a-lifetime ischaemic stroke in the Oxfordshire Community Stroke Project (Sanderco*ck et al. 1989).10 Source

Number

Percentage

Any atrial fibrillation without rheumatic heart disease with rheumatic heart disease Mitral regurgitation Recent ( 85

Fig. 6.15 The incidence of first-ever-in-a-lifetime ischaemic stroke and transient ischaemic attack (TIA) in the Oxfordshire Community Stroke Project. The flattening of TIA and, to some extent, stroke incidence in old age may be because cases did not come to medical attention or, when they did, they were not correctly diagnosed in the elderly.

281

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 282

Chapter 6 What caused this transient or persisting ischaemic event?

First-ever-in-a-lifetime ischaemic stroke Transient ischaemic attack

4.00

2.00 Relative risk of stroke

Male/female ratio of incidence

282

1.00

0.50

0.25

15– 44

45–54

55– 64

65–74

75 – 84

> 85

Age (years) Fig. 6.16 The ratio of male : female incidence of first-ever-in-alifetime ischaemic stroke and transient ischaemic attack in the Oxfordshire Community Stroke Project.

explained by the natural menopause, although bilateral oophorectomy without oestrogen replacement about doubles the risk of vascular events.317,318 Use of hormone replacement therapy in women after the menopause does not confer a benefit in terms of reduced cardiovascular morbidity, in fact there is evidence that such treatment is associated with an increased risk of acute coronary sydrome, stroke and venous thromboembolism319–321 (section 7.13.2).

6.6.3 Blood pressure In healthy populations of both sexes and independently of age, increasing blood pressure is strongly associated with overall stroke risk, and of all the main pathological types, including ischaemic stroke.322–324 The relationship between usual diastolic blood pressure and subsequent stroke is log–linear with no threshold below which stroke risk becomes stable, at least not within the ‘normal’ range of 70–110 mmHg (Fig. 6.17). The proportional increase in stroke risk associated with a given increase in blood pressure is similar at all levels of blood pressure. This risk almost doubles with each 7.5 mmHg increase in usual diastolic blood pressure in Western populations, and with each 5.0 mmHg in Japanese and Chinese populations.325–327 Because moderately raised blood pressure is so common in the middle-aged and elderly in developed countries, high blood pressure probably accounts for more ischaemic strokes than any other risk factor. The strength of the association between blood pressure and stroke is attenuated with increasing age, although

105

Approximate mean usual diastolic blood pressure (mmHg) in each of five categories, 1–5 Fig. 6.17 The risk of stroke related to the usual diastolic blood pressure in five categories, defined by usual baseline blood pressure, from a pooled analysis of seven prospective observational studies. Solid squares represent stroke risks relative to risk in the whole study population; their size is proportional to the number of strokes in each blood pressure category. The vertical lines represent 95% confidence intervals. Note the doubling scale of the y-axis. (With permission from MacMahon et al. 1990.325)

the absolute risk of stroke in the elderly is far higher than in the young.327 Nevertheless, hypertension is still a risk factor in the very elderly, although it is weaker because stroke may be associated with low blood pressure secondary to cardiac failure and other comorbid conditions.328 Moreover, in patients with bilateral severe carotid stenosis, stroke risk is higher at low blood pressures suggesting that aggressive blood-pressure lowering may be harmful in this group (section 16.3.1).329 The relationship between stroke and systolic blood pressure is possibly stronger than with diastolic pressure, and even ‘isolated’ systolic hypertension with a ‘normal’ diastolic blood pressure is associated with increased stroke risk.327,330–333 There is no doubt, as confirmed by the results of randomized controlled trials, that the relationship between increasing blood pressure and stroke risk is causal (section 16.3). However, it is not clear whether all types of stroke are prevented by reducing blood pressure, largely because in the clinical trials they have all been lumped together, haemorrhagic with ischaemic strokes of various types182,334–337 (section 16.3.2). Progression of carotid stenosis, at least as assessed by ultrasound, is slowed by treating hypertension.338

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 283

6.6 Risk factors for ischaemic stroke

numbers.311,344,371 So far, randomized trials have not shown that diabetic treatment definitely reduces the risk of stroke372 (section 16.9), although improved glycaemic control does lead to a reduction in overall rates of vascular complications of diabetes and a reduction in premature death. In the UKPDS study, overweight patients with type 2 diabetes who were treated with metformin had a 32% reduction in diabetes-related adverse outcomes and a 42% reduction in diabetes-related deaths.372

Hypertension seems to increase the risk of ischaemic stroke by increasing the extent and severity of atheroma (section 6.3) as well as the prevalence of ‘simple’ and ‘complex’ intracranial small vessel disease339–344 (section 6.4).

6.6.4 Smoking Cigarette-smoking is associated with approximately double the risk of ischaemic stroke in males and females, but less obviously in the elderly, and there is a dose–response relationship.345–350 There are not as many data on passive smoking and stroke as there are for coronary events where the association is surprisingly strong.351–354 As one would expect, most of the ultrasound and angiogram studies link carotid disease with smoking.344,355–357 Although cigar-smoking increases the risk of coronary events by about one-quarter, there are insufficient data to link either pipe- or cigar-smoking with stroke, perhaps because there are fewer people who still indulge in these habits.358 The risk of stroke gradually declines after stopping smoking so supporting a causal relationship, but a satisfactory randomized controlled trial proved impossible.345,350,359–361

6.6.6 Blood lipids Increasing plasma total cholesterol, increasing lowdensity lipoprotein-cholesterol and decreasing levels of high-density lipoprotein-cholesterol are all strong risk factors for coronary heart disease, whereas triglyceride levels are not. A long-term reduction of plasma cholesterol by 1 mmol/L should and does reduce the relative risk of coronary events by at least one-third,373 and also reduces coronary risk in the elderly.374–379 On the other hand, the relationship with ischaemic stroke is less clear. Very large systematic reviews of cohort studies have not revealed any association between all stroke types combined and increasing plasma total cholesterol at baseline, except perhaps under the age of 45 years326,380 (Fig. 6.18). Case–control studies provide less reliable measures of association because of their biases, particularly the changes in plasma lipids following stroke381 (section 6.8.1). Relating carotid intima-media thickness or stenosis with blood lipids is perhaps too far from the clinical consequences of atheroma to be relevant, or the studies have been too small to be reliable.310,312,342,344,382,383 Surprisingly therefore, the Heart Protection Study of cholesterol lowering in patients with known vascular disease or diabetes showed that simvastatin definitely reduced the risk of stroke on follow-up, although it did

6.6.5 Diabetes mellitus Any studies linking diabetes with fatal stroke will exaggerate the association, because diabetics who have a stroke are more likely to die of it than non-diabetics362 (sections 10.2.7 and 11.18.3). Indeed, the the risk of fatal stroke is higher in those with a higher HbA1C at diagnosis363 In fact, diabetes about doubles the risk of ischaemic stroke over and above confounding with hypertension and other risk factors.364–370 Diabetics also have thicker carotid arterial walls but the relationship with carotid stenosis is less clear, probably because of lack of patient

Fig. 6.18 Adjusted proportional risk of stroke (with 95% confidence intervals) by age and usual plasma cholesterol from a systematic review of 45 prospective observational studies. (With permission from the Prospective Studies Collaboration 1995.380)

Proportional stroke rate

1.2 1.0

1.2

1.2 1.1

1.1

1.0

1.0 1.1 1.0 0.9

0.8

0.7 0.7 0.6

4.5

5.0

5.5

Age at screening < 45 yr 45–64 65 yr+ 6.0

Usual total plasma cholesterol (mmol/L)

1.0

283

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 284

Chapter 6 What caused this transient or persisting ischaemic event?

not reduce the risk of recurrent stroke384,385 (section 16.4.2). This may have been because incident strokes occurred on average 4.6 years before the study onset and hence, at the time of the study, patients would have been at low risk of recurrent stroke but at high risk of coronary vascular disease. More recently, the Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) trial showed that atorvastatin in patients who had had a stroke or TIA within 1– 6 months before study entry did reduce stroke risk.386 However, statin treatment was associated with a small but significant increase in risk of haemorrhagic stroke. Interestingly, the same trend had been found in the Heart Protection Study (HPS) in the 3280 patients with previous stroke or TIA,384 in whom simvastatin 40 mg increased the risk of haemorrhagic stroke. Thus, the randomized evidence does suggest that there is a causal association between plasma LDL cholesterol and risk of ischaemic stroke – but more work is required to determine whether the increase in risk of haemorrhagic stroke is genuine. The marked contrast between coronary disease and ischaemic stroke in their association with plasma cholesterol is even more curious now it is clear that cholesterol lowering reduces the risk of myocardial infarction and stroke.384,386,387 It appears that the observational epidemiology has missed an ischaemic stroke–lipid connection, possibly because: • the seemingly negative association between increasing plasma cholesterol and intracranial haemorrhage has obscured a positive association with ischaemic stroke in studies where the pathological type of stroke was not accounted for (Fig. 6.19), which would tie in with the possibility that reducing plasma cholesterol with statins increases the risk of haemorrhagic stroke; • the over-representation of fatal, and therefore more likely to be haemorrhagic, strokes in some studies; • the narrow range of cholesterol levels examined in many studies; • the loss of stroke-susceptible individuals from the study population by prior death from coronary disease; • uncertainties about the effect of stroke itself on lipid levels in case–control studies; and • not differentiating ischaemic strokes likely to be caused by intracranial small vessel disease from those caused by large vessel atherothrombosis. Alternatively, the lack of a strong plasma cholesterol association with ischaemic stroke may be correct and perhaps the statins reduce stroke risk by some mechanism other than by cholesterol lowering.388,389 Both acute plaque-stabilizing effects of statins390 and neuroprotective effects391,392 have been postulated. Some support for potentially non-lipid effects comes from the

4.00 Non-haemorrhagic stroke (11 studies, 60 750 participants, 494 events) 2.00

1.00

0.50

Relative risk of stroke

284

0.25 4.0

4.00

4.5

5.0

5.5

6.0

Haemorrhagic stroke (11 studies, 60 750 participants, 404 events)

2.00

1.00

0.50

0.25 4.0

4.5

5.0

5.5

6.0

Approximate mean usual total plasma cholesterol concentration (mmol/L) Fig. 6.19 Overall adjusted relative risk (95% confidence intervals) of non-haemorrhagic and haemorrhagic stroke by usual plasma cholesterol from a systematic review of prospective observational studies in China and Japan. The size of the solid squares is proportional to the number of strokes in each cholesterol category. Note the doubling scale of the y-axis. (With permission from the Eastern Stroke & Coronary Heart Disease Collaborative Research Group, 1998.326)

apparent short-term effect of statins on risk of stroke after acute coronary syndromes.393 The inconsistency between the association between total or LDL cholesterol and ischaemic stroke and the effectiveness of statins in preventing ischaemic stroke also raises the possibility that other lipid subfractions that are affected by statins might be better predictors of ischaemic stroke.394

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 285

6.6 Risk factors for ischaemic stroke

Any association between plasma lipoprotein (a) and apolipoprotein E genotype with ischaemic stroke is still rather uncertain,395–397 but apolipoproteins (apo A1 and apo B), which make up the protein moiety of lipoproteins, do appear to be predictive of ischaemic stroke.398,399 Apo A1 is mainly found in HDL and apo B in LDL. Differences in prognostic value of apolipoproteins and LDL cholesterol are possible for several reasons. First, measurement of LDL cholesterol is an estimate of the mass of cholesterol in the LDL fraction of plasma, whereas measurements of apo B and apo A1 provide information on the total number of atherogenic (apo B) or anti-atherogenic (apo A1) particles. Second, the ratio of apo B : apo A1 best reflects the status of cholesterol transport to and from peripheral tissues. Third, unlike the relationships between total and LDL cholesterol and CHD, which weaken with age, apo B retains its predictive power in the elderly.400 Apo B is an established risk factor for coronary vascular events,400,401 and statin-mediated coronary event risk reduction has been attributed to apo B reduction.402 Data on apolipoproteins and ischaemic stroke risk are somewhat limited and conflicting, although there appears to be a relationship between raised apo B and an increased risk of ischaemic stroke, and an inverse correlation with apo A1 levels.398,399

6.6.7 Haemostatic variables Despite much effort, very few consistent associations have been found between coagulation parameters, fibrinolytic activity, platelet behaviour and vascular disease.403,404 Often these variables are altered by acute stroke so that most case–control studies are invalid, and it has not been practical to carry out very many long-term cohort studies. Plasma fibrinogen has a strong and consistent positive association with stroke and coronary events, including recurrent stroke.405,406 Cigarette-smoking is a confounding variable so the effect of cigarette-smoking on stroke may be mediated, at least in part, by increasing the fibrinogen level, thus accelerating thrombosis. Less important but still confusing confounding factors include age, hypertension, diabetes, hyperlipidaemia, lack of exercise, social class, social activity, season of the year, alcohol consumption and stress. It is not yet certain therefore whether increasing plasma fibrinogen really is a causal factor and, if so, whether it acts by increasing plasma viscosity or through promoting thrombosis. The confusion is compounded because there is no standard method for measuring plasma fibrinogen, it tends to rise after acute events including infections, it is not easy to lower plasma fibrinogen, and no satisfactory randomized controlled trials have been reported.344,381,405–412

Raised haematocrit is an uncertain risk factor for stroke and other acute vascular events, although it does seem to be associated with an increased case fatality in ischaemic stroke.413 The association is confounded by the fact that cigarette-smoking, blood pressure and plasma fibrinogen are all positively associated with haematocrit.414–416 Both raised haematocrit and raised plasma fibrinogen increase whole blood viscosity, another potentially causal risk factor.411 No randomized trials are available of lowering haematocrit, viscosity or fibrinogen. Raised plasma factor VII coagulant activity may be a risk factor for coronary events, and also polymorphisms of the factor VII gene, but there are very few data for stroke.417–419 Raised von Willebrand factor antigen is another possible risk factor for ischaemic stroke.417,420 Low blood fibrinolytic activity and high plasma plasminogen-activator type I are coronary risk factors and raised tissue plasminogen activator antigen may be associated with both coronary and stroke risk, perhaps because it is a marker of endogenous fibrinolytic activity.417,421–426 Abnormal platelet behaviour has not been convincingly linked with subsequent stroke in cohort studies, and any case–control studies after stroke have great potential for bias as stroke alters platelet function.427 Polymorphisms of platelet membrane glycoprotein IIIa were at first associated with stroke but, as seems to happen so often with genetic studies, then rejected on the basis of larger and more methodologically sound work.428

6.6.8 Plasma hom*ocysteine Because the rare inborn recessive condition of hom*ocystinuria is complicated by arterial and venous thrombosis (section 7.20.2), it is natural to imagine that mildly raised levels of plasma hom*ocysteine could be a risk factor for or even a cause of vascular disease in general. Raised plasma hom*ocysteine is positively associated with increasing age, abnormal lipids, smoking, diabetes, chronic renal failure and hypertension, and it is also raised in nutritional deficiencies and probably after stroke and myocardial infarction. The observational data linking coronary events, stroke, venous thromboembolism and carotid intima-media thickness and stenosis with increasing plasma hom*ocysteine (hyperhom*ocysteinaemia) are now fairly robust.429–435 In view of the strength of the epidemiological association between hom*ocysteine and risk of stroke, several trials of hom*ocysteine-lowering with folic acid and pyridoxine are being performed, some of which have now been reported, but so far the results are not promising and hyperhom*ocystinaemia cannot yet be regarded as a causal risk factor (section 16.8.1).

285

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 286

286

Chapter 6 What caused this transient or persisting ischaemic event?

6.6.9 Physical exercise Physical exercise somewhat reduces blood pressure, weight, plasma cholesterol, plasma fibrinogen and the risk of non-insulin-dependent diabetes mellitus and is associated with less cigarette-smoking.436,437 Therefore, not surprisingly, lack of exercise is associated with coronary events and there is also strong evidence of a similar association with stroke.438–443 So far there is insufficient evidence from randomized trials to be sure that deliberately increasing exercise levels, in sporting activities or as part of a generally healthy lifestyle, reduces the risk of vascular events in the general population444 (section 16.7.4). However, there is evidence (see below) that exercise can prevent progression to diabetes in patients with the metabolic syndrome as part of an overall lifestyle modification programme.445,446

6.6.10 Obesity Any relationship between obesity and stroke is likely to be confounded by the positive association of obesity with hypertension, diabetes, hypercholesterolaemia and lack of exercise, and the negative association with smoking and concurrent illness. Nevertheless, stroke is more common in the obese and so is overall mortality. How to measure obesity is itself somewhat controversial; a raised waist : hip ratio as a measure of central obesity, and perhaps change in body weight, may be stronger risk factors than the traditional measure of weight compared with height (see below).447–451

exercise;445,446 and angiotensin-converting enzyme (ACE) inhibitors or A2 blockers (ARB).455–457

6.6.12 Diet It is technically difficult to measure what people eat and how they cook it, relate any particular diet to events such as stroke that occur years later, and to be sure that any association is not better explained by a confounding variable.458 For example, people who eat a lot of fruit tend to have a generally ‘healthy’ lifestyle, not to smoke, and they may use less salt.459 Salt, by increasing systemic blood pressure, is probably associated with increased stroke risk but this issue is still controversial.460–463 In observational studies, diets low in the following may be associated with a raised risk of coronary disease and stroke:464 • potassium465–467 • calcium466,468 • fresh fruit and vegetables469,470 • fish469–473 • antioxidants such as vitamin E,474–476 vitamin C,477–479 beta-carotene480–482 and flavonols.483,484 On the whole, randomized trials of dietary interventions and vitamin supplementation have been disappointing (section 16.7.3). Either the theory is wrong, or the interventions were too little, too late, for insufficient time or the trials were too small.385,455,485–489 The effect of a vegetarian diet is unclear.490 Despite earlier enthusiasm, it now appears that there is no association between coffee consumption and vascular disease.491–494

6.6.11 The metabolic syndrome Obesity, specifically visceral obesity, plays a key role in the development of the metabolic syndrome and type 2 diabetes. Visceral adipocytes are insulin-resistant and highly metabolically active and promote dyslipidaemia, hypertension and reduced systemic thrombolysis, as well as leading to a relatively pro-inflammatory state.452 The constellation of metabolic abnormalities including central obesity, decreased HDL, raised triglycerides and blood pressure, and hyperglycaemia is known as the metabolic syndrome, and this is associated with a threefold increased risk of type 2 diabetes and a twofold increase in cardiovascular risk, including stroke.453,454 It is thought to be the main driver for the modern-day epidemic of diabetes and vascular disease. As well as the prevention of acute vascular events in patients with the metabolic syndrome, primary prevention of which is mostly the same as for other groups,453,454 one additional aim should be to prevent progression to frank diabetes. Two strategies have been shown to be effective in this regard – lifestyle modification with diet and

6.6.13 Alcohol The relationship between alcohol, ischaemic stroke and carotid atheroma is complex and may be U-shaped.495 While heavy alcohol consumption may be an independent and perhaps causal risk factor, it seems that modest consumers are protected to some extent compared with abstainers.496–504 Whether ‘binge’ drinking is associated with stroke is uncertain505,506 but it is possible that irregular drinking carries a higher risk.507 Much of the confusion in relating alcohol consumption with stroke is because:508 • people are not always truthful about their alcohol consumption; • it is difficult to measure alcohol consumption accurately, particularly over time; • there are varied ways of expressing alcohol consumption (per day, per week, grams, units, number of drinks, regular, binge, etc.); • different types of alcoholic drinks may have different effects;

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 287

6.6 Risk factors for ischaemic stroke

• pattern of drinking behaviour may change over time; • combining ex-drinkers, some of whom may have given up drinking because of symptoms of vascular or other diseases, with lifetime non-drinkers in the analyses; • the biases inherent in case–control studies; • publication bias; • small numbers and so imprecise estimates; • confounding with cigarette-smoking, hypertension and deprivation which are positively related, and with exercise which is negatively related with alcohol consumption; • confounding with unknown or unmeasurable factors which might link no drinking or heavy drinking with an excess risk of vascular disease (e.g. healthy vs risky lifestyle); • lumping ischaemic and haemorrhagic strokes together. It is difficult to think of the same biological reason for ischaemic stroke in non-drinkers and heavy drinkers. Possible causal explanations, at least in heavy drinkers, are that alcohol almost certainly increases the blood pressure; traumatic arterial dissection in the neck during an alcohol-related injury; dehydration, hyperviscosity and platelet activation perhaps; sleep apnoea and hypoxia; and that alcohol can cause both atrial fibrillation and cardiomyopathy with embolism to the brain.509–512 Perhaps it would be more productive to ask why modest drinking is protective, if it really is. Increased plasma high-density lipoprotein-cholesterol and lower fibrinogen levels are possibilities.513 It is most unlikely that a randomized trial of modest alcohol consumption, in the hope of reducing the risk of vascular disorders without increasing the risk of other disorders, will ever be feasible. Therefore, we are left trying to interpret, with difficulty, the available observational data (section 16.7.2).

6.6.14 Ethnicity There are no good population-based data on stroke incidence in developing countries, with yawning gaps particularly in South and South-East Asia and subSaharan Africa48,514,515 (section 18.2.2). Nonetheless, and notwithstanding the difficulty in defining ethnicity, stroke incidence does seem greater in black than white people living in Western countries, probably both ischaemic stroke and intracerebral haemorrhage.516–520 The high prevalence of hypertension, diabetes, sickle cell trait and social deprivation in black people may be part of the explanation.521–523 South Asian populations in the UK have a high stroke mortality as well as a high prevalence of coronary disease, central obesity, insulin resistance, diabetes mellitus and other risk factors.524,525 This is in part because they are genetically more at risk

than white populations by virtue of higher serum lipoprotein (a) concentrations.526–529 Also, Maori and Pacific people living in New Zealand have a higher stroke risk than white New Zealanders.530 Moreover, the fall in stroke incidence that has occurred in white New Zealanders has not been mirrored by any similar fall in the Maori and Pacific population.531

6.6.15 Specific genes for common types of ischaemic stroke Based on the assumption that at least some of the risk for stroke is genetic, large numbers of studies have now been done or are being done in an attempt to identify the genes involved.532 However, on the whole, methodological rigour – and so the reliability of the results of these studies – has been limited. It seems likely that the genetic component of stroke risk is modest and that many (probably hundreds) of genes are involved, each one contributing only a small increased risk. But, studies so far have generally not been large enough to detect reliably the sort of relative risks (i.e. the risk ratio comparing one genotype with another at a particular genetic locus) that might realistically be expected (probably about 1.2 to 1.5). Other methodological limitations include: poor choice of controls in case–control studies and/or failure to use methods designed to detect and control for selection bias; inadequate distinction between the different pathological types and subtypes of stroke for which genetic influences may differ; failure to replicate positive results in an independent and adequately sized study; and testing of multiple genetic or subgroup hypotheses with no adjustment of p-values for declaring statistical significance.533,534 Candidate gene studies Most genetic studies so far have been candidate gene studies, in which the frequency of different genotypes at a specific locus or loci within a gene or genes thought likely to be in some way connected with stroke risk are compared between stroke cases and stroke-free controls. Candidate genes have generally been selected on the basis of their known or presumed involvement in the control of factors or pathways likely to influence stroke risk: blood pressure, lipid metabolism, inflammation, coagulation, hom*ocysteine metabolism and so on.535,536 Rigorous meta-analyses of candidate gene studies, both in stroke and other vascular diseases such as coronary heart disease, have highlighted the various methodological issues discussed above. In particular, they have drawn attention to the problems of small study size (most studies have included a few hundred or less, rather than the required 1000 or more cases and controls),

287

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 288

288

Chapter 6 What caused this transient or persisting ischaemic event?

leading to the potential for publication bias (i.e. small positive studies are far more likely to be published and available for inclusion in meta-analyses than small negative ones), with very large studies tending to detect either much smaller relative risks or no evidence of altered risk at all.534,537,538 Large numbers of candidate gene studies have together identified a handful of genes (those encoding factor V Leiden, methylenetatrahydrofolate reductase, prothrombin and angiotensin-converting enzyme) which, on the basis of results from metaanalyses, seem likely to influence risk of ischaemic stroke modestly.536 However, since none of the contributing studies was large enough to be independently reliable, occult publication bias might still account for the apparent effects of these few genes, and confirmation in further very large studies is needed. Linkage studies As yet there have been far fewer stroke genetics studies that use more traditional genetic study designs, based on collecting information and DNA from related individuals with and without the disease of interest. This is at least partly because family members of stroke patients (who of course tend to be elderly) are often no longer still alive, and so obtaining information and samples for DNA extraction from large enough numbers of relatives is challenging.539 There are some exceptions, though. For example, the Icelandic population has excellent genealogical records allowing the construction of complex pedigrees, along with stored DNA and medical histories on large numbers of people in the population. Family or pedigree-based genetic study designs allow for linkage analysis, which has the attraction of not being driven by hypotheses about specific candidates, but instead sets out to identify which regions of the genome may influence disease risk, through analysing the degree of sharing of genetic information (at markers across the whole genome) between related individuals with and without disease. The Icelandic deCODE group has carried out a number of linkage studies to identify regions of the genome likely to influence risk of various common diseases, including stroke (see http://www.decode.com/ Population-Approach.php). Through this process, two candidate genes for ischaemic stroke, encoding the enzymes phosphodiesterase-4D and arachidonate 5-lipoxygenase-activating protein ALOX5AP, have been identified.540 There is still some debate, however, about their influence in non-Icelandic populations, since their effect on ischaemic stroke risk has been confirmed in some, but not all, subsequent studies, and the confirmatory studies have not been free of the methodological problems discussed above.541,542

Whole genome association studies The combination of technological developments allowing rapid throughput genotyping at multiple loci, the attraction of non-hypothesis-driven genetic studies, and the recognized limitations of traditional linkage approaches have together led to a new wave of genomewide association studies, where multiple (up to 500 000 or more) polymorphisms across the genome are genotyped and compared in cases and controls, looking for loci where significant differences may suggest genetic influences on disease risk. The statistical considerations required for such studies are considerable, since huge numbers of potential genetic risk factors, albeit not completely independent of each other (because segments of the genome tend to be transmitted together during meiosis – linkage dysequilibrium), are tested simultaneously.543,544 Thus, there is great potential for missing real genetic effects or for finding spurious ones. Nonetheless, this approach has had a few (perhaps lucky) successes in common diseases; one small, preliminary study in stroke has now been published, and no doubt others will follow.545 Genetic studies of other vascular diseases, known vascular risk factors and intermediate phenotypes The results of genetic studies of other vascular diseases (in particular coronary heart disease) may also prove useful in identifying likely genetic candidates for ischaemic stroke. In addition, since much of the genetic risk of stroke seems likely to be mediated through alreadyknown risk factors such as cholesterol and blood pressure (sections 6.6.3 and 6.6.6), studies of these may also provide useful information about likely stroke risk genes. Finally, genetic studies have started to emerge of so-called ‘intermediate phenotypes’, markers of predisposition to stroke (or other vascular diseases), which can be measured in large numbers of subjects both with and without vascular risk factors or disease. These intermediate phenotypes include carotid intima media thickness and leukoaraiosis (measured or graded on CT or MR brain scans), and both linkage and candidate gene approaches have been used.533,546 As with studies of clinically manifest disease, methodological rigour and adequate study size are crucial, and so systematic review and meta-analysis approaches will be helpful in interpreting the large numbers of studies and quantities of data emerging.547 We are still far from being able to quantify the genetic influences on stroke, or to identify reliably the genes involved. One serious challenge is that the available laboratory technology is currently outstripping our ability to properly analyse and interpret the results. In

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 289

6.6 Risk factors for ischaemic stroke

the future, only careful integration of information from all of the different approaches described above, together with the results of laboratory studies of gene expression and function, from the molecular to the animal model level, will shed useful light on the genetic influences on stroke, hopefully allowing us to understand better the way in which already-known risk factors influence stroke, and reliably to identify new ones.

6.6.16 Social deprivation Social deprivation, low socioeconomic status and unemployment are all inextricably linked and associated with increased stroke risk.548,549 This association is partly caused by a higher prevalence of vascular risk factors, stress and adverse health behaviours, such as smoking, poor diet and lack of exercise in deprived populations.550 For example, these factors appear to account for most of the north–south gradient in stroke mortality in the UK.551 Also, it has been suggested that poor health and nutrition in utero or infancy are associated with the development of vascular risk factors such as hypertension and adult vascular disease, including stroke.552–555 It is extremely difficult to disentangle early from later life influences on health and so, not surprisingly, the ‘early life’ or ‘Barker’ hypothesis has been refuted by some.556–559 The possibility of reducing stroke risk by tackling social deprivation is yet another reason to add to the many others to support what has to be political rather than medical action.

6.6.17 Infection There has been interest in the notion that infection contributes to the development of atheroma, the progression of the atherothrombotic plaque, and perhaps plaque instability. However, the evidence from observational epidemiological studies of infections in general,560 and of chronic dental infection,561 together with serological evidence of specific infectious agents, such as Chlamydia pneumoniae, Helicobacter pylori and cytomegalovirus, is not very convincing, even for coronary disease where there are – as usual – more data than for stroke.562–569 The generalized inflammatory response in patients with coronary artery disease may be too non-specific to be convincing evidence of infection, and there may well be the familiar problems of confounding and publication bias.412,563,564,570 Nonetheless, the fact that evidence of chlamydial infection can be found in atherosclerotic plaques is certainly interesting although reverse causality is a possibility, i.e. atheroma becomes infected more often than normal arterial wall.571 An early randomized trial to eliminate chlamydia was far

too small to be reliable despite being published in a high-impact journal.572 Further antibiotic intervention trials are ongoing.573,574

6.6.18 ‘Stress’ It is part of folklore that stress causes strokes: ‘I was so upset, I nearly had a stroke’. Indeed, there are some striking anecdotes. This 82-year-old lady was admitted for investigation of anaemia and hepatomegaly. She was found to have multiple liver metastases. When she was told the news, she stopped speaking and never spoke again. It took a day or so before the medical staff realized that her mute state was not ‘psychological’ but that she was aphasic. A brain CT scan showed a left cortical infarct. She died some days later from her malignancy. However, showing that long-term psychological stressors are associated with stroke, or that shorter-term ones precipitate stroke, is not easy. There is some evidence that severely threatening life events,575,576 anxiety and depression,577 high levels of anger expression578,579 and psychological stress577,580,581 may trigger the onset of stroke, perhaps in people already at risk of stroke.

6.6.19 Non-stroke vascular disease Because atheroma in one artery is likely to be accompanied by atheroma in others (section 6.3.1), and because embolism from the heart is a common cause of ischaemic stroke (section 6.5), non-stroke vascular disorders are associated with (i.e. are risk factors for) ischaemic stroke and transient ischaemic attack (Table 6.8). Coronary heart disease (e.g. angina or myocardial infarction) has been repeatedly associated with an increased risk of stroke in postmortem,582,583 twin,584 case– control190,585–587 and cohort studies.170,194,330 Therefore, it is not surprising that electrocardiogram abnormalities and cardiac failure are also associated with increased risk of stroke because they both so often reflect coronary Table 6.8 Degenerative vascular disorders outside the head associated with an increased risk of ischaemic stroke and transient ischaemic attack. Myocardial infarction/angina Cardiac failure Left ventricular hypertrophy Atrial fibrillation Cervical arterial bruit/stenosis Peripheral arterial disease

289

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 290

290

Chapter 6 What caused this transient or persisting ischaemic event?

disease and hypertension, as is left ventricular hypertrophy.330,588–592 Atrial fibrillation is considered in section 6.5.1. Cervical bruits (carotid or supraclavicular) are generally caused by stenosis of the underlying arteries (section 6.7.7), mostly as a result of atheroma, and so become more common with age; about 5% of asymptomatic people over the age of 75 years have bruits.593–595 Carotid bruits are clearly a risk factor for ischaemic stroke, but not necessarily in the same arterial territory as the bruit, and also for coronary events, because atheroma of one artery is likely to be accompanied by atheroma of other arteries in the same predisposed individual.596,597 The risk of these various vascular events increases with carotid stenosis severity (section 16.11.5). Atheroma affecting the leg arteries is associated with cerebrovascular and coronary disease in the same individuals so often that it is not surprising that claudicants have an increased risk of stroke and other serious vascular events.194,598–602 Little seems to be known about the prevalence of abdominal aortic aneurysms in ischaemic stroke/TIA patients. It is said to be quite high, about 10–20%, depending on the selection of patients and the size criterion for what constitutes an aneurysm.602–605 The incidence of aneurysm rupture in the general population is 5–10 times higher in men than in women,316 and so there might be an argument for screening men presenting with transient ischaemic attack or stroke, particularly perhaps if the cause is thought to be atherothromboembolism and bearing in mind the advantages of screening even in ‘normal’ elderly men from the Multicentre Aneurysm Screening Study (MASS).606

6.6.20 Other risk factors Innumerable other risk factors for coronary heart disease, and/or ischaemic stroke, have been suggested and supported by varying degrees of evidence. However, it may take many years, even decades, before a particular risk factor is accepted as causal, as has been the case for hypertension and smoking. So often initial discoveries of ‘new’ risk factors are found subsequently to be spurious.607 Even by 1981, 246 risk factors had been counted and by now there must be many more.608 Some may or may not be important in the causal pathway to stroke but, even if not, they may still conceivably be helpful in predicting future stroke in individuals and populations. This distinction is important: for example, claudication is clearly not on the causal pathway to stroke but is so strongly associated with stroke risk that it is included in many statistical models which predict later stroke (section 16.2).

6.7 From symptoms, signs and clinical syndrome to cause

Once the diagnosis of ischaemic stroke or transient ischaemic attack (TIA) has been made, the cause of the ischaemic event needs to be established since this has implications for treatment, prognosis and risk of recurrence. Four clinical syndromes, based on the history and examination, can be identified: total anterior circulation infarction, partial anterior circulation infarction, lacunar infarction and posterior circulation infarction. The clinical syndrome reasonably predicts the site and size of the brain lesion, which takes one a long way towards the likely cause of the ischaemic event (Fig. 6.20). Clinical localization is easier if the patient has had an established stroke with stable physical signs rather than being in the very early stages when the signs are still evolving (brain attack), or has had a TIA and any signs have disappeared. However, stroke localization based on the history and examination is by no means infallible: in about onequarter of cases where a recent lesion is visible on brain imaging, it is not in the expected place to explain the clinical syndrome.133 For example, although most pure motor strokes are caused by a lacunar infarct as a result of small vessel disease, in a few cases the CT or MR scan shows striatocapsular infarction which is likely to be caused by middle cerebral artery occlusion with good cortical collaterals (sections 4.2.2 and 6.7.2; fig. 6.20e). And in a recent study of 150 patients with minor stroke and a single acute lesion on DWI, agreement on localization to anterior or posterior vascular territory between DWI and the clinical judgements of three independent neurologists was only moderate (kappa = 0.5).609 Thus, brain imaging, in particular new techniques such as MR DWI, can and if possible should be used to refine lesion localization and thus aid in the search for the cause of the ischaemic stroke/TIA in an individual patient. In acute stroke, computed tomography or magnetic resonance imaging should be neither the first nor only way to classify patients on the basis of the size and site of any ischaemic brain lesion. Imaging is used to confirm and refine where the symptoms and neurological signs – in other words, the clinical syndrome – suggest the lesion is. From there, it is possible to narrow down the potential causes of the infarct.

6.7.1 Total anterior circulation infarction The acute ischaemic stroke clinical syndrome of a total anterior circulation infarction (TACI) comprises

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 291

6.7 From symptoms, signs and clinical syndrome to cause

a hemiparesis, with or without hemisensory loss, hom*onymous hemianopia, and a new cortical deficit such as aphasia or neglect. It is a good predictor of infarction of most of the middle cerebral artery (MCA) territory on brain CT, as a consequence of occlusion of either the MCA mainstem (or proximal large branch) or the internal carotid artery (ICA) in the neck (section 4.3.4) (Fig. 6.20c). Occasionally, a TACI can be caused by occlusion of the posterior cerebral artery, but the hemiparesis is usually rather mild. The cause of the arterial occlusion therefore is usually in the heart (e.g. embolism as a consequence of atrial fibrillation, recent myocardial infarction, etc.), or it is atherothrombosis complicated by embolism or occasionally propagating thrombosis of the ICA, or embolism from the aortic arch. Therefore, if the heart is clinically normal (history, examination, chest X-ray and electrocardiogram) and if there is no evidence of arterial disease in the neck (bruits, palpation perhaps, but mainly duplex sonography), then it is important to consider rarities, for example, infective endocarditis (echocardiogram, blood cultures) and carotid dissection (MR angiogram, and certainly recheck for past history of neck trauma, or associated neck or face pain). Although transcranial Doppler may confirm an MCA mainstem occlusion (but not if it has already recanalized), it will not help much in the search for a cause.134 A catheter angiogram might be diagnostically helpful if it could be justified on the basis of changing the patient’s management (e.g. traumatic carotid dissection could lead to later litigation, fibromuscular dysplasia could stop the search for other explanations, a giant aneurysm with contained thrombus might be surgically treatable, and so on). MRI and MR angiography, and CT angiography, are now widely and more appropriately used to show lesions, such as dissection and aneurysms, but they are still not always easily available in patients in the acute stage of stroke (sections 6.8.3 and 6.8.4).

6.7.2 Partial anterior circulation infarction Partial anterior circulation infarction (PACI) is a more restricted clinical syndrome with only two out of the three components of a TACI; or a new isolated cortical deficit, such as aphasia; or a predominantly proprioceptive deficit in one limb; or a motor/sensory deficit restricted to one body area or part of one body area (e.g. one leg, one hand, etc.) (section 4.3.5). This syndrome is reasonably predictive of a restricted cortical infarct caused by occlusion of a branch of the middle cerebral artery (MCA) or, much less commonly, of the anterior cerebral artery, as a result of embolism from the heart or from proximal sites of atherothrombosis (usually the

carotid bifurcation), or to any other cause of a TACI (Fig. 6.20b). Investigation is therefore similar to that for the patient with a TACI, except it is usually easier because the patient is fully conscious and less neurologically impaired. However, investigation must be quicker because of the higher risk of early recurrence (section 16.2.3)79,80 and because the patient has more to lose from a recurrence which might, next time, be a TACI. The potential for secondary prevention must be considered, particularly eligibility for carotid endarterectomy, and this requires early duplex sonography to find any severe carotid stenosis (section 6.8.5), and eligibility for anticoagulation if the patient is in atrial fibrillation. Transcranial Doppler (TCD) is unlikely to demonstrate the blocked cerebral artery because this is almost always distal to the MCA mainstem, at a point where TCD is not particularly sensitive.134 Occasionally, however, patients with a large PACI, but falling short of the full definition of a TACI, do have MCA mainstem occlusion, presumably because good collateral flow to the margins of the central infarcted area of brain restricts the clinical syndrome. This is particularly likely with striatocapsular infarction, which usually presents as a PACI (Figs 4.14, 6.20e). Some other PACI syndromes are caused by infarction in the centrum semiovale (section 4.2.2) and in boundary zones (section 6.7.5). Anterior choroidal artery infarcts may also present as a PACI (or a lacunar) syndrome and they seem to be caused by either embolism from proximal sites or intracranial small vessel disease (section 4.2.2). Total and partial anterior circulation infarction/transient ischaemic attacks are usually caused by occlusion of the mainstem or a branch of the middle cerebral artery, by occlusion of the anterior cerebral artery, or by occlusion of the internal carotid artery. Such occlusions are usually caused by embolism from the heart, embolism from proximal arterial sites of atherothombosis (the internal carotid artery origin, the aortic arch, etc.), and sometimes by thrombotic occlusion of severe internal carotid artery stenosis.

6.7.3 Lacunar infarction Lacunar syndromes, the vast majority of which are ischaemic rather than due to intracerebral haemorrhage, are almost always caused by small, deep, infarcts more likely to be seen on MRI than brain CT (section 4.3.2) (Fig. 6.20d). These small, deep, infarcts are mostly caused by a vasculopathy affecting the small perforating arteries of the brain, and not by embolism from proximal

291

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 292

292

Chapter 6 What caused this transient or persisting ischaemic event?

(a)

D C

C B

(b)

(c)

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 293

6.7 From symptoms, signs and clinical syndrome to cause

(d)

(e)

Fig. 6.20 (opposite & above) Various patterns of arterial occlusion cause different types of ischaemic stroke. Left-hand columns: axial CT brain scan at the level of the basal ganglia; middle columns, diagram to correspond with the CT brain scan with the area of infarction shaded; right-hand columns: diagram of the middle cerebral artery (MCA) and anterior cerebral arteries on a coronal brain section with the area of infarction shaded. A, main trunk of MCA; B, lenticulostriate perforating branches of the MCA; C, cortical branches of the MCA; D, cortical branches of the anterior cerebral arteries. (a) Normal arterial anatomy and CT scan. (b) Occlusion – usually embolic (straight arrow) from heart, aorta or internal carotid artery – of a

cortical branch of the MCA and restricted cortical infarct on CT (curved arrows); partial anterior circulation infarction (PACI). (c) Occlusion – usually embolic (straight arrow) as in (b) above – of MCA mainstem to cause infarction of entire MCA territory (curved arrows); total anterior circulation infarction (TACI). (d) Occlusion of one lenticulostriate artery (straight arrow) to cause a lacunar infarct (curved arrow); lacunar infarction (LACI). Note that the patient has an old lacunar infarct in the opposite hemisphere. (e) Occlusion of the MCA mainstem (straight arrow) but with good cortical collaterals from the anterior and posterior cerebral arteries to cause a striatocapsular infarct (curved arrows).

293

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 294

294

Chapter 6 What caused this transient or persisting ischaemic event?

arterial sources or the heart (section 6.4) and thus have a low risk of early recurrence71 There is not therefore the same urgency to rule out a cardiac source of embolism or severe carotid stenosis as there is for a partial anterior circulation infarction (PACI) (section 6.7.2). The vast majority of lacunar stroke syndromes are caused by ischaemia rather than haemorrhage. Most ischaemic lacunar strokes are the result of a small, deep, not a cortical, infarct. These small, deep infarcts are usually within the distribution of a small, perforating artery. The underlying vascular pathology is probably ‘complex’ small vessel disease, which differs from atheroma, but sometimes atheroma of the parent artery may occlude the mouth of the perforating artery. Lacunar infarcts are seldom caused by embolism from the heart or from proximal arterial sources. The capsular warning syndrome has a rather characteristic pattern. Over hours or days, there is cluster of transient ischaemic attacks (TIAs), consisting typically of weakness down the whole of one side of the body without any cognitive or language deficit (i.e. pure motor lacunar TIAs). These may be followed within hours or days by a lacunar infarct in the internal capsule. This syndrome is presumably caused by intermittent closure of a single lenticulostriate or other perforating artery, followed by complete occlusion, and one is unlikely to find a proximal arterial or cardiac cause, as in any other type of lacunar ischaemic stroke.346

6.7.4 Posterior circulation infarction Ischaemia and infarction in the brainstem and/or occipital region is aetiologically more heterogeneous than in the other three main clinical syndromes61,62 (section 4.3.3). Emboli from the heart may reach a small artery supplying the brainstem (e.g. superior cerebellar artery) to cause a fairly restricted deficit, block the basilar artery to produce a major brainstem stroke, travel on to block one or both posterior cerebral arteries to cause a hom*onymous hemianopia or cortical blindness, or any combination of these deficits. Similarly, embolism from the vertebral artery (as a result of atherothrombosis usually, but sometimes another disorder such as dissection) or from atherothrombosis of the basilar artery, aortic arch or innominate or subclavian arteries produces exactly the same neurological features as embolism from the heart.63,165,610 Even embolism from the carotid territory can, in some individuals with a dominant posterior communicating artery or a persistent trigeminal artery, cause occlusion of the posterior cerebral artery and even

brainstem infarction.611,612 Basilar occlusion, usually as a result of severe atherothrombotic stenosis, is likely to produce massive brainstem infarction. Obstruction, usually by atherothrombosis, of the origin of the small arteries arising from the basilar artery, can produce restricted brainstem syndromes, as can ‘complex’ small vessel disease within the brainstem; certainly some patients with a lacunar syndrome have a small infarct in the brainstem. A posterior circulation infarction (POCI) does not therefore provide much of a clue to the cause of the ischaemic event. An exception is the patient with simultaneous brainstem signs and a hom*onymous hemianopia, where embolism from the heart or a proximal artery must be the likely cause and not small vessel disease. Posterior circulation infarction/transient ischaemic attack can be due to almost any cause of cerebral ischaemia, which makes it very difficult to be certain of the exact cause in an individual patient if one knows no more than just the lesion localization. Cerebellar ischaemic strokes (section 4.2.3) are mostly caused by embolism from the heart, vertebral and basilar arteries, or by atherothrombotic occlusion at the origin of the cerebellar arteries; some are said to result from low blood flow alone.613–616 Thalamic infarcts (section 4.2.3) can be caused by: ‘complex’ small vessel disease affecting one of the small perforating arteries; atheromatous occlusion of these same arteries where they arise from the posterior cerebral and other medium-sized arteries; and occlusion of these latter arteries by embolism from the heart, basilar, vertebral and other proximal arterial sites.617–619

6.7.5 Ischaemic strokes and transient ischaemic attacks caused by low cerebral blood flow The pressure gradient across, and blood flow through, large arteries is not affected until their diameter is reduced by more than 50%, often not until by much more.620–622 Not surprisingly, therefore, even if there is severe disease of the carotid or vertebral arteries, cerebral perfusion pressure is usually normal. However, in some patients, as the stenosis becomes more severe, flow does fall, and eventually cerebral vasodilatation (autoregulation) cannot compensate for the low cerebral perfusion pressure. Regional cerebral blood flow (CBF) then falls, particularly if the collateral circulation is compromised because the circle of Willis, for example, is incomplete or diseased53,66,623–625 (section 4.2.2). At this stage of exhausted cerebral perfusion reserve, the ratio of cerebral blood flow : cerebral blood volume falls below about 6.0,

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 295

6.7 From symptoms, signs and clinical syndrome to cause

oxygen extraction fraction starts to rise and stroke risk probably also rises53,626–629 (section 12.1.2). Using transcranial Doppler (TCD) to demonstrate impaired cerebrovascular reactivity to a chemical rather than perfusion challenge is an indirect but more practical alternative to positron emission tomography (PET), but the correlation is not perfect624 (section 6.8.9). Isotopic measurement of the mean cerebral transit time,630 gradient echo and perfusion-weighted MRI,625,631 MR angiography,632 CT perfusion633,634 and near-infrared spectroscopy635 are other possibilities. Therefore, although the notion that ischaemic strokes may be caused by ‘hypotension’ goes back many years, low regional CBF alone is not particularly common and cannot easily explain more than a small fraction of strokes. Severe arterial stenosis or occlusion, or good evidence of a fall in systemic blood pressure just before onset, are simply not present in most ischaemic stroke cases.636 Most ischaemic strokes and transient ischaemic attacks (TIAs) must be caused, we believe, by embolic or in situ acute (usually thrombotic) occlusion of an artery to the brain causing blood flow to be suddenly cut off, so causing ischaemia in its territory of supply. Naturally, at times, focal ischaemia could also be caused just by low flow without acute vessel occlusion but usually only distal to a severely stenosed or occluded internal carotid (ICA) or other artery. This is where the vascular bed is likely to be maximally dilated and therefore where the brain is particularly vulnerable to any fall in perfusion pressure (even more so if arteries carrying collateral blood flow are also diseased). Under these circ*mstances, a small drop in systemic blood pressure might cause transient or permanent focal ischaemia without any acute occlusive event. Under normal circ*mstances quite a large fall in blood pressure does not cause cerebral symptoms, provided it is transient. This is because of autoregulation of CBF (section 12.1.2). If it does, the symptoms are much more likely to be nonfocal (faintness, bilateral blurring of vision, etc.) than focal637 (sections 3.2.1 and 3.4.12). Boundary-zone ischaemia and infarction Sometimes, ischaemia occurs not within but between major arterial territories in their boundary zones (section 4.2.4). Because this is where perfusion pressure is likely to be most attenuated, it is conceivable that ‘low flow’ as a result of low perfusion pressure, as well as the more common acute arterial occlusion caused by embolism, can cause ischaemia in these areas.638 The alternative term of watershed infarction is a misnomer based on geographical ignorance. A watershed is the line separating the water flowing into different river basins (i.e. the

high ground between two drainage areas), which is quite different from the pattern of arterial supply where flow is from larger to smaller vessels. The evidence that at least some boundary-zone infarcts are caused by low flow rather than acute arterial occlusion is that sudden, profound and relatively prolonged hypotension (e.g. as a result of cardiac arrest or cardiac surgery) sometimes causes infarction bilaterally in the posterior boundary zones, between the supply territories of the middle cerebral artery (MCA) and the posterior cerebral artery in the parieto-occipital region. The clinical features include cortical blindness, visual disorientation and agnosia, and amnesia. Unilateral posterior boundary-zone infarction causes contralateral hemianopia, cortical sensory loss and, if in the dominant hemisphere, aphasia. Also, distal to severe carotid stenosis or occlusion, unilateral infarction is well recognized in the anterior boundary zone between the supply territories of the MCA and anterior cerebral artery in the frontoparasagittal region, but this does not necessarily mean that the cause was low flow rather than embolism. The clinical features are contralateral weakness of the leg more than the arm and sparing the face, some impaired sensation in the same distribution, and aphasia if in the dominant hemisphere57,162,639 There is an internal or subcortical boundary zone in the corona radiata and centrum semiovale, lateral and/or above the lateral ventricle. This lies between the supply of the lenticulostriate perforating branches from the MCA trunk, and the medullary perforating arteries which arise from the cortical branches of the MCA and the anterior and, perhaps, posterior cerebral arteries. Infarction can occur within this internal boundary zone, usually causing a lacunar or partial anterior circulation syndrome, in association with severe carotid disease and sometimes an obvious haemodynamic precipitating cause.640 The diagnosis of ‘low flow’ as the cause of ischaemic strokes and TIAs It would be simplistic to presume that all ischaemic strokes are caused by acute arterial occlusion. However, the definitive separation of stroke resulting from ‘low flow’ from acute arterial occlusion is far from easy. It is probably best inferred from the circ*mstances surrounding the onset of the symptoms, and to some extent by their nature, and not very much from either the neurological signs or the site of any visible infarct on brain imaging. Unfortunately, any clinical guideance to ‘low flow’ ischaemic episodes cannot be validated against a ‘gold standard’ because at present there is none. Most ischaemic strokes and TIAs occur ‘out of the blue’ with no precipitating activity. However, on the basis of a

295

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 296

296

Chapter 6 What caused this transient or persisting ischaemic event?

number of convincing case reports, a fall in cerebral perfusion pressure as the cause – mostly resulting from a fall in systemic blood pressure – should be suspected if the symptoms start under certain circ*mstances: • on standing or sitting up quickly, even if postural hypotension cannot be demonstrated in the clinic; • immediately after a heavy meal; • in very hot weather; • after a hot bath or warming the face; • with exercise, coughing or hyperventilation; • during a Valsalva manoeuvre, but paradoxical embolism is another possibility; • during a clinically obvious episode of cardiac dysrhythmia (chest pain, palpitations, etc.), but embolism from the heart is also possible; • during operative hypotension (section 7.18); or • if the patient has recently been started on or increased the dose of any drug likely to cause hypotension, such as calcium blockers or vasodilators. In addition, there is usually very obvious evidence of severe arterial disease in the neck, i.e. bruits and/or absent pulsations.641–653 Transient ischaemic attacks (TIAs) caused by low flow may be atypical ‘limb shaking TIAs’ and tend to develop over minutes rather than seconds. These consist of stereotyped jerking and shaking of one arm and/or leg contralateral to the cerebral ischaemia and so are easily confused with focal motor seizures (section 3.3.4), or there is monocular or binocular visual blurring, dimming,

.. . .................................... Variable vascular supply ........................... or ‘boundary-zones’

fragmentation or bleaching, often only in bright light (section 3.3.6). Movement disorders including chorea have also been reported but involvement of the face is thought not to occur, perhaps because the blood flow to the facial motor cortex is not significantly compromised. ‘Limb shaking TIAs’ typically occur on standing up and may be abolished by a reduction in any hypertensive therapy, or carotid surgery, since they are frequently associated with severe occlusive carotid disease. There may be additional non-focal features, such as faintness, mental vagueness or even loss of consciousness.654–659 Low-flow ischaemic oculopathy is discussed in section 3.3.6. It is very important to note that boundary-zone infarction on brain imaging (or at postmortem) is not necessarily caused by low cerebral blood flow (without acute arterial occlusion) but this assumption has bedevilled much of the literature. The brain CT/MRI-defined site and size of any visible recent infarction is not an accurate way to diagnose a low-flow ischaemic stroke. This is, first, because some boundary-zone infarcts result from embolism.638,660,661 Second, there is much variation between individuals in where the boundary zones are, and they may even change with time in the same individual in response to changes in peripheral vascular resistance (Fig. 6.21). Third, however boundary-zone infarcts are defined on imaging, there is little difference between them and territorial presumed-embolic infarcts in patient demographic characteristics, vascular risk

Boundary of deep and superficial arteries

Fig. 6.21 The anterior and posterior boundary zones between the territories of the middle, anterior and posterior cerebral arteries. The maximum extent of these variable zones is shown on CT templates (see also Figs 4.11, 4.17, 4.23 and 4.25).

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 297

6.7 From symptoms, signs and clinical syndrome to cause

factors and even in the prevalence of arterial disease in the neck severe enough to cause low flow, which is far more often assumed than measured.662–665 However, there have not been many comparative studies, definitions of boundary zones vary and the numbers of patients have been small. On balance, although some boundary-zone infarcts may have a haemodynamic (i.e. low flow) cause, many others could be caused by embolism or acute occlusive thrombosis. After all, any arterial territory has a terminal zone which forms a boundary with adjacent arterial territories and which is probably particularly vulnerable to ischaemia. There is no reason to suppose that this zone is more susceptible to ischaemia caused by low flow without acute arterial occlusion, than as a result of acute arterial occlusion. Another possibility is that boundaryzone infarcts are caused by a combination of embolization to the margins of the territorial supply of a cerebral artery, as well as insufficient perfusion pressure to clear the emboli because of severe arterial disease in the neck, or operative hypotension.661 Clearly, in view of the diagnostic difficulties, it is quite conceivable that low flow is a more frequent cause of cerebral ischaemia, or less frequent, than currently believed. The exact sites of the boundary zones between the territories of supply of the major cerebral arteries are so variable between, and even within, individuals that the diagnosis of infarction in a boundary zone based on CT/MRI alone is all but impossible. Boundary-zone infarction can be caused by acute arterial obstruction, while low blood flow does not necessarily cause infarction only within boundary zones. Implications for treatment Being certain that an ischaemic episode is caused either by low flow alone, or by acute arterial obstruction, seldom really matters. It makes very little difference if ischaemic stroke or TIA caused by low flow is recognized as such because unless the precipitating factor(s) can be avoided or reversed (particularly over-treatment of hypertension), the management is exactly the same as for presumed embolic causes of ischaemia, i.e. antithrombotic drugs, statins, careful blood pressure lowering, management of any other causal vascular risk factors and surgical relief of any obstruction to blood flow if it is practical and safe to do so (Chapter 16). However, there is certainly a case for less aggressive treatment of hypertension if there is good evidence of low flow symptoms. And one must acknowledge that in a patient with or without known severe arterial disease in the neck, an ischaemic episode may occasionally be caused by low

flow rather than embolism or some other cause of acute arterial obstruction, and even that different episodes at different times in the same patient can be caused by different mechanisms.

6.7.6 Clues from the history The vast majority of transient ischaemic attacks (TIAs) and ischaemic strokes start suddenly, without any obvious provocation, and there are few if any symptoms other than those of a focal neurological or ocular deficit. Sometimes there can be clues to the cause in the history (Table 6.9), as well as to whether the patient has had a stroke or TIA in the first place (Chapter 3). These clues may require some tenacity to recognize, or perhaps just an ability to take a history instead of rushing to order lots of tests. There will be no clues from the history if no one bothers to take a history in the rush to organize a brain scan. Gradual onset Gradual onset of ischaemic stroke or TIA over hours or days, rather than seconds or minutes, is unusual but is becoming more recognized now that strokes are being seen much earlier (section 3.3.8). If the onset is gradual, and ischaemic stroke or TIA is not likely to be caused by low flow (section 6.7.5) or migraine (section 7.8), then the diagnosis should be considered particularly carefully and a structural intracranial (or ocular) lesion looked for again, or for the first time if brain imaging has not already been carried out (e.g. intracranial tumour, chronic subdural haematoma, cerebral abscess, section 3.4.4). Under the age of 50 years, multiple sclerosis should also be considered (section 3.4.10). However, a priori, focal neurological deficits which develop over hours, and even over 1 or 2 days, in elderly patients are still more likely to have a vascular than a non-vascular cause because in them vascular causes are so much more common than conditions such as brain tumours. It is only when progression occurs over a longer period that the likelihood of a non-vascular cause (such as chronic subdural haematoma) starts to rise. Precipitating factors The exact activity and time of onset may both be important (section 3.3.9). Anything to suggest a drop in cerebral perfusion or blood pressure may be relevant (section 6.7.5), as is any operative procedure (section 7.18).

297

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 298

298

Chapter 6 What caused this transient or persisting ischaemic event? Table 6.9 Important clues from the history which may suggest the cause of an ischaemic stroke or transient ischaemic attack, or that the diagnosis of cerebrovascular disease should be reconsidered. Gradual onset Low cerebral blood flow without acute occlusion (section 6.7.5) Migraine (section 7.8) Structural intracranial lesion (section 3.4.4) Multiple sclerosis (section 3.4.10) Precipitating factors Suspected systemic hypotension or low cerebral perfusion pressure (standing up or sitting up quickly, heavy meal, hot weather, hot bath, warming the face, exercise, coughing, hyperventilation, chest pain or palpitations, starting or changing blood pressure-lowering drugs) (section 6.7.5) Pregnancy/puerperium (section 7.14) Surgery (section 7.18) Head-turning (section 7.1.5) Hypoglycaemia (section 7.16) Valsalva manoeuvre (paradoxical embolism, section 6.5.12; or low flow, section 6.7.5) Recent headache Carotid/vertebral dissection (section 7.2.1) Migrainous stroke/transient ischaemic attack (section 7.8.1) Intracranial venous thrombosis (section 7.21.2) Giant-cell arteritis (or other inflammatory vascular disorders) (section 7.3.1) Structural intracranial lesion (section 3.4.4) Epileptic seizures Intracranial venous thrombosis (section 7.21.2) Mitochondrial diseases (section 7.19) Non-vascular intracranial lesion (section 3.4.4) Malaise Inflammatory arterial disorders (section 7.3) Infective endocarditis (section 6.5.9) Cardiac myxoma (section 6.5.13) Cancer (section 7.12) Thrombotic thrombocytopenic purpura (section 7.9.3) Sarcoidosis (section 7.3.16) Chest pain Myocardial infarction (section 7.10) Aortic dissection (section 7.2.3) Paradoxical embolism (sections 6.5.12)

Head-turning is an occasional cause (section 7.1.5). Recurrent attacks first thing in the morning or during exercise suggest hypoglycaemia, which is easy to think of in a diabetic patient on hypoglycaemic drugs but more difficult if there is a less obvious cause of hypoglycaemia, such as the very rare insulinoma or drugs such as pentamidine (sections 3.4.5 and 7.16). Onset during a Valsalva manoeuvre (e.g. lifting a heavy object) suggests a low flow ischaemic stroke (section 6.7.5) or paradoxical embolism (section 6.5.12), and so sets off a search for

Non-stroke vascular disease or vascular risk factors Heart disease (section 6.6.19) Claudication (section 6.6.19) Hypertension (section 6.6.3) Smoking (section 6.6.4) Drugs Oral contraceptives (section 7.13.1) Oestrogens in men (section 7.13) Blood pressure-lowering/vasodilators (section 6.7.5) Hypoglycaemic drugs (section 7.16) Cocaine (section 7.15.1) Amphetamines (section 7.15.1) Ephedrine (section 7.15.1) Phenylpropanolamine (section 7.15.1) ‘Ecstasy’ (section 7.15.1) Anti-inflammatory drugs (section 7.15.2) Antipsychotic drugs (section 7.15.3) Deoxycoformycin (section 7.3.21) Allopurinol (section 7.3.21) l-asparaginase (section 7.12) Injury Chronic subdural haematoma (section 3.4.4) Vertebral/carotid artery dissection (section 7.2.1) Cerebral air embolism (section 7.1.7) Fat embolism (section 7.1.8) Self-audible bruits Internal carotid artery stenosis (distal) (section 6.7.6) Dural arteriovenous fistula (section 8.2.8) Glomus tumour Caroticocavernous fistula (section 8.2.14) Raised intracranial pressure Intracranial venous thrombosis (section 7.21.2) Past medical history Inflammatory bowel disease (section 7.17) Coeliac disease (section 7.17) hom*ocystinuria (section 7.20.2) Cancer (section 7.12) Irradiation of the head or neck (section 7.12) Recurrent deep venous thrombosis (sections 7.3.4 and 7.9.11) Recurrent miscarriages (section 7.3.4) Recent surgery/long distance travel (section 6.5.12) Family history (Table 6.10)

deep venous thrombosis if there is evidence on echocardiography of a patent foramen ovale (section 6.9.3). Headache Headache at around the onset of ischaemic stroke or TIA occurs in about 25% of patients, is usually mild and, if localized at all, tends to be related to the position of the brain/eye lesion (sections 3.3.10 and 11.9). It is more common with vertebrobasilar than carotid

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 299

6.7 From symptoms, signs and clinical syndrome to cause

Table 6.10 Causes of familial stroke (including intracranial haemorrhage) and transient ischaemic attack. Connective tissue disorders Ehlers–Danlos syndrome (section 7.20.7) Pseudoxanthoma elasticum (section 7.20.7) Marfan syndrome (section 7.20.7) Fibromuscular dysplasia (section 7.4.1) Familial mitral leaflet prolapse (section 6.5.7) Haematological disorders Sickle cell disease/trait (section 7.9.8) Antithrombin III deficiency (section 7.9.11) Protein C deficiency (section 7.9.11) Protein S deficiency (section 7.9.11) Plasminogen abnormality/deficiency (section 7.9.11) Haemophilia and other inherited coagulation factor deficiencies (section 8.4.4) Others Familial hypercholesterolaemia Neurofibromatosis (section 7.20.5) hom*ocystinuria (section 7.20.2) Fabry’s disease (section 7.20.3) Tuberous sclerosis (section 7.20.4) Dutch and Icelandic cerebral amyloid angiopathy (section 8.2.2) Migraine (section 7.8) Familial cardiac myxoma (section 6.5.13) Familial cardiomyopathies (section 6.5.11) Mitochondrial diseases (section 7.19) CADASIL (section 7.20.1) Sneddon syndrome (section 7.3.5) Arteriovenous malformations (section 8.2.4) Cavernous malformations (section 8.2.5) Intracranial saccular aneurysms (section 8.2.3) CADASIL, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy.

distribution ischaemia, and less common with lacunar ischaemia.666–671 Severe pain unilaterally in the head, face, neck or eye at around or before the time of stroke onset is highly suggestive of carotid dissection, while vertebral dissection tends to cause unilateral or sometimes bilateral occipital pain (section 7.2.1). Migrainous stroke may be accompanied by headache (section 7.8.1) and patients with cerebral autosomal dominant arteriopathy with subcortical infarcts and leucoencephalo-pathy (CADASIL) usually have a history of migraine (section 7.20.1). In the context of the differential diagnosis of TIAs, migraine should be fairly obvious, unless there is no headache (section 3.4.1). It is important to note that vertebrobasilar ischaemia may cause similar symptoms to migraine with headache, gradual onset of focal neurological symptoms and visual disturbance. Although intracranial venous thrombosis usually causes either intracranial hypertension alone or a subacute

encephalopathy, a focal onset does occur and headache (occurring in around 75% of patients) can be a clue (section 7.21.2). Stroke or TIA in the context of a patient who has had a headache for days or weeks previously must raise the possibility of giant-cell arteritis and other inflammatory vascular disorders (section 7.3.1). Pain in the jaw muscles with chewing, which resolves with rest, strongly suggests claudication, which is caused by external carotid artery disease as a result of giant-cell arteritis far more often than atherothrombosis. Epileptic seizures Epileptic seizures, partial or generalized, within hours of stroke onset are distinctly unusual in adults (about 5%) and should lead to a reconsideration of non-stroke brain pathologies (section 3.4.4). They are rather more common in childhood stroke. They are more likely with haemorrhagic than ischaemic strokes and if the infarct is extensive and involves the cerebral cortex (section 11.8).672–676 They are also likely with venous infarction (up to 40%) (section 7.21.12) and mitochondrial disorders (section 7.19). Partial motor seizures can be confused with limb-shaking TIAs, but the former are more clonic and the jerking spreads in a typical Jacksonian way from one body part to another and the latter are supposed never to involve the face (sections 3.3.4 and 6.7.5). Very rarely, transient focal ischaemia seems to cause just partial epileptic seizures, but proving a causal relationship is seldom possible.677 Interestingly, onset of idiopathic seizures late in life is a powerful independent predictor of subsequent stroke.678 Because the diagnosis of stroke may be wrong if a tumour on CT is misinterpreted as an infarct (contrast enhancement can look very similar in both, section 5.4.2), partial seizures after a ‘stroke’ should always be an indication to re-examine the diagnosis and reassess the imaging. Also, seizures in the presence of a history of a few days of malaise, headache and fever should suggest encephalitis and the need for an electroencephalogram to show bilateral diffuse rather than unilateral focal slow waves, and cerebrospinal fluid examination (raised white cell count, but this can also occur in stroke, section 6.8.11). Malaise Stroke in the context of a patient who has been generally unwell for days, weeks or months suggests an inflammatory arterial disorder, particularly giant-cell arteritis (section 7.3.1), infective endocarditis (section 6.5.9), cardiac myxoma (section 6.5.13), cancer (section 7.12), thrombotic thrombocytopenic purpura (section 7.9.3) or even sarcoidosis (section 7.3.16).

299

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 300

300

Chapter 6 What caused this transient or persisting ischaemic event?

Chest pain Chest pain may suggest a recent myocardial infarction with complicating stroke (section 7.10); aortic dissection, particularly if the pain is also interscapular (section 7.2.3); and pleuritic pain suggests pulmonary embolism and the possibility of paradoxical embolism (section 6.5.12). Vascular risk factors Vascular risk factors (section 6.6) and diseases should be sought. It is most unusual for an ischaemic stroke or TIA to occur in someone with no vascular risk factors, unless they are very old, or are young with some unusual cause of stroke (Table 6.3). Heart disease of any sort may be relevant (source of embolism to the brain, dysrhythmias causing low-flow ischaemia, etc.) and cardiac symptoms should be specifically elicited in the history: angina, shortness of breath, palpitations, and so on. Drugs and drug users Drugs may well be relevant: oral contraceptives and hormone replacement therapy in women and oestrogens in men (section 7.13); anything which lowers the blood pressure (section 6.7.5); hypoglycaemic agents (section 7.16); and illicit drugs (section 7.15.1). Injury Any injury in the days and weeks before ischaemic stroke or TIA onset is crucial information. A head injury might have caused a chronic subdural haematoma (highly unlikely if more than 3 months previously) although this should have been considered earlier at the stroke vs non-stroke stage (section 3.4.4). Of possible relevance is an injury to the neck in the hours, days or even few weeks before onset, because this may cause carotid or vertebral dissection (section 7.2.1). After long bone fracture, fat embolism may cause a generalized encephalopathy, but occasionally there are additional focal features (section 7.1.8).679–682 It is therefore essential to ask about any injury, strangulation, car crash, unusual yoga exercises, neck manipulation and so forth, in any unexplained stroke (Table 6.2). Self-audible bruits Pulsatile self-audible bruits are rare. They can be differentiated from tinnitus because they are in time with the pulse. They may be audible to the examiner on auscultation of the neck, orbit or cranium. They are unlikely

to be caused by carotid bifurcation atherothrombosis because the source of the sound is too far from the ear. They are much more likely to indicate distal internal carotid artery stenosis (due to dissection or, rarely, atherothrombosis), dural arteriovenous fistula near the petrous temporal bone, glomus tumour, caroticocavernous fistula, intracranial venous thrombosis, symptomatic and idiopathic intracranial hypertension, a loop in the internal carotid artery, or just heightened awareness of one’s own pulse.683,684 Past medical history Past medical history of inflammatory bowel disease (section 7.17), coeliac disease (section 7.17), irradiation of the head and neck (section 7.12), cancer (section 7.12) or even hom*ocystinuria (section 7.20.2) may be important. Recurrent deep venous thrombosis (DVT) suggests thrombophilia (section 7.9.11), particularly if there is a family history, or the antiphospholipid syndrome (section 7.3.4). Recurrent miscarriage is another feature of the antiphospholipid syndrome. Any reason for a recent DVT (e.g. a long cramped journey, admission to hospital for an acute medical disorder, or surgery) should raise the question of paradoxical embolism (sections 6.5.12 and 6.9.3). If a patient has, or has had, deep venous thrombosis in the legs, then consider paradoxical embolism to the brain, a familial clotting factor problem or the antiphospholipid syndrome. Previous strokes and/or transient ischaemic attacks Previous strokes and/or TIAs in different vascular territories are more likely with a proximal embolic source in the heart, or arch of the aorta, than with a single arterial lesion in the neck or head. Attacks going back months or more make some causes unlikely (e.g. infective endocarditis, arterial dissection). Family history There are several rare familial conditions that may be complicated by ischaemic stroke and TIAs (Table 6.10). There is also increasing interest in complex genetic disorders thought to be caused by multiple gene interactions, presumably influenced by environmental factors (section 6.6.15).535,685–687 However, family history of stroke is only a modest risk factor for ischaemic stroke.687 Moreover, much of the association appears to be secondary to hereditability of risk factors for stroke such as hypertension and diabetes.687,688 On the other hand, these

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 301

6.7 From symptoms, signs and clinical syndrome to cause

same risk factors are clearly influenced by the environment (for example, diets rich in fat and in salt tend to raise the plasma cholesterol and blood pressure, respectively). Just how easy it will be to separate out shared genes from shared environment in a disease as common as stroke remains to be seen. Disentangling the interactions and working out the pathway from genotype to phenotype will be a monumental task (section 6.6.15). Whatever the mechanism(s), one can at least reassure patients with TIA or stroke that a family history of stroke is associated with little or no increase in the risk of a future stroke.689,690

• remember that comorbidity, such as pneumonia, sedative drugs, infection and hypoglycaemia, may all make the neurological deficit seem worse than it really is (section 11.4). If the neurological deficit is mild and yet the patient is drowsy, then consider chronic subdural haematoma, cerebral vasculitis, non-bacterial thrombotic endocarditis, intracranial venous thrombosis, mitochondrial disorders, thrombotic thrombocytopenic purpura, sedative drugs, hypoglycaemia, familial hemiplegic migraine and comorbidity, such as pneumonia or other infections.

6.7.7 Clues from the examination Eyes Neurological examination Neurological examination is primarily to localize the brain lesion; of course, in patients with transient ischaemic attacks (TIAs), or those seen some days after a minor stroke, there will probably be no signs at all (section 3.3.1). Occasionally, however, there may be a clue to the cause. A Horner syndrome ipsilateral to a carotid distribution infarct (i.e. not as the result of a brainstem stroke, where it might be expected) suggests dissection of the internal carotid artery (ICA) or sometimes acute atherothrombotic carotid occlusion (section 7.2.1). Lower cranial nerve lesions ipsilateral to a hemispheric cerebral infarct can also occur in carotid dissection and, like Horner syndrome, are caused by stretching and bulging of the arterial wall in relation to the affected nerves, or ischaemia. Ocular ischaemia, as well as third, fourth and sixth cranial nerve palsies – sometimes with orbital pain – has been described ipsilateral to acute ICA occlusion and stenosis, presumably caused by ischaemia of the nerve trunks.691,692 In a total anterior circulation infarct or brainstem stroke some drowsiness is expected, but with more restricted infarcts, consciousness is normal. Therefore if consciousness is impaired and yet the ‘stroke’ itself seems mild, it is important to: • reconsider the differential diagnosis (particularly chronic subdural haematoma) (section 3.4.4); • consider the diffuse encephalopathic disorders which have focal features and which may masquerade as stroke, e.g. cerebral vasculitis of some sort (section 7.3), endocarditis (sections 6.5.9 and 6.5.10), intracranial venous thrombosis (section 7.21), mitochondrial disorders (section 7.19), thrombotic thrombocytopenic purpura (section 7.9.3), familial hemiplegic migraine693 and Hashimoto’s encephalitis;694

The eyes may provide general clues to the cause of a stroke (e.g. diabetic or hypertensive retinopathy), or may reveal papilloedema which would make the diagnosis of ischaemic stroke, or even intracerebral haemorrhage, most unlikely. In addition, it is worth searching thoroughly for evidence of emboli which are very often completely asymptomatic695,696 (section 3.3.6): • Fibrin–platelet emboli are dull greyish-white amorphous plugs but are rarely observed, perhaps because they move through the retinal circulation and disperse; they suggest embolism from the heart or proximal sources of atherothrombosis. • Cholesterol emboli quite often stick at arteriolar branching points, usually without obstructing the blood flow, and appear as glittering orange or yellow bodies reflecting the ophthalmoscope light; obviously these strongly suggest embolization from proximal atheromatous plaques, but they are often asymptomatic. • ‘Calcific’ retinal emboli appear as solid, white and non-reflective bodies and tend to lodge near the edge of the optic disc; they suggest embolism from aortic or mitral valve calcification. Localized areas of periarteriolar sheathing, seen as opaque white obliteration of segments of the retinal arterioles, suggest embolism, usually cholesterol, in the past. Roth spots in the retina are very suggestive of infective endocarditis (section 6.5.9). Dislocated lenses should suggest Marfan syndrome (section 7.20.7) or hom*ocystinuria (section 7.20.2); angioid streaks in the retina suggest pseudoxanthoma elasticum (section 7.20.7); and in hyperviscosity syndromes there is a characteristic retinopathy (section 7.9.10). Dilated episcleral vessels are a clue to abnormal anastomoses between branches of the external carotid artery (ECA) and orbital branches of the internal carotid artery (ICA), distal to severe ICA disease.697 With very

301

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 302

302

Chapter 6 What caused this transient or persisting ischaemic event?

severe ICA disease, usually accompanied by severe disease of the ipsilateral ECA, the eye may occasionally become so ischaemic that venous stasis retinopathy develops, although arterial disease does not invariably underlie this condition698 (section 3.3.6). Haemorrhages are scattered around the retina with microaneurysms, and the retinal veins are dilated and irregular. The retinal blood flow is extremely impaired, as demonstrated by lightly compressing the eye with one finger while observing the fundus and noting collapse of the central retinal artery. With more extreme ischaemia, ischaemic oculopathy may develop with impaired visual acuity, eye pain, rubeosis of the iris (dilated blood vessels), fixed dilated pupil, ‘low pressure’ glaucoma, cataract and corneal oedema.699,700 Pre-existing raised intraocular pressure, i.e. glaucoma, makes the eye more susceptible to low blood flow and ischaemia as a result. Arterial pulses It is always worth feeling both radial pulses simultaneously. Any inequality in timing or volume suggests subclavian or innominate stenosis or occlusion, and this is further supported if there is an ipsilateral supraclavicular bruit or lower blood pressure in the arm with the weak or delayed pulse. An elderly patient presented with a sudden left-sided hemiparesis and no other symptoms. She had right carotid and supraclavicular bruits. Brain CT was normal. Three days later the duplex examination showed narrowing of the right common carotid artery which appeared to be caused by dissection of the aortic arch. Only then did she admit to some mild chest pain before the stroke. Unequal pulses and blood pressures were found in her upper limbs, and chest CT confirmed the aortic dissection. The lessons are that any pain in or around the chest may be relevant and should have been more thoroughly sought, and in all stroke patients both radial pulses must be felt routinely before, not after, arterial imaging. Normally, the internal carotid artery pulse is too deep and rostral to be felt in the neck. Therefore, any loss of the ‘carotid’ pulsation reflects common carotid artery (CCA) or innominate occlusion or severe stenosis, both rather rare situations or, perhaps more likely, the artery is too deep to be felt or the neck too thick. The arterial pulse felt in the neck comes from the common, not the internal, carotid artery. The superficial temporal pulses should be easily felt and symmetrical. If there is unilateral absence or delay,

this suggests external carotid artery (ECA) or CCA disease. Tenderness of any of the branches of the ECA (occipital, facial, superficial temporal) points towards giant-cell arteritis. Tenderness of the carotid artery in the neck (i.e. the CCA) can occur in acute carotid occlusion but is more likely to be a sign of dissection, or possibly arteritis. Absence of several neck and arm pulses in a young person suggests Takayasu’s arteritis (section 7.3.2). Other causes of widespread disease of the aortic arch are atheroma, giant-cell arteritis, syphilis, subintimal fibrosis, arterial dissection and trauma.264,701,702 Delayed or absent leg pulses suggest coarctation of the aorta or, much more commonly, peripheral vascular disease (PVD) which is very common in patients with TIA and ischaemic stroke (Table 6.3) and may need treating in its own right. Furthermore, PVD is an important predictor of future serious vascular events (Table 16.4). The state of the femoral artery is important to assess before cerebral angiography via the femoral route (section 6.8.5) and, if after angiography the leg pulses disappear, then it was a complication of the angiography. Obviously, any evidence of systemic embolism would direct the search towards a source of emboli in the heart (section 6.5). Finally, while the hand is on the abdomen, aortic aneurysm should be considered while searching for any masses or hepatosplenomegaly. Although the prevalence of aortic aneurysm in these stroke/TIA patients is unknown, it could well be quite high, particularly in men and if the patient has carotid stenosis (section 6.6.19). Diagnosis is important because of the benefits of surgery in patients with larger aneurysms.606 Cervical bruits Listening to the neck is a favourite occupation for inquisitive physicians, and acquisitive surgeons, and can lead to some useful information (Fig. 6.22). A localized bruit, occasionally palpable, over the carotid bifurcation (i.e. high up under the jaw) is predictive of some degree of carotid stenosis, but very tight stenosis (or occlusion) may not cause a bruit at all (Fig. 6.23) (Table 6.11). External carotid stenosis can also cause a bruit in the same place. An innocent carotid bruit is more common in women,703 probably due to sex differences in carotid bifurcation anatomy.704 Bruits transmitted from the heart become attenuated as one listens further up the neck towards the angle of the jaw, thyroid bruits are bilateral and more obviously over the gland, a hyperdynamic circulation tends to cause a diffuse bruit, and venous hums are more continuous and roaring, and are obliterated by light pressure over

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 303

6.7 From symptoms, signs and clinical syndrome to cause Table 6.11 The source of neck bruits.

Bruit from carotid bifurcation

Bruit from subclavian artery, or vertebral artery origin

Transmitted bruit from the heart

Fig. 6.22 The sites of various cervical bruits. Note that a bruit arising from the carotid bifurcation is high up under the angle of the jaw. Localized supraclavicular bruits are caused either by subclavian or vertebral origin artery stenosis.

Carotid bifurcation arterial bruit Internal carotid artery origin stenosis External carotid artery origin stenosis Supraclavicular arterial bruit Subclavian artery stenosis Vertebral artery origin stenosis Can be normal in young adults Diffuse neck bruit Thyrotoxicosis Hyperdynamic circulation (pregnancy, anaemia, fever, haemodialysis) Transmitted bruit from the heart and great vessels Aortic stenosis/regurgitation Mitral regurgitation Patent ductus arteriosus Coarctation of the aorta Venous hum

Carotid bruits are neither sufficiently specific nor sensitive to diagnose carotid stenosis severe enough to consider surgery. Arterial imaging is needed. Physicians and surgeons desperate to use their stethoscopes would do better to measure the blood pressure and listen to the heart. Cardiac examination

No bruit

Carotid bruit

Carotid bruit (%)

100 80 60 40 20 0

Normal

1–24

25–49

50–74

75–99 Occluded

Diameter stenosis of symptomatic carotid artery (%) Fig. 6.23 The percentage of patients with a localized bruit over the symptomatic carotid bifurcation for various degrees of stenosis as estimated (using the European Carotid Surgery Trial method) from 298 carotid angiograms. (Adapted with permission from Hankey and Warlow, 1990.790)

the ipsilateral jugular vein.593 An arterial bruit in the supraclavicular fossa suggests either subclavian or proximal vertebral arterial disease, but a transmitted bruit from aortic stenosis must also be considered. Normal young adults quite often have a short supraclavicular bruit; the reason is unknown.

Cardiac examination is important, particularly to look for any cardiac source of embolism (section 6.5). If physicians feel under-confident about their cardiological abilities, then they should get properly trained or a cardiologist will have to be consulted. Atrial fibrillation (AF) will already have been suspected from the radial pulse; left ventricular hypertrophy suggests hypertension or aortic stenosis, and most major cardiac sources of embolism are fairly obvious clinically (e.g. AF, mitral stenosis, prosthetic heart valve). Fever Fever is distinctly unusual in the first few hours after stroke onset. Any raised temperature at this time must therefore be taken seriously and endocarditis or other infections, inflammatory vascular disorders, deep venous thrombosis or cardiac myxoma considered. Later on, fever is quite common and usually reflects some complication of the stroke (section 11.12). Skin and nails The skin and nails occasionally provide clues to the cause of ischaemic stroke or transient ischaemic attack (Table 6.12).

303

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 304

304

Chapter 6 What caused this transient or persisting ischaemic event?

Finger-clubbing

Right-to-left intracardiac shunt (section 6.5.12) Cancer (section 7.12) Pulmonary arteriovenous malformation (section 6.5.12) Infective endocarditis (section 6.5.9) Inflammatory bowel disease (section 7.17) Splinter haemorrhages Infective endocarditis (section 6.5.9) Cholesterol embolization syndrome (section 7.7) Vasculitis (section 7.3) Scleroderma Systemic sclerosis (section 7.3.13) Livedo reticularis Sneddon syndrome (section 7.3.5) Systemic lupus erythematosus (section 7.3.4) Polyarteritis nodosa (section 7.3.6) Cholesterol embolization syndrome (section 7.7) Lax skin Ehlers–Danlos syndrome (section 7.20.7) Pseudoxanthoma elasticum (section 7.20.7) Skin colour Anaemia (section 7.9.7) Polycythaemia (section 7.9.1) Cyanosis (right-to-left intracardiac shunt, pulmonary arteriovenous malformation) (section 6.5.12) Porcelain-white papules/scars Kohlmeier–Degos disease (section 7.3.15) Skin scars Ehlers–Danlos syndrome (section 7.20.7) Petechiae/purpura/bruising Thrombotic thrombocytopenic purpura (section 7.9.3) Fat embolism (section 7.1.8) Cholesterol embolization syndrome (section 7.7) Ehlers–Danlos syndrome (section 7.20.7) Orogenital ulceration Behçet’s disease (section 7.3.11) Rash Fabry’s disease (section 7.20.3) Systemic lupus erythematosus (section 7.3.3) Tuberous sclerosis (section 7.20.4) Epidermal naevi Epidermal naevus syndrome Café-au-lait patches Neurofibromatosis (section 7.20.5) Thrombosed superficial veins, Intravenous drug users (section 7.15.1) needle marks

Getting to the bottom of the cause of an ischaemic stroke or transient ischaemic attack requires much more than just neurological skills. Stroke medicine, like neurology, is part of general internal medicine. It is important therefore that doctors looking after stroke patients have a good general internal medical training.

6.8 Investigation

Investigations are mainly to help unravel the pathological type of stroke (ischaemic stroke vs intracerebral haemorrhage vs subarachnoid haemorrhage, as discussed in Chapter 5) and then to determine the cause of the cerebral ischaemia (or intracranial haemorrhage; Chapters 8 and 9), particularly a cause that will influence

Table 6.12 Clues to the cause of ischaemic stroke/transient ischaemic attack from examination of the skin and nails.

immediate treatment or long-term management. Investigation may also provide important prognostic information, e.g. severe carotid stenosis on ultrasound (section 16.11.8). In addition, the many patients who also have angina or other cardiac symptoms, claudication or suspected aortic aneurysm may well need specific investigations directed at these problems with a view to appropriate treatment. Ideally, any investigation should be practical, feasible, accurate, safe, non-invasive, inexpensive and, most importantly, informative in the sense that the result (positive or negative, high or low, etc.) will influence patient management and outcome. Idle curiosity or financial gain are not good reasons to order any test.

6.8.1 Routine investigations Although there are no absolute rules, all ischaemic stroke/transient ischaemic attack (TIA) patients, unless

9781405127660_4_006.qxd 10/13/07 2:11 PM Page 305

6.8 Investigation

they are already heavily dependent or institutionalized, or have already been recently investigated for a previous event or some other problem, should have basic noninvasive first-line investigations within a few hours of presentation. None of these necessarily require hospital admission, although brain imaging does require attendance at hospital (Table 6.13). The chance of picking up a relevant abnormality (yield) may be very low for some tests – e.g. full blood count and erythrocyte sedimentation rate (ESR) – but these are cheap and the consequences of missing a treatable disorder, such as giant-cell arteritis, are serious. There is a higher chance of picking up a treatable abnormality with the blood glucose, urine analysis and electrocardiogram (ECG). Depending on the definition, many or even most patients are hypercholesterolaemic, and how this should be acted on is discussed in section 16.4; immediately after stroke, but probably not TIA, there is a transient fall in plasma cholesterol, which will underestimate the usual level.705,706 Brain imaging is discussed in sections 5.4, 5.5 and 6.8.3. If patients have the basic tests and if the results are read, written in the records and acted upon, Table 6.13 First-line investigations for ischaemic stroke/transient ischaemic attack.

Investigation Full blood count

this may well do more good than the inappropriate ordering of a huge range of further tests while missing some crucial clue from one of the routine investigations (such as an ESR of 100 mm in the first hour). All patients should have a full blood count, erythrocyte sedimentation rate, plasma glucose, urea, electrolytes and cholesterol, urine analysis, and electrocardiogram. Most should also have a CT and/or MR brain scan.

6.8.2 Second-line investigations for selected patients Second-line investigations (Table 6.14) are usually more costly, invasive and/or dangerous, so they must be targeted on patients most likely to gain from a useful change in management as a consequence of the test result. The likelihood of a relevant result depends on the selection of patients for the investigation – a balance must be struck between over-investigation (inconvenience, high cost, possibly high-risk, low yield, false-positive results

Disorders suggested

Anaemia, polycythaemia, leukaemia, thrombocythaemia, heparin-induced thrombocytopenia with thrombosis, infections Erythrocyte Vasculitis, infections, cardiac myxoma, sedimentation rate hyperviscosity, cholesterol embolization syndrome, non-bacterial thrombotic endocarditis Plasma glucose Diabetes mellitus, hypoglycaemia Urea and electrolytes Diuretic-induced hypokalaemia, renal failure, hyponatraemia Plasma cholesterol Hypercholesterolaemia Syphilis serology Syphilis, anticardiolipin syndrome Urinalysis Diabetes, renal disease, infective endocarditis, vasculitis, Fabry’s disease ECG Dysrhythmia, left-ventricular hypertrophy, silent myocardial infarction Unenhanced CT Intracerebral haemorrhage, non-vascular brain scan intracranial mimic of stroke-syndrome

Yield* (%) 1

5 3 45 10% >20%

in the completeness of the assessment of stroke patients following the introduction of a stroke clerking or admission form.55,56 The use of the form also makes it much easier to access the information subsequently and to identify relevant items that are missing. However, one has to acknowledge the practical difficulties of introducing disease-specific assessment forms in situations – e.g. accident and emergency departments – where patients with a wide range of medical problems are being assessed. For example, it may not be clear until the end of the assessment what the most likely diagnosis is and therefore which form should have been used. One solution might be a ‘core form’ for all patients, with supplementary sheets for specific common conditions. There is reasonable evidence that patient-specific reminders (e.g. ‘if ischaemic stroke, prescribe aspirin’), which can be included in such forms (perhaps as part of an integrated care pathway), improve adherence to management guidelines although randomized trials have not yet demonstrated improvements in patient outcomes.57,58 A clerking or admission form will improve the completeness and relevance of the initial assessment and will facilitate communication and audit, but the introduction of diagnosis-specific forms is not without its problems. Although a physician is often the first person to assess the patient, it is important to emphasize that assessment should often involve other members of the multidisciplinary team (section 10.3.5). It is often very valuable, even on the day of the stroke, to involve the nurse, the physiotherapist and a speech and language therapist. Advice on the patient’s risk of pressure sores, lifting, handling and positioning and the patient’s ability to

Fig. 10.14 ABCD scale to estimate the risk of stroke early after a transient ischaemic attack. DBP, diastolic blood pressure; SBP, systolic blood pressure.

swallow safely are all relevant to their care from the moment of hospital admission, or indeed from the moment of assessment at home. The components of the initial assessment and their use are summarized in Table 10.8. It is important to emphasize again that assessment may well be staged; it is not a ‘once only’ process, but should continue throughout the course of the patient’s recovery from stroke. Assessment is not a ‘once only’ process, but should continue throughout the course of the patient’s recovery from stroke. Health professionals involved in the initial assessment of stroke patients learn that, although the patient may be a useful source of information, other people often provide more information that is essential to planning treatment. This is particularly important when the patient, for a variety of reasons, cannot communicate. It is usually valuable to spend a little time interviewing the family, neighbours, general practitioner, ambulance technicians or nursing staff, using a telephone if necessary (Fig. 10.15). To complete an assessment it is usually valuable to talk to family, neighbours, or family doctor; essential information can often be collected by a telephone call to the appropriate person. Social environment The patient’s social environment is a very important factor in determining the overall effect that a stroke will have on an individual and his or her family. Accurate knowledge of social networks is therefore critical when

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 521

10.3 Delivering an integrated management plan Table 10.8 Information that should be collected during the initial assessment of a stroke patient, and the potential value. Type of information

Diagnosis

Cause

Prognosis

Hyper-acute treatment

Demographics History of onset Risk factors Coexisting disease Medication Social details Pre-stroke function General examination Neurological examination Investigations

+ + + +

+ + + + + +

+ + + +

+ +

+ + +

+ +

+ + + + +

+ +

+ + +

Problems

Secondary prevention +

+ + + + + + + +

+ + + + + + +

because one of the major aims of rehabilitation is to minimize handicap. For example, it may be more appropriate to put greater energy and resources into occupational therapy for a craftsman than for a schoolteacher, who might require relatively more speech therapy. The patient should, if possible, have a major role in deciding such priorities. It is so often difficult to see the real person behind the facial weakness, severe aphasia, hemiplegia and incontinence.

Fig. 10.15 The telephone is a valuable diagnostic tool.

setting longer-term goals for rehabilitation and in planning discharge from hospital. Moreover, this allows one to build up a picture of the patient as a person rather than as ‘just another stroke’ or ‘that fibrillating hemiplegic in bed six’. It is often difficult to see the real person behind the facial weakness, severe aphasia, hemiplegia and incontinence. Finding out about the patient’s prestroke life might encourage members of the team who are caring for the patient to treat him or her with more understanding and sympathy. Also, it is this background information that allows one to judge the likely effect that the individual’s disabilities will have on their role in society, i.e. the likely handicap. This is important,

Although a lot of this background information can be collected over a longer period while the patient is recovering, it can be useful early on and may be more easily collected at the initial assessment. So often, the family – who may be the only source of this sort of information – disappear within a few hours of the admission and may then not reappear to be asked the relevant questions. It is therefore vital to seize the opportunity when a patient is first admitted, or at least during the next day or two, to obtain as much information from the family and friends as possible. Clearly, this may not be regarded as a medical priority, but professionals other than doctors, in particular the nursing staff, are often well placed to collect it. However, often this is not done very well. A complete picture of the patient’s pre-stroke life will be useful when deciding how aggressively to manage a patient (e.g. neurosurgery for obstructive hydrocephalus). Because this background information is so important, but may not be available at the initial assessment, one needs to have some method of identifying which items of data are missing so they can be sought later on. The clerking or admission form, or a patient record that is shared by the different professions involved (the so-called ‘combined’ or ‘single-patient’ record), has the potential if properly

521

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 522

522

Chapter 10 A practical approach to the management of stroke and transient ischaemic attack patients

used to fulfil this role. Many stroke units have introduced ‘integrated care pathways’, which usually include an admission form, multidisciplinary records and guidance on how to manage common problems (section 17.14). It is just as important to know the home and social circ*mstances of a stroke patient for early decision making (such as the desirability of emergency operation) as for later rehabilitation and discharge from hospital. By the end of the initial assessment, which may be punctuated by giving a hyperacute treatment, one should have collected enough information to produce a diagnostic formulation, including certainty of stroke diagnosis, site and size of the brain lesion, and the likely causes. This leads on to choices about the amount of further investigation required (e.g. echocardiography or not?), and enables the physician to talk to the patient and/or the family about the likely diagnosis, prognosis and management. At this stage, it is valuable to consider, and even list, the particular problems the patient has, to ensure they have all been identified and addressed. It may also be useful to use a checklist of the most common ones that occur at this early stage (Table 10.9).

10.3.3 Identifying problems and setting goals The patient’s assessment, both initially and subsequently, should identify any major problems, and it is these that Table 10.9 Important things to think about during the first day after a patient has had a severe stroke. Maintenance of a clear airway Treatment of co-existing or underlying disease Need to review the patient’s usual medication(s) Investigation and avoidance of fever Adequacy of oxygenation Swallowing ability Hydration Nutritional status Exclusion of urinary retention Management of urinary incontinence Exclusion of fractures if the patient may have fallen Prevention of: Deep venous thrombosis Pressure ulcers Aspiration Trauma Faecal impaction Protection of a flaccid shoulder Obtaining information from and giving information to the patient and family

determine the patient’s management. Problems can occur at every level of the patient’s illness – i.e. pathology, impairment, disability and handicap (section 10.1). For example, a problem at the pathology level might be ‘diagnostic uncertainty’ or ‘raised erythrocyte sedimentation rate’, which should lead on to further investigation if appropriate; while a problem at the level of handicap might be that the patient provides the only income for a large family, who now have no money to feed themselves. Having identified a problem, one can then formulate a plan to solve or at least alleviate that problem. Thus, a problem list can be turned into an action plan for the individual patient. Some problems can be dealt with very simply (e.g. antibiotics for a urinary tract infection). These are the ‘short-loop problems’ (section 10.1.2). The goal here is simply to remove the problem. Other problems such as immobility – which are more complex, respond more slowly to therapy and may require several different types of intervention – are the ‘long-loop problems’ (section 10.1.2); here it is useful to set a long-term goal of removing or alleviating the problem, but it is also helpful to set intermediate goals that allow one to judge whether progress is being made towards the long-term goal. Why a goal-orientated approach has several advantages Setting goals allows forward planning and provides a useful focus for multidisciplinary team meetings (section 10.3.7). Intermediate goals allow members of the team to coordinate their work, assuming that goals are achieved on time, and so improve efficiency. For example, if patients are to dress the lower half of their body, they must be able to stand. If the physiotherapist can estimate when the patient will be able to stand independently, the occupational therapist can plan when to start working on dressing the lower half. Setting longer-term goals can, for example, allow advanced planning of a pre-discharge home visit and final discharge to the community, which can reduce the patient’s length of stay in hospital by the number of days or weeks needed to plan a home visit or to make any necessary adaptations to the patient’s home before discharge (e.g. stair rails). If realistic goals are set, they can then be used to help motivate patients, especially if they have been involved in choosing or setting the goal. Recovery from a stroke may be very slow, so slow that the patient and even the therapists are unable to discern any progress being made towards the long-term goal (e.g. to achieve independent mobility). If one sets and achieves intermediate goals (e.g. sitting balance), progress is more easily perceived and morale maintained. The management of patients

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 523

10.3 Delivering an integrated management plan

Fig. 10.16 The ‘road’ to recovery after a hemiplegic stroke, showing some mobility ‘milestones’: 1, sitting balance for 1 min; 2, standing for 10 s; 3, 10 steps unaided; 4, timed 10-m walk (Smith & Baer 1999).59

with stroke in hospital is often allowed to drift without direction or leadership. The clinician who is responsible for the patient waves to them on a weekly ward round, in the belief that the therapists are actively rehabilitating the patient. If the head of the multidisciplinary team maintains discipline, and encourages the setting of both intermediate and long-term goals, drift can be avoided. This discipline benefits the patients, the team members and the service as a whole. The greater efficiency reduces length of stay in hospital and allows a greater proportion of stroke patients to be treated in the stroke unit without the need for extra staff or beds. Describing goals Where the goal is simply the removal of a problem, such as a urinary tract infection, it is fairly easy to describe it and then measure progress, e.g. relief of symptoms and sterile urine on a repeat culture. Similarly, longterm goals are often easy to describe, but should take into account the patient’s need for accommodation, physical and emotional support, how they might fill their time and what role they play in society. Judgements about whether these long-term goals have been achieved are relatively straightforward. For example, if the goal is to get a patient home to live with their family, or to return to work, one does not need any complex measures of outcome. In some areas, it is even fairly easy to set intermediate goals. For example, many patients with stroke

have problems with mobility, but usually the patient will achieve certain physical milestones on the road to recovery59 (Fig. 10.16). Further improvement in mobility can be measured by recording the time it takes the patient to walk 10 m.60 Thus, it is a fairly simple process to set an intermediate goal, which the patient or carer can understand, in terms of the level of function and the date by which it should be achieved. If goals are to be reproducible, and therefore useful, a standard method for assessing attainment or not needs to be established and used. It is more complex to set goals for problems such as language or activities of daily living (ADL) than for mobility. For example, a goal for independent dressing a patient’s top half would need to take account of important factors including the clothes worn, fasteners involved and the amount of prompting allowed (section 11.32.5). One could use a score on any one of the huge number of measures of language function or ADL as intermediate goals, but these scores are not easily understood by patients and carers, or even by the professionals using them. For example, it is unlikely to mean much to a patient, or even the team members, to aim for a Barthel index score of 12 (out of a maximum of 20) within the next 2 weeks. Despite these difficulties, the team should attempt to identify problems, specify intermediate and long-term goals, and introduce some meaningful measures to determine whether progress is being made towards achieving each of them.

523

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 524

Chapter 10 A practical approach to the management of stroke and transient ischaemic attack patients

In every stroke patient, intermediate and long-term goals should be described so that progress towards them can be measured. Moreover, everyone will feel a sense of achievement when the goals are met.

10.3.4 Goal setting: a diagnostic tool Goals may also be useful in identifying new or previously unrecognized problems. If a patient is not achieving the goals that have been set, it may be due to a number of reasons, which can be divided into team factors, patient factors and carer factors.

8 Function

524

Complication

6 4 Failing to achieve goals

Time (weeks) Expected Observed

Team factors If the goals are too ambitious because of inaccurate diagnosis, inadequate assessment or uncertainty about the prognosis, then patients may fail to achieve them and this will have a detrimental effect on the patient’s and the team’s morale. If goals are too easy, then progress may be slower than is, in fact, possible. Also, goal setting must be realistic if one is to use it to coordinate care (section 10.3.3). To set realistic goals in such a way that they are more often achieved than not requires an understanding of the prognosis and of the likely effectiveness of potential interventions. We have already seen that accurate predictions of progress and outcome are difficult in individual patients (section 10.2.7), but informal judgements made by the team may be more accurate, because they are based on past experience and observation of the patient’s progress over a period of time. Another reason why a patient may not be achieving a goal might be lack of appropriate treatment. Thus, progress may be hampered by too little therapy, or by the wrong sort of therapy. However, since there is currently so little information about the optimum amount or the relative effectiveness of most interventions, it is difficult to sort this out. Team members will much more often have to modify their therapy based on their own experience rather than on evidence from properly conducted randomized controlled trials. Goals should be meaningful and challenging, but achievable. Patient and carer factors New, unrecognized medical or psychological problems (e.g. infection, recurrent stroke or depression) may not present overtly – especially in their early stages – but may develop in a less specific way, causing a patient’s progress to slow, stop, or even reverse. One can draw a parallel

Fig. 10.17 ‘Failure to thrive’ after a stroke. Deviation from the expected recovery pattern might be due to any number of factors, including recurrent stroke, infections, depression, etc.

with the concept of ‘failure to thrive’ in paediatric practice (Fig. 10.17). If a patient is failing to achieve his or her goals (or milestones), then one needs to identify the likely reason or reasons (Tables 11.5 and 11.7). When a patient is failing to achieve his or her goals (or milestones) then one must identify the reason, and if possible do something about it. Sometimes the patient or the carer may have different goals from those of the team looking after the patient. This is ‘goal mismatch’. For example, a patient who does not want to live alone but would prefer to live with his or her daughter may not achieve the level of independence expected. Therefore, when setting goals ideally one should agree them with the patient and carers – although they may be reticent about discussing such matters openly. It is also important to involve the patient, and perhaps the carer, in setting goals, to ensure that the goals are really relevant to them. Most stroke patients are retired, so that leisure activities may be particularly important to their quality of life (section 11.33.3). The patient may be less interested in a goal aiming at achieving self-care in dressing than in being able to read or do the gardening. In hospital practice, where activities of daily living (ADL) abilities often determine the length of stay, too much emphasis can be placed on ADL-related goals because of the pressures on the team to make beds available for new patients and to minimize costs by discharging the patient as early as possible. There are however important practical difficulties in involving patients in goal setting, including their cognitive,

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 525

10.3 Delivering an integrated management plan

communication and visuuospatial problems and the clinician’s lack of experience or time.61,62 Goal setting may sometimes involve just an individual professional, but ideally it should involve the rest of the team, the patient and sometimes the patient’s family.

10.3.5 The stroke team Although a physician usually has overall responsibility for the management of the patient, other members of the multidisciplinary team play an essential part. Stroke unit care reduces the case fatality, physical dependency and need for institutionalization compared with care on a general medical ward (section 17.6). The main difference between the two models of care is that stroke unit care is coordinated with a multidisciplinary team. Because stroke patients have such a broad range of problems, their care demands input from several professions. For coordination of the professionals’ input, it is important that at least some of them should work as a core team, with regular meetings to discuss patients’ progress and problems (Table 10.10). Other professionals (Table 10.11), who may not be regular members of the team, should be available for consultation about individual patients. Although well established in rehabilitation settings, the team has an important role in all phases of treatment, even on the day of the stroke. Of course, the type and intensity of input from different members of the team will vary at different stages of the patient’s illness. By working closely together and sharing information and skills, some blurring of the boundaries between the roles of the professions becomes possible, and this can provide greater flexibility and efficiency. For example, if the nursing staff are trained by the speech and language therapist to screen for dysphagia, this will provide every patient with early screening (even at weekends), enabling the speech and language therapist to focus on patients with definite swallowing or communication problems (sections 11.17 and 11.30). Some have suggested that we should develop a hybrid therapist who could take on several roles, but this interesting idea

Table 10.10 The core stroke team. Physician Nurse Physiotherapist Occupational therapist Speech and language therapist Social worker

Table 10.11 Other professionals who may be helpful in the management of particular stroke patients. Others who may be consulted

Example of problem

Clinical psychologist Psychiatrist Neurosurgeon Vascular surgeon Radiologist

Antisocial behaviour Severe depression Obstructive hydrocephalus Peripheral artery embolus Unusual CT scan appearance Painful shoulder Fractured neck of femur Refractive problems Persistent diplopia Shortened leg, foot drop Weight loss Formulations for dysphagia Ill-fitting dentures

Rheumatologist Orthopaedic surgeon Optometrist Ophthalmologist Orthotist Dietician Pharmacist Dentist CT, computerized tomography.

Physiotherapist Occupational therapist

Nurse

Patient and carer

Social worker

Physician Speech therapist

Fig. 10.18 The traditional model of care or rehabilitation, in which each member of the multidisciplinary team interacts independently with the patient and/or carer.

has, not surprisingly, met with considerable opposition from the existing professions. This concept may have particular merits in situations in which it is difficult to coordinate the activities of a multidisciplinary team, for example in a rural community rehabilitation setting. Figures 10.18 and 10.19 illustrate two models of how members of the stroke team can provide input to patients and their carers. In the first, each professional predominantly works directly with the patient and/or carer, while in the second each professional has less direct patient contact, but influences the care given by a primary nurse. The two models represent the extremes, and it is likely that in the real world care is provided in an intermediate way. Although the model represented in Fig. 10.19 realistically means that patients must be treated in a stroke unit, it has other important advantages.

525

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 526

526

Chapter 10 A practical approach to the management of stroke and transient ischaemic attack patients

Speech therapist

Occupational therapist

Physician Physiotherapist

Nurse Physician

Social worker

Patient and carer Fig. 10.19 The model of care or rehabilitation that can be adopted on a geographically defined stroke unit, in which each member of the multidisciplinary team influences the nursing input to the patients and carers, as well as having direct interaction with them.

Stroke patients, even on a stroke unit, are directly involved in therapeutic activity with therapists for only a small proportion of their time.63 Training the nurses to practise with the patients the activities initiated by the therapists should encourage consistency and increase the total amount of therapy received by patients. After all, the nurses are caring for patients on a ward throughout the 24-h period. If, for example, the physiotherapists are teaching a patient how to transfer from bed to wheelchair, it is important that the patient should continue to do this in the same way between physiotherapy sessions. Otherwise, many of the skills the patient acquires during physiotherapy sessions may not be used during activities on the ward or, more importantly, at home. The same principles apply to input from other therapists. Whether therapy delivered by a non-specialist is as effective as therapy delivered by a highly trained specialist needs to be evaluated in randomized trials. One of the important functions of the team meeting is to harmonize the activity of the therapists and the nurses who provide much of the daily input to the patients. This may be at least one of the reasons why stroke units seem to achieve better outcomes (section 17.6).

10.3.6 The roles of the team members This section outlines the main functions of the core members of the team. In practice, and certainly where a team is functioning well, there will be some blurring of the roles of the team members with each performing the less specialized functions of the others. The roles described and the professional labels used reflect those in our own clinical practice but they inevitably vary in different healthcare settings.

Few doctors really understand the important role that they can play in the care of stroke patients beyond making the initial diagnosis, searching for a treatable cause, giving acute medical therapies and initiating secondary preventive measures. The physician needs to be much more involved in the activities of the team for several reasons: • Team leader: Usually, rightly or wrongly, political power in health services lies principally with doctors, and therefore if stroke patients are to have access to adequate facilities, the doctor must be involved. • Source of knowledge: The physician should be knowledgeable about the disease processes – both stroke and non-stroke – that underlie the patients’ functional problems, and should thus understand the prognosis and likely effect of interventions. This is essential for predicting outcome and setting appropriate goals. • Medical care underpinning rehabilitation: Stroke patients commonly have coexisting and complicating medical problems (e.g. diabetes, heart failure, deep venous thrombosis), which need to be identified and treated.4 • Ability to chair meetings effectively: The physician will often have the broadest knowledge about stroke and is therefore in a good – although not necessarily unique – position to coordinate the team and to chair team meetings. This role is important, because team meetings can become very time-consuming and may lose direction without a strong chairperson (section 10.3.7). Overly long meetings are boring, demoralizing and inefficient. • Legal responsibility: The legal responsibility for the patient is the doctor’s, and therefore the doctor must be involved at all stages of the patient’s care. In the UK, to encourage specialist physician involvement in stroke care and management, there has been a move towards establishing stroke medicine as a subspecialty with its own training schemes (http://www. basp.ac.uk/2004specialtraining.pdf). Stroke physicians are trained to deal with the whole patient journey, from the hyperacute phase through to rehabilitation and longterm follow-up, and to manage most post-stroke problems (Chapter 11). In other countries these roles are often distributed among other specialists including neurologists, internal medicine specialists, rehabilitation specialists and geriatricians. Nursing staff The nursing staff probably have the broadest role in the management of patients with stroke, which includes at least four major components:

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 527

10.3 Delivering an integrated management plan

• A daily assessment of the patient’s problems, both existing and new, and of their abilities and disabilities. Every week, our nurses assess the patients using the Barthel index, which focuses their attention on important functional issues. Based on this assessment, they evolve a care plan that aims to meet the individual’s needs. • Provision of all the basic needs of dependent patients after a stroke (e.g. feeding, washing, dressing, toileting, turning and transferring). • Provision of skilled nursing to prevent the development of complications such as pressure ulcers (section 11.16), painful shoulders (section 11.24), other injuries (section 11.26) and aspiration pneumonia (sections 11.12 and 11.17). This involves correctly positioning and handling the immobile patient (section 11.11). If one adopts a model of teamworking approximating to that shown in Fig. 10.19, then the nurse’s role will include many of the functions of other members of the team. • Supporting the patient and family. With increasing emphasis by managers and politicians on improving easily measured clinical outcomes such as case fatality and disability, other aspects of caring for and about patients are easily forgotten. But these are very important to patients and close family members64 – informing, reassuring, encouraging, advising, supporting and sympathizing. The nurses usually have greatest contact with the patients and family and therefore have an important role in this area. In some healthcare systems trained nurses are not readily available, and much of the patient’s day-to-day needs are met by relatives. In many countries specific training for stroke unit nurses is not yet available (http:// www.nationalstrokenursingforum.com/). To fully realize the potential of stroke unit nurses requires considerable investment in specialist education and training. Physiotherapist The physiotherapist has several important roles in caring for stroke patients, depending on their individual needs and the stage of the illness (http://www.csp.org.uk/). Soon after a severe stroke, the physiotherapist may be involved in at least seven functions: • Providing a detailed assessment of the motor and sensory problems of patients to help estimate their prognosis. • Assessing and treating chest problems, including pneumonia and retention of secretions. Speech and language therapists often find it useful to have a physiotherapist present at swallowing assessments, to help position the patient and to deal promptly with any aspiration (section 11.17).

• Advising nurses and other carers on the best way to position patients to prevent unhelpful changes in muscle tone that may lead ultimately to contractures and further limitation of function (section 11.20). • Teaching the nurses and informal carers the best way to handle the patient to avoid pain or injury to the patient or carer. This will often involve teaching proper methods of transferring, lifting, standing and walking the patient. • Providing therapy to relieve the symptoms associated with painful shoulders or swollen limbs (sections 11.24 and 11.25). • Providing therapy to improve the patient’s mobility and arm function (section 11.21). • Advising on walking aids and splints (section 11.32.1) that may sometimes improve a patient’s function. In some countries physiotherapists have a more limited role. They may not be so involved in assessment and formulation of treatment but simply deliver the interventions prescribed by a rehabilitation specialist. Occupational therapist Occupational therapists fulfil several roles in the management of stroke patients (http://www.cot.org.uk/ specialist/nanot/forums/stroke.php). These are usually fairly limited in the very early period after a severe stroke, but become more important as the patient recovers and as self-care becomes more relevant. These roles include the following five components: • An early assessment of patients to find out how each impairment is likely to restrict their function. This requires an assessment of what the patients were able to do before the stroke, what they can do now and what their home circ*mstances are, e.g. ease of access to the front door, bedroom, toilet, or bathroom; or the circ*mstances into which the patients will be discharged, e.g. a relative’s home or nursing home. • An assessment of the patient’s visuospatial functioning. It is important to remember that many evaluations of the objective tests of visuospatial functioning use an occupational therapist’s assessment as the ‘gold standard’65 (section 11.28). • Training the patient and the carer to carry out everyday activities, despite the patient’s impairments, is a crucial role (section 11.32). This involves finding the best way of achieving a particular activity for that individual patient. Most input is spent on the activities of daily living (Table 11.45), although in specialized units or those dealing with younger patients, therapy aimed at return to an occupation (section 11.33.2) or leisure activity (section 11.33.3) may be available.

527

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 528

528

Chapter 10 A practical approach to the management of stroke and transient ischaemic attack patients

• Provision of aids and adaptations to allow patients to function better. In the UK, this includes wheelchair provision, feeding and kitchen aids, and bathroom aids, among many others (section 11.32). • Assessment of patients’ ability to function in their own home, e.g. home visit which is often done before discharge with patients who have been admitted to hospital, and is an important element of discharge planning. It is important to identify problems that are specific to the patient’s own home environment and which may have to be solved by further training or by provision of aids or adaptations. Speech and language therapist Speech and language therapists have several roles in the care of stroke patients (http://www.rcslt.org/). They include: • Assessment of swallowing safety both initially and as the patients improve, so that their diet and fluid intake matches their swallowing abilities (section 11.17). • Teaching the patient, nurses or family who are involved with feeding, techniques that help overcome a swallowing impairment and avoid aspiration. • Teaching the patients exercises that may increase the rate of recovery of swallowing problems (section 11.17). • Diagnosis and assessment of the patient’s communication problems (section 11.30). • Informing both formal and informal carers of the nature of the patient’s communication problems. Patients, their families and even the professionals caring for the patient often have great difficulty in understanding what is wrong with a dysphasic patient. They may think the patient is ‘confused’, ‘demented’, or simply ‘mad’. Because problems with communication so often lead to emotional distress, most speech and language therapists also take on a counselling role. • Teaching patients, carers and even volunteers, strategies to allow the patients to communicate effectively using language (spoken or written), gesture, or communication aids where appropriate. • Providing therapy that may enhance the recovery of communication difficulties (section 11.30). Social worker The role of the social worker is bound to vary in different societies. In the UK, Australia and the Netherlands (which are the countries where we have direct experience), the social worker is likely to: • Provide patients and their families with practical advice and help at all stages of the patients’ illness. For example, arranging subsidized transport for the family to

visit the hospital, or extra home care for a dependent relative if the stroke patient was the main carer before admission. Social workers often help with any financial problems that have arisen because the main breadwinner has had a stroke, e.g. by applying for allowances or grants. • Be involved helping to plan the patient’s discharge from hospital, or change of accommodation. Social workers often spend a lot of time identifying the wishes and needs of the patient and family and then, with the rest of the team, trying to meet them. Where patients are unable to make decisions for themselves and have no close family, the social worker may need to act as the patient’s advocate and make arrangements for financial affairs to be taken care of, perhaps even arranging any change of accommodation (e.g. transfer to a nursing home). • Follow up patients and families after the patient’s discharge from hospital, to identify their changing need for support and to make adjustments to any care package. • Provide counselling that may be helpful in allowing patients and families to come to terms with the change in circ*mstances brought about by the stroke. Some organize groups of patients, carers, or both to help solve problems. One function that should be shared by all members of the team is that of monitoring the patient’s condition. Nurses and therapists, in particular, spend a lot of their time handling patients and observing the patient’s performance, so they are often in an ideal position to notice the relatively minor changes that may be an early sign of a complication, which might benefit from early treatment. All members of the team should be alert to changes in the patient’s condition that may indicate the development of a complication. Are the interventions of the team members effective? Although organized stroke care improves outcomes, some question the effectiveness of physiotherapy, occupational therapy, speech and language therapy and social work. They point to the lack of research evidence to support the effectiveness of these professionals (but interestingly, without questioning their own effectiveness), or they emphasize particular studies that appear to demonstrate the ineffectiveness of other professions. Unfortunately, these ‘negative’ studies have, in general, asked the wrong questions, measured the wrong outcomes, been too small, and have not evaluated the intervention in the context of a well-organized stroke team. This has possibly resulted in ‘false-negative’ studies,

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 529

10.3 Delivering an integrated management plan

MULTIDISCIPLINARY TEAM MEETINGS Addressograph

WARD: Date:

For CPR: Yes:

Week No:

No:

Barthel:

Milestones:

None, Sit, Stand, Steps, 10m

Falls Risk? YES/NO

Nursing: present /absent

Weight:

Medical: Consultant present /absent

Nutrition: Dietician present /absent

Physiotherapy: present /absent

Speech & Language therapy: present /absent

Occupational therapy: present /absent

Social Work: present /absent

Action Plan: 1. 2. 3. Discharge within one week? Y/N Signed:

Designation:

Fig. 10.20 A proforma for recording the results of a multidisciplinary team meeting.

resulting in rejection of valuable team members. But, as everyone must realize, failure to demonstrate a definite effect cannot be taken as proof of a lack of effect.66 There may be uncertainty about the effectiveness, optimum ‘dose’ and duration of some of the specific therapeutic

manoeuvres that therapists use (in common with many medical practices) but, as we have shown, this type of input forms only a small part of their contribution to the care of stroke patients. In fact, there are an increasing number of well-conducted randomized trials that have

529

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 530

530

Chapter 10 A practical approach to the management of stroke and transient ischaemic attack patients

established that interventions by therapists do improve patient outcome, as discussed in Chapter 11.67–69

10.3.7 Multidisciplinary team meetings The team must meet regularly if it is to work effectively. All the stroke units included in the recent systematic review of the randomized controlled trials held at least weekly meetings of the multidisciplinary team, separate from conventional ward rounds.70 These meetings have several important functions: • The entire team can be introduced to new patients and their problems. • Existing patients can be reviewed, and if individual team members have noted a change in their condition or a new problem, this can be communicated to the other members. • After reviewing each patient’s progress and any developing problems, realistic goals can be set jointly, and an appropriate course of action to meet those goals can be agreed (section 10.3.3). • Verbal reports of individual therapists’ assessments, and in particular the results of pre-discharge home visits, can be discussed and detailed plans for discharge can be made. • Meetings have an educational role for the team members, students and visitors. For example, we often show the patient’s brain CT and discuss the likely pathogenesis of the stroke and any rationale for treatment or therapy. So that important details are not overlooked, it may be useful to agree a formal structure to the discussions which can be reinforced by using a standard form on which team discussions are recorded (Fig. 10.20). The structure we have adopted is discussed below. This is not rigidly adhered to, but provides a framework for discussions about individual patients. Structure of team discussion The medical details of any new patients are presented by the physician, including a brief account of the patient’s symptoms and any relevant past history, including risk factors and the presumed cause of the stroke. The patient’s social background is also presented briefly by the physician, but other members of the team are often able to contribute extra details. The patient’s neurological impairments are discussed, and the therapists have often identified some that may not have been obvious to the physician. By the time of the meeting, the nursing staff can report on the functional consequences of these impairments (i.e. disability), and using the information about pre-stroke activities, some estimate of the stroke’s

likely effect on the patient’s life after the stroke can be made (i.e. predicted handicap). At each subsequent meeting, we introduce the patient briefly with a résumé of the date of stroke, clinical type and presumed cause. We then summarize the problems, goals and actions that were agreed at the last meeting. Each member of the team is then invited to update the rest of the team on the patient’s progress, what problems and goals have been solved or achieved, what new goals they plan to set, and how they plan to achieve them. Usually we do this in a set order that tends to reflect the way we think about patients, i.e. pathology, impairment, disability and handicap. The physician starts by updating the team on any changes in the patient’s medical condition or the results of any important investigations that may be relevant to other team members. The nurses then give an overall report on the patient’s progress, including current performance on the Barthel index. This usefully highlights many of the patient’s functional problems, as well as giving an objective indication of progress. The therapists follow: first the physiotherapist, then the speech therapist and lastly the occupational therapist. The first two tend to focus more on impairments and the patient’s basic functions such as mobility, swallowing and communication, while the occupational therapist usually focuses on the broader range of disabilities and how these are likely to affect the patient’s everyday life. Lastly, the social worker reports on any problems which close contact with the family may have revealed and progress regarding discharge planning. This sequence also has the advantage that the occupational therapist and social worker can make use of the information from the others in formulating their own goals and actions. The discussion is then opened up, and longer-term goals such as timing of home visits, discharges and case (family) conferences are set, and we decide who will do what by the next meeting including who will communicate with the patient and family to ensure that they are involved with the goal-setting process and that their views are being taken into account. Multidisciplinary teams may lack effectiveness if each member is not given an equal chance to contribute. Inevitably, teams will include some more assertive individuals and other less outgoing members, the former tending to dominate discussions. It is important that the team leader – often the physician – ensures that individual members do not monopolize the discussions to the extent that others are not heard. The framework we have adopted lessens the likelihood of this arising, although of course not every member will necessarily have something useful to contribute in every case.

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 531

10.4 Some difficult ethical issues

It is important that the person chairing the team meetings ensures that no individual member, including the chairperson, monopolizes the discussion to the extent that others are not heard.

We find it useful to have separate team meetings where patients are not discussed when we discuss general management issues, how the team is working, what the problems are, and any changes that might be made in the way we work.

Involving patients and carers in team meetings Neither the patient nor the family have been mentioned yet as having a direct input into the team meetings, although their input is obviously crucial in setting goals and planning discharge from hospital. We do not invite them to attend each meeting, but we do seek their views on certain issues in advance of the meeting. We also try to ensure that the nature and conclusions of the discussion at the meeting are communicated back to the patient and/or family by the most appropriate team member, most often the nurse. Although it is sometimes vital to involve the patient and families in discussions, we feel that their attendance at each weekly meeting would inhibit the discussion and might be distressing for them. It would also make meetings longer and more unwieldy. We hold separate team meetings (case conferences or family meetings) to which the patient, family and any other people who are involved – such as the district (community) nurse and home care organizer – are invited. These meetings are useful in planning hospital discharge in complex cases and in resolving differences of opinion between the team members and the patient or family. The timing of these meetings is tailored to the individual patient. Team building It is useful to have separate team meetings at which patients are not discussed. On these occasions, we discuss general management issues, how the team is working, what the problems are, and any changes that might be made in the way we work. We discuss critical incidents to ensure that we learn from these. These sessions also allow individuals to tell the rest of the team about certain specialized aspects of their jobs. This, we believe, not only encourages some blurring of the distinctions between their functions but also aids team building. Each member can more fully appreciate the contribution made by other members. We also have weekly meetings, involving the consultants and trainees, to discuss patients entering our stroke register where we review their clinical and radiological details. These provide an additional opportunity for education, training, quality control and discussion of difficult cases with consultant colleagues.

Recording the work of the team We have not yet found a totally satisfactory method of recording the work of the team. Traditionally, members have each kept their own records so that they may refer to them, but this inevitably leads to much unnecessary duplication. However, a unified record has disadvantages, the main one being that different members of the team require information of varying detail. Thus, the physiotherapist’s record contains very detailed notes on the patient’s motor functioning that are irrelevant to the social worker. Also, the records may be needed simultaneously by the physiotherapist in the gymnasium and by the pharmacist on the stroke unit. In our units, each profession has therefore kept its own records, but they do all make entries in a combined medical and nursing record. The nurses base their care plans on the discussions at the team meetings. This has been helped by their adoption of a problem-orientated approach, so that the nurses regularly record the patients’ problems, the goals and actions aimed at attaining the goals. A compromise in which core details are collected once and shared, while more detailed records are kept by each team member, might provide the best and most efficient solution. We also increasingly use structured records in which the results of the most commonly used assessments (e.g. swallowing, continence) and actions taken require only a tick, date and signature. In the future, increased use of information technology may facilitate a combined patient record that avoids some of these problems but even now we do ensure that changes in the patient’s condition, key decisions and proposed actions (by whom and by when) are recorded in the medical notes. This is important to avoid unnecessary and time-wasting repetition of discussions at future meetings and to ensure that decisions lead on to actions.

10.4 Some difficult ethical issues

When patients are judged unlikely to achieve a reasonable quality of life if they were to survive, many

531

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 532

532

Chapter 10 A practical approach to the management of stroke and transient ischaemic attack patients

clinicians decide not to strive to ensure survival. However, this is an extremely difficult decision since, as we have already discussed, there are considerable difficulties in predicting outcome reliably at an early stage (section 10.2.7). Also the value that the patient would place on survival in a given state is impossible to judge. Patients who have had a major stroke are often not in a position to make decisions about their treatment because, due to a depressed conscious level, language or visuospatial problems, they are unable to understand or communicate. Under these circ*mstances, the family or potential carers should be included in the decision-making process. The family can often provide information allowing judgements to be made about what the patient’s likely wishes are. However, some relatives may have their own interests and not just those of the patient to consider. Younger relatives may gain financially from the death of an older relative, and death removes the obligation to care for a severely disabled person. Patients may even have written down what they would like to happen in the event of a life-threatening illness (i.e. a living will). The legal status of these so-called ‘living wills’ varies from country to country, however. Also, it is not at all clear whether patients’ judgements made while in good health about the value they might place on their life if they were in a dependent state remain true when they have become disabled. We have encountered many people who have become severely disabled and yet have accepted their disability and felt that their quality of life is acceptable or even very good. It is difficult to know how to react to patients’ previously expressed view that they would never, ever want to live in a dependent state or go to a nursing home, when they then seem to cling to life after a major stroke. One has to be particularly careful in assessing the likely prognosis of an individual patient with an apparently severe stroke. A relatively minor stroke that is complicated by another medical problem (e.g. pneumonia) may be impossible to differentiate from a severe stroke (section 11.4). If the complicating medical problem can be identified and treated then the patient may achieve a very satisfactory clinical outcome despite the apparently gloomy initial prognosis. It is not always easy to distinguish between a patient who has had a catastrophic stroke and one who has had a less severe stroke that is complicated and worsened by infection, epileptic seizures or metabolic problems. The prognosis is much worse for the former patient than the latter one.

Perhaps the most common dilemma is whether or not to give fluids or nutritional support to a patient with an ‘apparently severe stroke’ who is unable to swallow. Early nutritional support is likely to improve the proportion of patients surviving although not perhaps the functional status of survivors.71 Few would argue against giving parenteral fluids to a conscious patient to prevent dehydration, thirst and discomfort. However, if the patient is unconscious and thus probably unaware of any discomfort, then there is more uncertainty about what to do for the best. Also, if the patient develops an infection very soon after the onset of the stroke, apart from the difficulties mentioned above in assessing the severity of stroke, the question arises as to how aggressively one should treat the patient. Are intravenous antibiotics, physiotherapy or even artificial ventilation and inotropic support appropriate in this situation? There are some facts worth considering before making these difficult decisions. Most patients who are unconscious soon after a stroke die in the first few days from the direct neurological effects of the stroke on brainstem function, and not from dehydration or infection (section 10.2.3). Therefore, by simply giving fluids, it is unlikely that the lives of many patients will be greatly prolonged, but this may give more time to make an accurate assessment of prognosis. If one elects to give a patient fluids, antibiotics or even nutritional support via a feeding tube, this clearly does not have to be continued on a long-term basis. However, in some countries there are laws that prevent clinicians from withdrawing treatment of this kind, and many clinicians feel more uncomfortable about stopping a treatment that they have started than about not starting it in the first place. Apart from worrying about medicolegal issues, one has to consider the reaction of the family. However, in our experience, if the physician has spent enough time informing the family of the patient’s state and likely prognosis, and has involved them in the decisions, then withdrawal of fluids, feeding or antibiotics is seldom a major problem. Unfortunately, in some countries there is a trend for such decisions – especially in severely brain-damaged children and patients in a persistent vegetative state – to be delegated to the legal profession. We cannot hope to have the answers to all these difficult problems, but perhaps it is worth suggesting a general approach that we have found practicable: • First, collect accurate and unbiased information about the patient’s life before the stroke. It is not good enough to base one’s decisions on the severity of just the stroke, since this may be difficult to estimate accurately, and other factors will influence the prognosis (section 11.4). The patient’s pre-stroke functional level and social activity are crucial in making these

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 533

References

decisions. Unfortunately, this information is often not collected in sufficient detail during the acute phase, because its relevance to acute care is not recognized (section 10.3.2). One might be less aggressive in treating patient A, who was previously handicapped, miserable and in pain from severe arthritis and living in an institution, than patient B, who has an equally severe stroke but was previously living at home independently with his or her family. Usually, the patient’s pre-stroke handicap sets the ceiling on the potential post-stroke outcome, so that these two patients have very different outlooks. It is still a difficult – some would say an impossible – judgement to say whether patient A’s life is worth prolonging. • Next, it is important to make a detailed assessment of the patient’s condition in order to make the best possible judgement about the prognosis. • Third, formulate the various alternative management plans that might be adopted. These should include those from the most aggressive to the most conservative. • Finally, all options must be discussed with other members of the team and any close family members. The family needs to be given accurate information about the diagnosis, problems, likely outcomes and treatment possibilities. It is probably unfair to suggest that the decisions are all theirs, but any decision the clinician makes should ideally be compatible with their views. If one does not take the family’s wishes into account one is inviting trouble, complaints – and nowadays litigation. It is worth making it clear to them that few decisions are irreversible (except perhaps turning off a ventilator), since circ*mstances change. One is often in a much better position to make an accurate prognosis having observed the patient’s progress over a few days than on the day of the stroke, so what appeared to be an appropriate intervention on day 1 may appear less appropriate after a few days. It can also be useful to involve an independent and experienced clinician to provide a second opinion. The development of better tools to help predict prognosis, and studies that give a better understanding of the effect of our supportive interventions, should make these difficult decisions somewhat easier in the future. Professional bodies often provide specific advice which is applicable to particular countries (e.g. http://www. gmc-uk.org/standards/default.htm).

Involve other members of the team and the patient’s family in any decisions about how aggressively one should strive to keep a patient alive. Emphasize that most decisions can be reviewed – and reversed – in the light of the patient’s progress.

References

1 WHO. International Classification of Impairments, Disabilities and Handicaps. Geneva: World Health Organization; 1980. 2 WHO. International Classification of Functioning, Disability and Health (ICF). Geneva: World Health Organization; 2001. 3 Bornman J. The World Health Organization’s terminology and classification: application to severe disability. Disabil Rehabil 2004; 26(3):182–8. 4 Langhorne P, Stott DJ, Robertson L, MacDonald J, Jones L, McAlpine C et al. Medical complications after stroke: a multicenter study. Stroke 2000; 31(6):1223–9. 5 Bamford J, Sanderco*ck P, Dennis M, Burn J, Warlow C. A prospective study of acute cerebrovascular disease in the community: the Oxfordshire Community Stroke Project 1981–86. 2. Incidence, case fatality rates and overall outcome at one year of cerebral infarction, primary intracerebral and subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry 1990; 53(1):16–22. 6 Rothwell PM, Coull AJ, Giles MF, Howard SC, Silver LE, Bull LM et al. Change in stroke incidence, mortality, case-fatality, severity, and risk factors in Oxfordshire, UK from 1981 to 2004 (Oxford Vascular Study). Lancet 2004; 363(9425):1925–33. 7 Bonita R, Broad JB, Beaglehole R. Changes in stroke incidence and case-fatality in Auckland, New Zealand, 1981–91. Lancet 1993; 342(8885):1470–3. 8 Peltonnen M, Stegmayr B, Asplund K. Marked improvement since 1985 in short-term and long-term survival after stroke. Cerebrovasc Dis 1999; 9(Suppl. 1): 62. 9 Gresham GE, Kelly-Hayes M, Wolf PA, Beiser AS, Kase CS, D’Agostino RB. Survival and functional status 20 or more years after first stroke: the Framingham Study. Stroke 1998; 29(4):793–7. 10 Bamford J, Dennis M, Sanderco*ck P, Burn J, Warlow C. The frequency, causes and timing of death within 30 days of a first stroke: the Oxfordshire Community Stroke Project. J Neurol Neurosurg Psychiatry 1990; 53(10):824–9. 11 Burn J, Dennis M, Bamford J, Sanderco*ck P, Wade D, Warlow C. Long-term risk of recurrent stroke after a firstever stroke. The Oxfordshire Community Stroke Project. Stroke 1994; 25(2):333–7. 12 Van GJ. Subarachnoid haemorrhage. Lancet 1992; 339(8794):653–5. 13 Dombovy ML, Basford JR, Whisnant JP, Bergstralh EJ. Disability and use of rehabilitation services following stroke in Rochester, Minnesota, 1975–1979. Stroke 1987; 18(5):830–6. 14 Kwakkel G, Kollen B, Lindeman E. Understanding the pattern of functional recovery after stroke: facts and theories. Restor Neurol Neurosci 2004; 22(3–5):281–99. 15 Gray CS, French JM, Bates D, Cartlidge NE, James OF, Venables G. Motor recovery following acute stroke. Age Ageing 1990; 19(3):179–84.

533

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 534

534

Chapter 10 A practical approach to the management of stroke and transient ischaemic attack patients

16 Duncan PW, Goldstein LB, Matchar D, Divine GW, Feussner J. Measurement of motor recovery after stroke. Outcome assessment and sample size requirements. Stroke 1992; 23(8):1084–9. 17 Ashburn A. Physical recovery following stroke. Physiotherapy 1997; 83:480–90. 18 Kwon S, Hartzema AG, Duncan PW, Min-Lai S. Disability measures in stroke: relationship among the Barthel Index, the Functional Independence Measure, and the Modified Rankin Scale. Stroke 2004; 35(4):918–23. 19 Dennis MS. Outcome after brain haemorrhage. Cerebrovasc Dis 2003; 16 Suppl 1:9–13. 20 Barber M, Roditi G, Stott DJ, Langhorne P. Poor outcome in primary intracerebral haemorrhage: results of a matched comparison. Postgrad Med J 2004; 80(940):89–92. 21 Lipson DM, Sangha H, Foley NC, Bhogal S, Pohani G, Teasell RW. Recovery from stroke: differences between subtypes. Int J Rehabil Res 2005; 28(4):303–8. 22 Feigin VL, Lawes CM, Bennett DA, Anderson CS. Stroke epidemiology: a review of population-based studies of incidence, prevalence, and case-fatality in the late 20th century. Lancet Neurol 2003; 2(1):43–53. 23 Tanaka H, Ueda Y, Date C, Baba T, Yamash*ta H, Hayashi M et al. Incidence of stroke in Shibata, Japan: 1976–1978. Stroke 1981; 12(4):460–6. 24 Asplund K, Bonita R, Kuulasmaa K, Rajakangas AM, Schaedlich H, Suzuki K et al. Multinational comparisons of stroke epidemiology. Evaluation of case ascertainment in the WHO MONICA Stroke Study. World Health Organization Monitoring Trends and Determinants in Cardiovascular Disease. Stroke 1995; 26(3):355–60. 25 Feigin VL, Nikitin I, Kholodov VA, Shishkin SV, Novokhatskaia MV, Belenko AI, Khatsenko VN. [The epidemiology of cerebral stroke in Siberia based on registry data]. Zh Nevropatol Psikhiatr Im S S Korsakova 1996; 96(6):59–65. 26 Petty GW, Brown RD, Jr., Whisnant JP, Sicks JD, O’Fallon WM, Wiebers DO. Ischemic stroke: outcomes, patient mix, and practice variation for neurologists and generalists in a community. Neurology 1998; 50(6):1669–78. 27 Wijdicks EF, Rabinstein AA. Absolutely no hope? Some ambiguity of futility of care in devastating acute stroke. Crit Care Med 2004; 32(11):2332–42. 28 Counsell C, Dennis M. Systematic review of prognostic models in patients with acute stroke. Cerebrovasc Dis 2001; 12(3):159–70. 29 Meijer R, Ihnenfeldt DS, van Limbeek J, Vermeulen M, de Haan RJ. Prognostic factors in the subacute phase after stroke for the future residence after six months to one year. A systematic review of the literature. Clin Rehabil 2003; 17(5):512–20. 30 Meijer R, Ihnenfeldt DS, de Groot IJ, van Limbeek J, Vermeulen M, de Haan RJ. Prognostic factors for ambulation and activities of daily living in the subacute phase after stroke. A systematic review of the literature. Clin Rehabil 2003; 17(2):119–29. 31 Gladman JR, Harwood DM, Barer DH. Predicting the outcome of acute stroke: prospective evaluation of five

multivariate models and comparison with simple methods. J Neurol Neurosurg Psychiatry 1992; 55(5):347–51. Taub NA, Wolfe CD, Richardson E, Burney PG. Predicting the disability of first-time stroke sufferers at 1 year. 12month follow-up of a population-based cohort in southeast England. Stroke 1994; 25(2):352–7. Wardlaw JM, Lewis SC, Dennis MS, Counsell C, McDowall M. Is visible infarction on computed tomography associated with an adverse prognosis in acute ischemic stroke? Stroke 1998; 29(7):1315 –9. Nuutinen J, Liu Y, Laakso MP, Karonen JO, Roivainen R, Vanninen RL et al. Assessing the outcome of stroke: a comparison between MRI and clinical stroke scales. Acta Neurol Scand 2006; 113(2):100–7. Hendricks HT, Zwarts MJ, Plat EF, van Limbeek J. Systematic review for the early prediction of motor and functional outcome after stroke by using motor-evoked potentials. Arch Phys Med Rehabil 2002; 83(9):1303–8. Kwakkel G, Wagenaar RC, Kollen BJ, Lankhorst GJ. Predicting disability in stroke: a critical review of the literature. Age Ageing 1996; 25(6):479–89. Counsell C, Dennis M, McDowall M, Warlow C. Predicting outcome after acute and subacute stroke: development and validation of new prognostic models. Stroke 2002; 33(4):1041–7. Weingarten S, Bolus R, Riedinger MS, Maldonado L, Stein S, Ellrodt AG. The principle of parsimony: Glasgow Coma Scale score predicts mortality as well as the APACHE II score for stroke patients. Stroke 1990; 21(9):1280–2. Wright CJ, Swinton LC, Green TL, Hill MD. Predicting final disposition after stroke using the Orpington Prognostic Score. Can J Neurol Sci 2004; 31(4):494–8. Horgan NF, Cunningham CJ, Coakley D, Walsh JB, O’Neill D, O’Regan M et al. Validating the Orpington Prognostic Score in an Irish in-patient stroke population. Ir Med J 2005; 98(6):172, 174–5. Weir NU, Counsell CE, McDowall M, Gunkel A, Dennis MS. Reliability of the variables in a new set of models that predict outcome after stroke. J Neurol Neurosurg Psychiatry 2003; 74(4):447–51. Dennis MS, Lewis SC, Warlow C. Routine oral nutritional supplementation for stroke patients in hospital (FOOD): a multicentre randomised controlled trial. Lancet 2005; 365(9461):755–63. Counsell C, Dennis MS, Lewis S, Warlow C. Performance of a statistical model to predict stroke outcome in the context of a large, simple, randomized, controlled trial of feeding. Stroke 2003; 34(1):127–33. Frankel MR, Morgenstern LB, Kwiatkowski T, Lu M, Tilley BC, Broderick JP et al. Predicting prognosis after stroke: a placebo group analysis from the National Institute of Neurological Disorders and Stroke rt-PA Stroke Trial. Neurology 2000; 55(7):952–9. Weimar C, Konig IR, Kraywinkel K, Ziegler A, Diener HC. Age and National Institutes of Health Stroke Scale Score within 6 hours after onset are accurate predictors of outcome after cerebral ischemia: development and external validation of prognostic models. Stroke 2004; 35(1):158–62.

9781405127660_4_010.qxd 10/13/07 11:04 AM Page 535

References

46 Muir KW, Weir CJ, Murray GD, Povey C, Lees KR. Comparison of neurological scales and scoring systems for acute stroke prognosis. Stroke 1996; 27(10):1817–20. 47 Lai SM, Duncan PW, Keighley J. Prediction of functional outcome after stroke: comparison of the Orpington Prognostic Scale and the NIH Stroke Scale. Stroke 1998; 29(9):1838– 42. 48 Tilling K, Sterne JA, Rudd AG, Glass TA, Wityk RJ, Wolfe CD. A new method for predicting recovery after stroke. Stroke 2001; 32(12):2867–73. 49 Tilling K, Sterne JA, Wolfe CD. Multilevel growth curve models with covariate effects: application to recovery after stroke. Stat Med 2001; 20(5):685–704. 50 Counsell C, Dennis M, McDowall M. Predicting functional outcome in acute stroke: comparison of a simple six variable model with other predictive systems and informal clinical prediction. J Neurol Neurosurg Psychiatry 2004; 75(3):401–5. 51 Harbison J, Hossain O, Jenkinson D, Davis J, Louw SJ, Ford GA. Diagnostic accuracy of stroke referrals from primary care, emergency room physicians, and ambulance staff using the face arm speech test. Stroke 2003; 34(1):71–6. 52 Nor AM, McAllister C, Louw SJ, Dyker AG, Davis M, Jenkinson D, Ford GA. Agreement between ambulance paramedic- and physician-recorded neurological signs with Face Arm Speech Test (FAST) in acute stroke patients. Stroke 2004; 35(6):1355–9. 53 Rothwell PM, Giles MF, Flossmann E, Lovelock CE, Redgrave JN, Warlow CP, Mehta Z. A simple score (ABCD) to identify individuals at high early risk of stroke after transient ischaemic attack. Lancet 2005; 366(9479):29–36. 54 Hill MD, Weir NU. Is the ABCD score truly useful? Stroke 2006; 37(7):1636. 55 Davenport RJ, Dennis MS, Warlow CP. Improving the recording of the clinical assessment of stroke patients using a clerking pro forma. Age Ageing 1995; 24(1):43–8. 56 Hanco*ck RJ, Oddy M, Saweirs WM, Court B. The RCP stroke audit package in practice. J R Coll Physicians Lond 1997; 31(1):74 –8. 57 Bero LA, Grilli R, Grimshaw JM, Harvey E, Oxman AD, Thomson MA. Closing the gap between research and practice: an overview of systematic reviews of interventions to promote the implementation of research findings. The Cochrane Effective Practice and Organization of Care Review Group. Br Med J 1998; 317(7156):465–8. 58 Kwan J, Sanderco*ck P. In-hospital care pathways for stroke. [update of Cochrane Database Syst Rev 2002; (2):CD002924; PMID: 12076460]. Cochrane Database Syst Rev 2004; (4):CD002924. 59 Smith MT, Baer GD. Achievement of simple mobility milestones after stroke. Arch Phys Med Rehabil 1999; 80(4):442–7. 60 Wade DT, Wood VA, Heller A, Maggs J, Langton HR. Walking after stroke. Measurement and recovery over the first 3 months. Scand J Rehabil Med 1987; 19(1):25–30.

61 Parry RH. Communication during goal-setting in physiotherapy treatment sessions. Clin Rehabil 2004; 18(6):668–82. 62 Wressle E, Eeg-Olofsson AM, Marcusson J, Henriksson C. Improved client participation in the rehabilitation process using a client-centred goal formulation structure. J Rehabil Med 2002; 34(1):5–11. 63 Bernhardt J, Dewey H, Thrift A, Donnan G. Inactive and alone: physical activity within the first 14 days of acute stroke unit care. Stroke 2004; 35(4):1005–9. 64 Pound P, Bury M, Gompertz P, Ebrahim S. Stroke patients’ views on their admission to hospital. Br Med J 1995; 311(6996):18–22. 65 Stone SP, Wilson B, Wroot A, Halligan PW, Lange LS, Marshall JC, Greenwood RJ. The assessment of visuo-spatial neglect after acute stroke. J Neurol Neurosurg Psychiatry 1991; 54(4):345–50. 66 Altman DG, Bland JM. Absence of evidence is not evidence of absence. Br Med J 1995; 311(7003):485. 67 Greener J, Enderby P, Whurr R. Speech and language therapy for aphasia following stroke. Cochrane Database Syst Rev 2000; (2):CD000425. 68 Walker MF, Leonardi-Bee J, Bath P, Langhorne P, Dewey M, Corr S et al. Individual patient data meta-analysis of randomized controlled trials of community occupational therapy for stroke patients. Stroke 2004; 35(9):2226–32. 69 Steultjens EM, Dekker J, Bouter LM, Van de Nes JC, Cup EH, Van den Ende CH. Occupational therapy for stroke patients: a systematic review. Stroke 2003; 34(3):676–87. 70 Langhorne P, Pollock A. What are the components of effective stroke unit care? Age Ageing 2002; 31(5):365–71. 71 Dennis MS, Lewis SC, Warlow C. Effect of timing and method of enteral tube feeding for dysphagic stroke patients (FOOD): a multicentre randomised controlled trial. Lancet 2005; 365(9461):764 –72. 72 Sackett DL, Haynes RB, Guyatt GH, Tugwell MD. Clinical Epidemiology: A Basic Science for Clinical Medicine. 2nd ed. Boston: Little, Brown, 1991. 73 Al-Shahi R, Vousden C, Warlow C. Bias from requiring explicit consent from all participants in observational research: prospective, population based study. Br Med J 2005; 331(7522):942. 74 Counsell C. The Prediction of Outcome in Patients with Acute Stroke. Cambridge: University of Cambridge; 1998. 75 Dennis MS, Burn JP, Sanderco*ck PA, Bamford JM, Wade DT, Warlow CP. Long-term survival after first-ever stroke: the Oxfordshire Community Stroke Project. Stroke 1993; 24(6):796–800. 76 Langhorne P, Murray GD, Stott DJ, McAlpine C, MacDonald J. Profiles of Recovery after Stroke: Their Value in Monitoring Recovery and Predicting Outcome. Chief Scientists Office report. K/RED/4/C284, in press.

535

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 537

What are this patient’s problems? A problem-based approach to the general management of stroke

11.1

Introduction

11.2

Airway, breathing and circulation

11.3

Reduced level of consciousness

11.4

Severe stroke vs apparently severe stroke

11.5

Worsening after a stroke

11.6

Coexisting medical problems

11.7

High and low blood pressure after stroke

11.8

Epileptic seizures

11.9

Headache, nausea and vomiting

11.10 Hiccups

538 538

543 544

545 548 549

551 553

553

11.11 Immobility and poor positioning 11.12 Fever and infection

553

555

11.13 Venous thromboembolism

557

11.14 Urinary incontinence and retention

560

11.15 Faecal incontinence and constipation 11.16 Pressure ulcers

564

11.17 Swallowing problems

567

11.18 Metabolic disturbances 11.19 Nutritional problems

572

575

11.20 Spasticity and contractures

579

11.21 Limb weakness, poor truncal control and unsteady gait 11.22 Sensory impairments

11.23 Pain (excluding headache) 11.24 Painful shoulder

587

588

11.25 Swollen and cold limbs 11.26 Falls and fractures 11.27 Visual problems

590

591 592

11.28 Visuospatial dysfunction 11.29 Cognitive dysfunction

593 595

11.30 Communication difficulties 11.31 Psychological problems

598

599

11.32 Dependency in activities of daily living 11.33 Social difficulties 11.34 Carer problems

583

586

603

614 617

537

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 538

538

Chapter 11 What are this patient’s problems?

11.1 Introduction

The patient’s general management, as distinct from the specific treatment of their stroke (Chapters 12–14), is primarily aimed at identifying and solving existing problems, as well as anticipating and preventing potential problems, at different stages of their illness. This chapter will cover the common problems which occur in patients who have had a stroke. Each section is loosely structured as follows: • General description of each problem which includes a definition, its frequency, causes and clinical significance, and prognosis. • Assessment, including methods of detection, simple clinical assessments and measures which may be appropriate for goal setting, audit or research. • Prevention and treatment, including interventions which may reduce the risk of a problem developing or hasten recovery. Post-stroke problems have seldom been systematically identified in community-based incidence studies (section 17.11.1) so their frequency in unselected populations is often unknown. Moreover, compared with other aspects of treatment there are few large randomized controlled trials (RCTs) or systematic reviews addressing these problems. This is partly because there are major

methodological difficulties in performing RCTs and systematic reviews to evaluate non-pharmacological interventions (Table 11.1). However, the evidence base is improving. Existing trials are being systematically appraised (www.effectivestrokecare.org) and more trials are being performed. Where there is no high-quality evidence the content of this chapter reflects our own clinical experience. Many of the topics are not specific to stroke medicine but might be found in any textbook of internal medicine or surgery. Therefore, we have not attempted to provide comprehensive reviews, but instead have concentrated on those aspects of assessment and management which are particularly relevant to stroke patients.

11.2 Airway, breathing and circulation

Inadequate airway, breathing or circulation are lifethreatening. Urgent resuscitative measures must be taken. Even if not an immediate threat to survival, it seems sensible to optimize the delivery of oxygen and glucose to the brain to minimize brain damage and so achieve the best possible outcome for the patient (section 12.1).

Table 11.1 Problems in performing randomized controlled trials and systematic reviews of physiotherapy interventions. Lack of theoretical model or rationale founded on basic science for many interventions Ethical problems of performing randomized controlled trials due to therapists’ certainty about the effectiveness of their own therapy in individual patients Difficulty in getting patients and therapists, to accept the possibility of ‘no treatment’ Strong patient preferences based on their belief in the benefits and acceptability of therapy Difficulty in blinding patients to treatment allocation Difficulty in designing convincing placebo treatments, or establishing appropriate ‘controls’ Difficulties in defining a therapy (in terms of type, dose, frequency, timing and duration) which has to be tailored to the individual patient; this leads to difficulties in applying the results in practice Failure to identify key components or interactions between components of an intervention Moderate treatment effects mean that large numbers of patients have to be randomized The need to randomize large numbers of patients may necessitate multicentre trials but these raise difficulties in standardizing and monitoring treatment Therapy is very labour-intensive and therefore expensive, so that bodies which fund research may not be willing to fund the therapy Difficulties agreeing suitable measures of outcome which are sensitive to both what therapists expect to influence, and what is relevant to the patient and family In unblinded trials, patient loyalty to a therapist may bias their responses to subjective outcome scales Various different outcome measures which hinders systematic review of trials Problems due to complex interactions with other therapies given simultaneously

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 539

11.2 Airway, breathing and circulation

11.2.1 Obstructed airway Patients with a decreased level of consciousness, impaired bulbar function or who have aspirated may have an obstructed or partially obstructed airway. Central cyanosis, noisy airflow with grunting, snoring or gurgling, an irregular breathing pattern and indrawing of the suprasternal area and intercostal muscles may indicate an obstruction. Transient obstruction is common in the acute phase of stroke during sleep (section 11.2.2) and it is, therefore, important that apnoeic spells due to an obstructed airway are not mistakenly attributed to

Fig. 11.1 It is essential to maintain a patent airway in a patient with a decreased level of consciousness. (a) In comatose patients with their head in a resting or flexed position the tongue falls backwards to obstruct the hypopharnyx (arrow) and the epiglottis obstructs the larynx. (b) Tilting the head back by lifting the chin stretches the anterior neck structures, which opens the airway.

(a)

Fig. 11.2 Airways. (a) Oropharyngeal (Guedel) airway. To check that the airway is the correct size for the patient hold the airway across the cheek from corner of mouth to tip of ear lobe (the angle of the jaw bone may also be used). If this matches then the airway is the correct size. If too long or too short select another size. Incorrect sizing can result in the airway being blocked rather than kept open. Inserted upside down, they are moved along the roof of the mouth and rotated into position

periodic respiration and so ignored (e.g. Cheyne–Stokes; section 11.2.2). If an obstructed airway is suspected, the oropharynx should be cleared of any foreign matter with a sweep of a gloved finger, the patient’s jaw pulled forward, and the neck extended to stop the tongue falling back to obstruct the airway (Fig. 11.1). Positioning the patient in the coma (i.e. recovery) position may be enough to keep the airway clear, although in some situations an oropharyngeal or nasopharyngeal airway, or even endotracheal intubation, may be required (Fig. 11.2).

(b)

sitting behind the tongue and lifting it forward slightly. (b) Nasopharyngeal airway. To measure an appropriately sized nasopharyngeal airway, it is traditionally taught that the correct size is related to the patient’s little finger or nostril. However these measures may not relate to the internal anatomy of their nose so it may more appropriate to size the airway by the patient’s sex and height.457,458

539

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 540

540

Chapter 11 What are this patient’s problems?

Thorax

≥10s Abdomen

Nasal Flow

≥10s

Fig. 11.3 Diagram showing Cheyne– Stokes breathing (courtesy of A Rowat); upper two traces show muscle activity, lower trace air flow.

Do not mistakenly attribute apnoeic spells due to intermittent airway obstruction to periodic or Cheyne–Stokes respiration.

simply reflects co-existing severe cardiorespiratory disease occurring in severe strokes.3 Abnormal patterns of breathing

11.2.2 Inadequate breathing Strokes may weaken the intercostal muscles and the diaphragm, leading to reduced ventilation, poor cough and an increased risk of pneumonia.1 Furthermore, the brain lesion itself may directly, or more often indirectly, impair the function of the respiratory centre in the medulla which results in various disordered patterns of breathing (Fig. 11.3) while the patient is awake, but much more commonly during sleep.2 Hypoxia Hypoxic episodes (defined as saturations of < 90% for > 10% of monitoring period) have been described in about one-fifth of stroke patients within the first few hours of admission.3 They occur more commonly during transfers between hospital departments and are associated with greater stroke severity and pre-existing cardiorespiratory disease. In patients with cardiorespiratory disease, hypoxia may be more marked in the supine position and reduced by sitting.4 Hypoxia is associated with poorer survival but it is unclear whether this is an effect of hypoxia exacerbating the brain damage, or

The abnormal patterns of breathing associated with stroke include obstructive and central sleep apnoea, periodic respiration (Cheyne–Stokes), hyperventilation (‘forced respiration’), irregular (ataxic) breathing, apneustic (held in expiration) breathing and, ultimately, complete apnoea.5 Sleep apnoea is the commonest abnormality and has been identified in up to two-thirds of hospital-admitted patients, depending on the definitions and detection methods used. It is usually ‘obstructive’ although may occasionally be ‘central’. Obstructive sleep apnoea has been associated with greater age, obesity, large neck circumference, and in some studies greater stroke severity.2 It is associated with poorer survival and functional outcome but it is unclear whether this is due to incomplete adjustment for stroke severity, the associated hypoxia or fluctuations in blood pressure and cerebral perfusion.6,7 Periodic respiration (Cheyne–Stokes), where there are regular alternating phases of hyperventilation and hypoventilation, occurs in about a quarter of patients although it is seldom recognized by healthcare staff.8 Other neurological functions, e.g. wakefulness, often vary in phase with breathing. The mechanisms of periodic respiration are still debated but it probably reflects a change in the sensitivity of the brainstem respiratory

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 541

11.2 Airway, breathing and circulation

centre to the arterial pressure of carbon dioxide and/or slowed central circulation, with a resultant delay in the feedback loop controlling respiration.9,10 Changes in oxygenation, pH and cerebral blood flow occur in periodic respiration but their importance is unclear. Periodic respiration is associated with greater stroke severity and poorer outcomes although it may be present in patients who are alert and who subsequently make a reasonable recovery.8 Periodic respiration (Cheyne–Stokes) does not necessarily imply a hopeless prognosis. Co-existing cardiopulmonary disease (e.g. chronic obstructive pulmonary disease, pneumonia) is probably a more frequent cause of inadequate ventilation than the abnormal breathing patterns due to the stroke itself. Its detection depends on a thorough initial assessment including an adequate history, physical examination and some simple investigations, e.g. chest X-ray and electrocardiogram (section 10.3.2). Assessment The adequacy of ventilation should be assessed clinically by checking for central cyanosis and examining the chest. We increasingly use pulse oximetry to assess and monitor our patients to detect hypoxaemia. Finger probes are more accurate than those on the ear, and it probably does not matter whether the probe is placed on the paretic or non-paretic hand.11,12 Arterial blood gas analysis is useful in monitoring therapy in patients with type II respiratory failure, those requiring artificial ventilation and those with severe metabolic disturbances such as acidosis. It is easy to miss abnormalities in breathing pattern. Encourage all members of the team to observe the breathing pattern while they are assessing the patient.

Treatment There are currently no data to support the routine administration of increased concentrations of oxygen to patients who are not hypoxic so do not automatically give oxygen to every admitted stroke patient.13 However, if the patient has an oxygen saturation of < 95% in the first couple of days after stroke onset it seems sensible (although not supported by strong evidence) to increase incrementally the inspired oxygen concentration to achieve a saturation of > 95% and so reduce the the likelihood that any ischaemic but still viable brain tissue is

adversely affected. Patients prefer oxygen delivered via nasal prongs rather than with a face mask.14 In patients with chronic lung disease and a tendency to hypercapnoea, 24% oxygen should be used initially with careful monitoring of the patient’s respiration, neurological function (section 11.5) and arterial blood gases. Although continuous positive airways pressure (CPAP) has been shown to reduce symptoms of obstructive sleep apnoea in general, in patients with recent stroke compliance with treatment tends to be very poor and there is no reliable evidence of benefit.15–17 Treatment of any coexisting lung or cardiac disease should of course be given. Sitting the patient up may help to improve oxygenation.4 If patients are sat out of bed, it is important that they are well supported in an upright position and not slumped in the chair, which worsens ventilation. Sitting may also reduce the intracranial pressure but may cause other problems (section 11.11). Sedative drugs, given to promote sleep, to facilitate imaging or to control seizures, may precipitate periodic respiration or respiratory failure and should generally be avoided if at all possible. Tracheal intubation and mechanical ventilation may be used to maintain ventilation, or to reduce intracranial pressure (section 12.7.4). Indications vary between centres but generally include: deteriorating conscious level, severe hypoxia or hypercapnoea, and inability to maintain an airway. Depending on the indications and case mix, reported frequency of in-hospital case fatality among patients undergoing mechanical ventilation varies between 50 and 90%. Older age, evidence of brainstem dysfunction and comorbidities are predictably associated with worse outcomes; however, a small proportion of survivors make a reasonable recovery.18–20 Often, such aggressive management is considered inappropriate because of the patient’s low probability of regaining a good quality of life (section 10.4). If patients are nursed in a sitting position, it is best if they are well supported and upright in a chair rather than slumped in bed or a chair, which makes breathing more difficult.

11.2.3 Poor circulation Both severe brainstem dysfunction and massive subarachnoid haemorrhage can cause major circulatory problems such as neurogenic pulmonary oedema, cardiac dysrhythmias and erratic blood pressure with severe hypertension or hypotension (section 14.9.3 and 14.9.4). More frequently, circulatory failure, with or without hypotension or hypoxia related to pulmonary oedema, is due to

541

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 542

542

Chapter 11 What are this patient’s problems?

coexistent heart disease (e.g. congestive cardiac failure, myocardial infarction, atrial fibrillation), hypovolaemia (e.g. dehydration, bleeding) or severe infection. Cardiac dysrhythmias Cardiac dysrhythmias are quite common after an acute stroke, although they rarely cause major problems.21 The most clinically important, in terms of frequency and effect on management, is atrial fibrillation which occurs in about one-fifth of patients with ischaemic stroke and one in ten of those with intracerebral haemorrhage.22,23 In most cases, atrial fibrillation precedes the stroke and in some it is the most likely cause of the stroke (section 6.5.1). Of course, a small proportion of ischaemic strokes, perhaps 5%, occur in the context of a recent (within 6 weeks) myocardial infarction (section 7.10).24 Routine monitoring of the patient’s cardiac rhythm, either at the bedside or preferably with ambulatory systems which do not inhibit early mobilization, will identify abnormalities, most commonly paroxysmal atrial fibrillation. The yield will depend on the duration of monitoring. A 24-h ECG monitor might detect previously undected atrial fibrillation in up to 5% and 7-day monitoring in a further 5% of hospitalized patients with ischaemic stroke.25,26 The detection of atrial fibrillation will influence decisions about anticoagulation for secondary prevention (section 16.6). The feasibility and cost of routinely monitoring rhythm inevitably varies between centres and it is unclear what the yield is in terms of altering management, perhaps by starting anticoagulation and preventing recurrences. Certainly, as a minimum we would advise monitoring those with palpitations, syncope, unexplained breathlessness or recent myocardial infarction. Cardiac dysrhythmias, other than atrial fibrillation, are quite common after an acute stroke, but seldom seem to be a problem unless they are due not to the stroke itself but to a recent myocardial infarction.

Myocardial injury It is important to identify the presence and likely cause of any circulatory failure, by clinical assessment of the cardiovascular system backed up with relevant investigations (e.g. cardiac enzymes, electrocardiogram (ECG) and echocardiography). Repolarization abnormalities and ischaemic-like changes on the ECG occur in about three-quarters of patients with subarachnoid haemorrhage but are far less common with other types of stroke.

In SAH the changes are most often secondary to the cerebral insult while in other stroke types they usually reflect associated cardiac disease.27 The specificity of ECG changes for acute myocardial infarction in the context of an acute stroke is low.28 Levels of plasma creatine kinase (CK) and of its more cardiac-specific isoenzyme CK muscle–brain (MB) are quite often raised but the interpretation may be difficult where the patient has fallen, been injured, lain on a hard floor or had a generalized seizure which may all increase the CK level. Indeed even when CK-MB is raised, troponins, which are more specific markers of myocardial injury, are often normal indicating that even CK-MB may not be of cardiac origin.29 Raised troponins have been reported in up to a third of patients with stroke and may be markers of a concurrent myocardial infarction or small areas of myocardial necrosis, which have been identified at autopsy.30 Raised troponins are associated with poorer survival even having adjusted for age and stroke severity.27,31,32 Patients with circulatory failure, atrial fibrillation and recent myocardial infarction have a poor prognosis for survival and functional recovery after stroke.22,33 Active treatment of these problems is likely to improve the outcome but a detailed account is beyond the scope of this book. Gastrointestinal haemorrhage About 3% of stroke patients have an upper gastrointestinal bleed in the first few weeks.34 Bleeding most commonly occurs from petechiae, erosions and ulcers.35 Bleeds are more common in the elderly, in more severe strokes and those with intracerebral haemorrhage; not surprisingly, therefore, gastrointestinal haemorrhage has been associated with high case fatality. Bleeding appears to be more common in patients fed via an enteral tube and may be more common with nasogastric, rather than percutaneous endoscopic, gastrostomy.36 Clearly, resulting hypotension and anaemia might both exacerbate cerebral ischaemia and worsen the neurological outcome. One should obviously be aware of this bleeding potential when prescribing antithrombotic and antiinflammatory medication.37 We try to avoid using nonsteroidal anti-inflammatory drugs (NSAIDs) unless there is a strong indication. Proton pump inhibitors (PPI), double dose H2 receptor antagonists and misoprostol have all been shown to reduce gastric and duodenal ulcers in patients (not stroke) taking NSAIDs chronically.38 A meta-analysis of 20 trials with a total of 2624 mainly Chinese patients suggested an 80% odds reduction in stress-related gastrointestinal haemorrhage among stroke patients and even a reduction in mortality associated with PPI and H2 receptor antagonists.39 However, the rate of bleeding in the control groups of

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 543

11.3 Reduced level of consciousness Table 11.2 The frequency of various clinical features among 675 patients with a first-ever-in-a-lifetime stroke at their initial assessment by a study neurologist (from the Oxfordshire Community Stroke Project, unpublished data). Clinical feature

Number (%)

Not assessable (%)

Glasgow Coma Scale (motor < 6) Glasgow Coma Scale (eyes < 4) Glasgow Coma Scale (eyes + motor < 10) High blood pressure (systolic > 160 mmHg) Very high blood pressure (systolic > 200 mmHg) High blood pressure (diastolic > 100 mmHg) Very high blood pressure (diastolic > 120 mmHg) Epileptic seizures within 24 h Facial weakness Arm or hand weakness Leg weakness Unilateral weakness of at least two of face, arm and leg Sensory loss (proprioception) Sensory loss (spinothalamic) hom*onymous visual field defect Gaze palsy Mental test score < 8/10433 Visuospatial dysfunction Dysphasia Dysarthria

86 (13) 98 (15) 111 (16) 311 (46) 90 (13) 131 (19) 23 (3) 14 (2) 256 (38) 344 (51) 307 (45) 331 (49) 101 (15) 196 (29) 113 (18) 50 (7) 85 (13) 81 (12) 122 (19) 135 (20)

15 (2) 16 (2) 16 (2) 19 (3) 19 (3) 22 (3) 22 (3) 0 (0) 40 (6) 32 (5) 34 (5) 0 (0) 168 (25) 140 (21) 134 (20) 36 (5) 135 (20) 179 (27) 111 (16) 127 (19)

these trials was far higher than that reported in observational studies in white populations. Management of active gastrointestinal haemorrhage is the same as in other situations.

been argued that differences in pressure between brain compartments may be more important than the overall pressure. However, treatable complications of the stroke may also depress conscious level (Table 11.3). Level of consciousness is an important indicator of the severity of the stroke and a valuable prognostic variable44 (section 10.2.7).

11.3 Reduced level of consciousness

Table 11.3 Remediable causes of a reduced level of consciousness after stroke.

About 15% of stroke patients have a reduced level of consciousness during the first few days after the stroke (Table 11.2). The causes are not entirely clear (Table 11.3). It may arise from direct damage (haemorrhage or infarction) to the brainstem or more commonly indirect damage from supratentorial lesions associated with brain swelling and midline shift.40,41 However, we not infrequently see patients with large ischaemic, apparently unilateral, hemispheric strokes who become drowsy within hours of onset, before one would expect enough oedema to have developed to raise intracranial pressure markedly, and with little or no apparent mass effect on brain imaging.42 Indeed, in one study, only a minority of patients whose conscious level deteriorated after a large hemispheric ischaemic stroke had globally raised intracranial pressure on invasive monitoring.43 It has

Causes General Hypoxia Hypotension Severe infection Electrolyte imbalance Hypoglycaemia Drugs: e.g. benzodiazepines, opiate analgesics Neurological Epileptic seizures Raised intracranial pressure Obstructive hydrocephalus (e.g. due to cerebellar haematoma) Cerebral ischaemia after subarachnoid haemorrhage

Section

11.2.2 11.2.3 11.12 11.18.1 and 11.18.2 11.18.4

11.8 12.7, 13.3.2 13.3.6, 14.8 14.5, 14.7

543

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 544

544

Chapter 11 What are this patient’s problems?

Assessment The most commonly used measure of level of consciousness is the Glasgow Coma Scale (GCS) (Table 3.12), which reliably documents the patient’s spontaneous actions as well as those in response to verbal and painful stimuli.45 Although useful, the GCS, which was originally designed for patients with head injury, has a number of problems when applied to stroke. The most important is that many patients with stroke are aphasic and therefore have a reduced score on the verbal response scale. Thus, a patient may have a reduced total score on the GCS but a normal level of consciousness. Another problem is that inexperienced users may apply the motor scale to the affected arm in a patient with a hemiparesis and obtain a reduced motor response rather than the ‘best’ response. The GCS communicates most information if the component scores are recorded separately. Alternative measures include the reaction level scale and components of so-called ‘stroke scales’ (sections 10.2.7 and 17.12.5).46 Patients with a reduced level of consciousness are at greater risk of developing various complications (Table 11.4) so expert nursing care is needed to minimize these risks. Also, a falling level of consciousness may alert one to an important and potentially reversible condition (Table 11.3). Therefore, patients with severe strokes, in whom active treatment would be considered, should be monitored with the GCS for the first few days, especially where staff change frequently, to avoid delays in initiating investigation and treatment of any worsening in their condition. We usually reduce the frequency of monitoring, e.g. hourly, 2, 3 or 6-hourly, over the first few days after the stroke. The speed at which we do this will depend on the team’s judgement about the likelihood of deterioration and the appropriateness and urgency of any intervention aimed at reversing the deterioration. Also, the availability of nursing staff to perform the measurements and the patient’s need for Table 11.4 Common complications in patients with a depressed level of consciousness after stroke. Complication

Section

Airway obstruction Aspiration Fever and infection Dehydration Malnutrition Urinary incontinence Faecal incontinence or constipation Pressure ulcers

11.2.1 11.17 11.12 11.18.1 11.19 11.14 11.15 11.16

unbroken sleep have to be taken into account. Therefore, one might monitor a patient’s GCS hourly during the first 2 days after a cerebellar haematoma (in whom decompressive surgery would be considered) but only 6-hourly in a patient who is stable 2 days after a mild ischaemic stroke. The Glasgow Coma Scale is a useful tool for monitoring patients only if it is applied properly. Staff need to be trained in its proper use. Our patients are frequently said by medical and nursing staff to be confused or aphasic with a GCS of 15/15. Clearly this is nonsense! The investigation and treatment of patients with a falling level of consciousness are similar to those with other indicators of worsening and are discussed in section 11.5. The specific treatment of patients with raised intracranial pressure is discussed in sections 12.7 and 13.3.2.

11.4 Severe stroke vs apparently severe stroke

It is important to be aware that severe concurrent illness, or stroke complications, may make a stroke appear much worse than it really is (Table 11.5). The clinical features which might indicate a severe stroke (e.g. reduced level of consciousness), and so a poor prognosis, may in fact be due to the complicating illness rather than the stroke itself. For example, a patient with a pure motor stroke (lacunar syndrome) (section 4.3.2) and a severe chest infection may be drowsy or confused. It is then difficult to distinguish this clinical picture from that due to a major middle cerebral artery territory infarction (section 4.3.4). Of course, an infection is more easily treated than a large volume of necrotic brain, so with appropriate therapy the prognosis of the two types of patient will be quite different. A thorough general examination will identify signs such as fever, confusion, increased respiratory rate and reduced oxygen saturation, and usually reveals any relevant coexisting disorder. Simple investigations such as a white blood cell count, C-reactive protein, urea and electrolytes, urine microscopy and culture, chest X-ray, ECG and blood cultures are useful, not only to identify the cause of the stroke (sections 6.8.1 and 6.8.2), but also to alert one to serious coexisting non-stroke disease. Seizures that have occurred since stroke onset make the

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 545

11.5 Worsening after a stroke Table 11.5 Causes of apparently severe strokes. Section Non-neurological Infection Respiratory Urinary Septicaemia Metabolic Dehydration Electrolyte disturbance Hypoglycaemia Drugs Major and minor tranquillizers Baclofen Lithium toxicity Anti-epileptic drug toxicity Anti-emetics Hypoxia Pneumonia/chest infection Pulmonary embolism Chronic pulmonary disease Pulmonary oedema Hypercapnoea Chronic pulmonary disease Limb or bowel ischaemia in patients with a cardiac or aortic arch source of embolism Neurological Obstructive hydrocephalus in patients with stroke in the posterior fossa, or subarachnoid haemorrhage Epileptic seizures, including complex partial seizures

11.12.1 11.12.2

11.18.1 11.18.2, 11.18.3 11.18.4 11.29 11.20 11.8 11.9 11.12.1 11.13 11.6 11.2.3 11.2.2

13.3.6, 12.7.3 and 14.8 11.8

assessment of stroke severity particularly difficult (section 11.8). It is also important not to overlook the possibility that any decreased conscious level is due to sedative drugs. The treatment of apparently severe strokes will depend on the specific problem identified or suspected. A 70-year-old man was found unconscious at home and admitted to hospital. In the receiving unit he had a partial seizure affecting his right side with secondary generalization. Subsequent neurological examination revealed a reduced conscious level, deviated gaze and a flaccid paralysis of the right side. Brain CT scan on the day of admission was normal. A diagnosis of a severe ischaemic stroke was made and the decision not to resuscitate was recorded in the notes. He was given intravenous fluids but no anti-epileptic drugs. He was later reviewed by the stroke physician who, having found occasional twitching movements of the right hand,

reversed the decision not to resuscitate and arranged an EEG which showed status epilepticus. Following intravenous phenytoin the patient improved and 4 days later was discharged from hospital without residual neurological deficits. Therefore, do not ‘write off’ people with apparently severe strokes and recent seizures.

11.5 Worsening after a stroke

Although stroke onset is usually abrupt, the patient’s neurological condition may worsen hours, days or, rarely, weeks after the initial assessment (as well as improve). Patients may have a reducing level of consciousness, worsening of existing neurological deficits, or new deficits indicating dysfunction in another part of the brain. A large number of terms including ‘strokein-evolution’, ‘stroke-in-progression’, ‘early neurological deterioration’ and ‘progressing stroke’ have in the past been applied to this situation, which probably reflects the considerable level of interest in the problem and uncertainty as to its causes.47 After all, here is a situation where one might be able to intervene to prevent a major stroke. One well-validated definition of ‘progressing stroke’ is based on a deterioration in the Scandinavian Stroke Scale48 (Table 11.6) of at least two points on the conscious level, arm, leg or eye movement subscales and/or at least three points on the speech subscale between baseline and day 3.49,50 The same change between consecutive measurements within the first 3 days was referred to as an ‘early deterioration episode’. Clearly, if we accept that the neurological deficit commonly increases over minutes or hours, then the earlier in the course of the stroke we first see the patient, the more likely we are to observe subsequent worsening. Indeed, it is likely that, as we attempt to introduce acute treatments for stroke, we will become much more aware of the ‘normal’ worsening over the first few hours after stroke onset. We will probably see patients earlier in the development of their stroke and monitor them more closely because they will be in clinical trials, or given a specific treatment for acute stroke. Not surprisingly therefore, estimates of the frequency of worsening in the first day or two have varied considerably but have been reported in up to 40% of hospital-admitted patients.51 Early worsening is more likely to be due to a neurological mechanism than a systemic complication of the stroke.52 The factors underlying the progression of neurological deficits in the first day or two after stroke

545

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 546

546

Chapter 11 What are this patient’s problems? Table 11.6 Scandinavian Stroke Scale.48 Consciousness: – fully conscious – somnolent, can be awaked to full consciousness – reacts to verbal command, but is not fully conscious – no reaction to verbal command Eye movement: – no gaze palsy – gaze palsy present – conjugate eye deviation Arm, motor power:* – raises arm with normal strength – raises arm with reduced strength – raises arm with flexion in elbow – can move, but not against gravity – paralysis Hand, motor power *: – normal strength – reduced strength in full range – some movement, fingertips do not reach palm – paralysis Leg, motor power *: – normal strength – raises straight leg with reduced strength – raises leg with flexion of knee – can move, but not against gravity – paralysis Orientation: – correct for time, place and person – two of these – one of these – completely disorientated Speech: – no aphasia – limited vocabulary or incoherent speech – more than yes/no, but not longer sentences – only yes/no or less Facial palsy: – none/dubious – present Gait: – walks 5 m without aids – walks with aids – walks with help of another person – sits without support – bedridden/wheelchair TOTAL

6 4 2 0 4 2 0 6 5 4 2 0

6 4 2 0

6 5 4 2 0

onset are unclear.53 For example, what causes the progression of deficits in some patients with lacunar infarction?54 Early worsening is likely to reflect complex interactions between biochemical and haemodynamic factors which are known to be important in the development of ischaemic stroke (sections 12.1.3 and 12.1.4). A history of diabetes or coronary heart disease, low and high arterial blood pressure, early computed tomography (CT) signs of infarction, evidence of siphon or middle cerebral artery occlusion, various biochemical (e.g. glucose and glutamate) and haematological (e.g. D-dimers) parameters have all been associated with a greater risk of early worsening.47,55–57 Greater age, initial stroke severity and the presence of cerebral oedema on scanning have been associated with late deterioration.47 Worsening has a number of potential causes, some reversible, so it is important to detect and treat them early. The literature has tended to emphasize the neurological causes (Table 11.7) but it is important not to miss the non-neurological ones which tend to occur after the first couple of days and which are more treatable. There is, not surprisingly, considerable overlap with the causes of ‘apparently severe stroke’ (Table 11.5). The outcome of patients whose neurological condition worsens after initial presentation is predictably worse than that of patients who remain stable or improve rapidly.

Table 11.7 Causes of worsening after stroke.

6 4 2 0

Section

10 6 3 0 2 0

12 9 6 3 0

*Motor power is assessed only on the affected side.

Neurological Progression/completion of the stroke Extension/early recurrence Haemorrhagic transformation of an infarct Development of oedema around the infarct or haemorrhage* Obstructive hydrocephalus in patients with stroke in the posterior fossa, or after subarachnoid haemorrhage* Epileptic seizures* Delayed ischaemia* (in subarachnoid haemorrhage) Incorrect diagnosis Intracranial tumour* Cerebral abscess* Encephalitis Chronic subdural haematoma* Subdural empyema* Non-neurological* See Table 11.5

16.2 5.7 12.1.5 13.2.6, 12.7.3 and 14.8 11.8 14.5, 14.7 3.4

*Remediable causes of worsening.

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 547

11.5 Worsening after a stroke Table 11.8 Parameters to monitor and which can detect worsening. Section Conscious level, i.e. Glasgow Coma Scale Pupillary responses Eye movements Limb movements Stroke scale (e.g. Scandinavian Stroke Scale) Temperature Pulse rate Blood pressure Respiratory rate Pulse oximetry Fluid balance

3.3.2, 11.3 3.3.6 3.3.6 3.3.4 Table 11.6 11.12 11.2.3 11.7 11.2.2, 11.12 11.2.2 11.18.1

Assessment To ensure that clinical worsening is identified as early as possible, i.e. when the potential for therapeutic reversal is usually greatest, it is important to monitor the patient’s condition. Regular measurements such as those in Table 11.8 will usually alert one to any problem. However, experienced nursing staff who have regular close contact with the patient may detect a problem at an early stage before it becomes obvious to other members of the team. Family members, who often spend long periods with ill patients, may also detect subtle but important changes in the patient’s condition and they should be listened to. It is important that the physician caring for the patient encourages free communication so this sort of information is made known to those directing the patient’s care. Although there is a long list of potential causes of worsening, it is important to consider first those that are most readily reversible (Table 11.7). Most can be diagnosed by a clinical assessment supplemented by simple laboratory investigations. An urgent repeat CT brain scan is advisable in some situations, for example deterioration in conscious level in a patient with a cerebellar stroke or subarachnoid haemorrhage who may be developing obstructive hydrocephalus amenable to neurosurgical intervention. In a patient treated with thrombolytic drugs, an abrupt change in blood pressure, new onset of headache and/or vomiting or neurological deterioration should prompt an urgent CT scan to detect intracranial bleeding due to the treatment. The detection of haemorrhagic transformation of infarction in a patient receiving antithrombotic drugs may influence the decision to continue the medication or not, although there is no reliable evidence for the best policy

in this situation (sections 5.7, 12.4.2 and 12.5.7). A CT or MRI brain scan that shows that the deterioration was due to an early recurrent ischaemic stroke, especially one in a different arterial territory to the initial stroke, may encourage one to investigate further for a proximal arterial or cardiac source of embolism. Treatment Clearly, any treatment will depend on the reason for the worsening (see relevant sections) but it should be emphasized that at present there is little evidence from RCTs to support the use of treatments such as anticoagulants, thrombolysis, neuroprotection, haemodilution and manipulation of blood pressure for the treatment of worsening (Chapter 12). A small trial which randomized 98 patients to either routine care or intensive physiological monitoring showed that the more intensively monitored patients received more intensive treatment (e.g. supplementary oxygen) and deteriorated less frequently.58 This preliminary finding, which needs to be confirmed in larger studies, could have important implications for the management of patients with acute stroke (section 17.6.2). A 65-year-old woman had an acute ischaemic stroke affecting her cerebellum and brainstem. This occurred on a background of a previous occipital ischaemic stroke, impaired renal function and hypertension. After a stormy early course requiring drainage of obstructive hydrocephalus and ventilation on the intensive care unit, she made good progress. She could sit independently, help with washing herself and dressing, and could take a soft diet and fluids safely. She had diplopia, poor balance and complained of vertigo and vomiting which was exacerbated by movement. She was started on regular oral metoclopramide with some relief of these symptoms. She was transferred to a separate rehabilitation ward on a Friday afternoon. Over the weekend she deteriorated. Her speech became very unclear, her swallowing unsafe and she was unable to sit. Infection, metabolic abnormalities and recurrent hydrocephalus were excluded and a repeat MR scan showed no new stroke lesion although this seemed the most likely explanation for her deterioration. The family were advised that the prognosis was poor. The speech therapist, assessing the patient prior to insertion of a percutaneous endoscopic gastrostomy, remarked that the patient’s tongue movements were like those of a patient with Parkinson’s disease – the penny dropped! The drug chart was reviewed, the metoclopramide was withdrawn and the patient returned to her previous functional state over the next week. She was eventually discharged home, walking and requiring minimal help with everyday activities.

547

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 548

548

Chapter 11 What are this patient’s problems? Table 11.9 Frequency of coexisting medical problems among 675 patients with a first-ever-in-a-lifetime stroke (from the Oxfordshire Community Stroke Project, unpublished data).

Previous angina Previous myocardial infarction Cardiac failure Intermittent claudication Diabetes mellitus Previous epileptic seizures Previous malignancy Dependent before stroke (Rankin > 2)

(%)

106 112 52 112 63 19 74 103

(16) (17) (8) (17) (9) (3) (11) (15)

No data were available on respiratory or musculoskeletal problems.

which contribute to disability, progress. Thus, a patient may reach their optimal functional recovery some months after a stroke and then deteriorate due to progression of a comorbid condition. If this kind of deterioration can be anticipated, it may allow for a more flexible package of care to cope with such fluctuations. There can be few things more dispiriting than to strive to discharge a patient into one form of accommodation and then hear that within a few months their condition has deteriorated to such an extent that the accommodation is no longer suitable. This can shatter the morale of the patient and their carer. Functional deterioration months after a stroke is unlikely to be due to the initial stroke and much more likely to be caused by a recurrent stroke or the progression of some comorbid condition such as angina, arthritis or intermittent claudication.

11.6 Coexisting medical problems Assessment Coexisting medical problems are common in stroke patients because they are usually elderly and have associated vascular disease (Table 11.9). Although we have already mentioned the importance of cardiorespiratory diseases in the immediate management of the patient (section 11.2), these and other conditions can be important for many other reasons, not least that they may require treatment in their own right. Severe non-stroke illness can make a mild stroke appear severe and thus lead to an inaccurate prognosis and possibly inappropriate treatment (section 11.4). Coexistent cardiorespiratory (e.g. angina, cardiac failure, chronic obstructive pulmonary disease), musculoskeletal (e.g. arthritis, back pain, amputation) and psychiatric conditions (e.g. depression, anxiety) often compound stroke-related disability. So, for example, after several months of rehabilitation, one might have taught a patient with a severe hemiparesis to walk again, but if he or she also has chronic obstructive pulmonary disease the added effort of walking with a hemiplegic gait may mean that the patient cannot walk any useful distance. It is important to be aware of the limitations on rehabilitation imposed by any pre-existing disease before spending months trying to make a patient walk. It may be more realistic to teach the patient to be independent in a wheelchair. Even if one cannot estimate the impact of non-stroke disease on recovery, knowledge of its existence may explain why patients are not achieving their rehabilitation goals (section 10.3.4). Although most patients who survive a stroke improve over weeks or months, many of the coexisting problems,

A thorough history and examination at the time of admission, with reassessment when the patient is more active, should identify the main coexisting medical problems – the patient’s medication will often provide a clue. An assessment of the patient’s pre-stroke functional status (i.e. what could they do and not do?) is invaluable, not only in predicting outcome (section 10.2.7) but also for identifying comorbidities. Unfortunately, this is often not recorded in the medical records unless specifically prompted.59,60 One approach might be to estimate routinely a pre-stroke Barthel Index (section 17.12.5) since this covers most of the important activities of daily living (ADL). This sort of ADL checklist is also useful in making a prognosis and in setting rehabilitation goals since, depending on the cause and duration, the pre-stroke functional impairment will determine the best achievable post-stroke functional status. If a patient has, as a result of arthritis, been immobile for 10 years before the stroke, it is ridiculous to try to get the patient to walk after the stroke. Unfortunately, this is often attempted simply because nobody has obtained an accurate picture of the patient’s pre-stroke function. This is more likely to happen where the patient has difficulties with communication and no carer is available. An assessment of the patient’s pre-stroke function may also be very important in making decisions even within the first few hours after stroke. Although it may be impossible for anyone to judge a person’s quality of life, other than that person themselves, one may deduce something from the patient’s function. This may be important where, for example, one is considering antibiotics

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 549

11.7 High and low blood pressure after stroke

Clearly, one should aim to minimize the effect of any comorbidity by giving as effective treatments as possible. Where it is not possible to influence the disease directly, it is still important to take account of comorbidity in one’s overall approach to the patient. One’s intermediate and long-term goals also have to take this into account (section 10.3.3).

11.7 High and low blood pressure after stroke

Dead within 14 days (%)

Treatment

68 66

64 20

60 58

56 54

5 160/90 mmHg) before the onset. They will tend to have higher blood pressures than those without previous hypertension and are more likely to have evidence of end-organ damage, e.g. hypertensive retinopathy, impaired renal function and left-ventricular hypertrophy. High blood pressure in the acute phase of both haemorrhagic and ischaemic strokes is associated with a poor outcome, an effect which appears to be independent of age and stroke severity (Fig. 11.4).61,63–68 Assessment There is considerable variation in the methods used and the frequency of monitoring blood pressure after acute stroke. Traditionally, the blood pressure has been measured in the standard way with a sphygmomanometer and an appropriately sized cuff kept at the level of the

patient’s heart. However, increasingly, semi-automated non-invasive systems are used which allow more frequent monitoring (even when there are few nursing staff) and thus earlier intervention. These systems must be calibrated to avoid misleading readings.69 The presence of atrial fibrillation reduces the reproducibility of both manual and semi-automated blood pressure recordings.70,71 In the intensive care unit, intra-arterial monitoring is frequently used although this can very occasionally cause peripheral ischaemia. Perhaps not surprisingly, more intensive monitoring often leads to more manipulation of blood pressure, the value of which is currently unclear (see below). Semi-automatic blood pressure monitors should be properly calibrated to avoid misleading results. In patients with very low or high blood pressure, where the reading is likely to lead to a significant change in management, we would normally check the reading with a traditional sphygmomanometer. There is no consistent difference between the blood pressure measured in the weak vs the unaffected arm, although there is often a difference between arms which is unrelated to the side of the stroke and probably reflects occlusive arterial disease affecting one arm.72 Therefore, ideally the blood pressure should be checked in both arms on at least one occasion, and if the blood

549

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 550

550

Chapter 11 What are this patient’s problems?

pressures are different, it should be monitored consistently in the arm giving the highest reading to avoid a spurious label of labile hypertension. If the blood pressure is raised on admission it should be monitored to establish whether or not it falls spontaneously and the patient examined to identify evidence of end-organ damage. Treatment There is considerable uncertainty about the risks and benefits of lowering the blood pressure in the acute phase of stroke. Treatment may theoretically reduce the likelihood of rebleeding in intracerebral and subarachnoid haemorrhage, and of brain oedema and haemorrhagic transformation in cerebral infarction. However, lowering the blood pressure may reduce cerebral perfusion where cerebral autoregulation is impaired and thus further increase ischaemic damage73 (section 12.1.2). Intravenous calcium channel blockers which, apart from their potential neuroprotective action, also lower arterial blood pressure have been associated with worse outcomes in several randomized controlled trials (RCTs)74 (section 12.8.3). Small trials have not established how blood pressure should be manipulated in the acute phase of stroke and larger trials are ongoing to determine how and when we should do so61 (http://www.strokecenter.org/ trials/). Until RCTs are available, we offer the following advice: • If the blood pressure remains raised (i.e. > 140/90 mmHg) for more than 1 week, or where there is evidence of end-organ damage, we would give the patient general advice (i.e. salt restriction, weight loss and moderation of alcohol intake) and start an antihypertensive drug, accepting that this timing is arbitrary. The issues of when to start antihypertensive treatment, the blood pressure level and the choice of agent, for the purposes of secondary prevention, are discussed in section 16.3. • Our management is similar whatever the pathological type of stroke although we would be less inclined to lower blood pressure if the patient has severe stenoses or occlusions of the carotid and/or vertebral arteries. • Where the patient is already on antihypertensive treatment, pending the results of ongoing RCTs, it seems reasonable to continue it as long as the patient can swallow the tablets safely and has not become hypovolaemic, which may increase the drug effect and lead to marked and potentially damaging hypotension. • If the blood pressure is very high (i.e. > 220/ > 120 mmHg) there may be evidence of organ damage which would prompt earlier initiation of bloodpressure-lowering drugs (Table 11.10).

Table 11.10 Circ*mstances in which we would consider lowering the blood pressure immediately after an acute stroke. Papilloedema or retinal haemorrhages and exudates Marked renal failure with microscopic haematuria and proteinuria Left ventricular failure diagnosed on clinical features and supported by evidence from the chest X-ray and/or echocardiogram Features of hypertensive encephalopathy, e.g. seizures, reduced conscious level Aortic dissection Note: even these features may be misleading in acute stroke because left heart failure may frequently be due to coexistent ischaemic heart disease, and seizures and drowsiness may be due to the stroke itself.

• The aim of therapy should be a moderate reduction in blood pressure over a day or so, not minutes. An angiotensin-converting enzyme inhibitor (in the absence of contraindications) is a reasonable first choice. Thiazides probably have little effect in the acute phase although they remain a very reasonable choice later.75,76 Transcutaneous glyceryl trinitrate is effective in lowering blood pressure, and the patches can easily be applied in dysphagic patients and withdrawn relatively quickly if blood pressure drops too far.77 Intravenous labetalol, with very careful intraarterial monitoring of blood pressure, may be useful in resistant cases. Do not lower the blood pressure in the first few days after a stroke unless there is evidence of accelerated hypertension or end-organ damage, or the patient is in of one of the ongoing multicentre randomized controlled trials (http://www.strokecenter.org/trials/).

11.7.2 Low blood pressure It is difficult to define low blood pressure since, although one could set an arbitrary lower value for the systolic or mean blood pressure, clinically significant hypotension is that which leads to dysfunction of one or more organs. The level at which this occurs will depend on the patient’s age, their usual blood pressure, the state of their arterial tree and whether cerebral autoregulation is intact or impaired. There is good evidence that autoregulation in the brain is impaired after stroke so that even if a patient has the same blood pressure after stroke as before, cerebral perfusion might still be reduced.

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 551

11.8 Epileptic seizures

Assessment The monitoring of blood pressure after stroke has been discussed in section 11.7.1. The assessment of the clinical importance of hypotension should comprise a clinical examination and some simple investigations (e.g. blood urea, arterial blood gases) to identify the features of ‘shock’ mentioned already. Having established that the patient has low blood pressure which is associated with under-perfusion of tissues, it is then important to establish the cause. Is the patient hypovolaemic, due to dehydration (section 11.18.1) or blood loss (section 11.2.3)? Has the patient had a pulmonary embolus (section 11.13), or are they in heart failure (section 11.2.3) or septic (section 11.12)? Is the patient on drugs which could lower blood pressure excessively? These questions can normally be answered following a thorough clinical examination including assessment of the jugular venous pressure, review of the drug and fluid balance charts, and some simple investigations including haemoglobin and haematocrit, neutrophil count, C-reactive protein, urine and blood cultures, cardiac enzymes, an ECG and chest X-ray. Occasionally, further investigation with, for example, a CT pulmonary angiogram, an echocardiogram or measurement of right atrial or pulmonary wedge pressures is required to sort out the cause.

blood pressure in acute stroke.78 Trials of sympathomimetics to raise blood pressure are in progress (http://www.strokecenter.org/trials/).

11.8 Epileptic seizures

11.8.1 Early seizures About 5% of patients have an epileptic seizure within the first week or two of their stroke (so-called onset seizures), most occurring within 24 h. Inevitably, estimates of the frequency of onset seizures vary because of differences in case selection, diagnostic criteria, lack of witnesses and methods of follow-up. Most onset seizures are partial (focal) although often with secondary generalization (section 3.3.10) (Table 11.2). Onset seizures are more common in severe strokes, in haemorrhagic stroke and stroke involving the cerebral cortex.79,80

11.8.2 Later seizures In population-based cohorts, which are relatively unaffected by hospital referral bias, the risk of having a first seizure, excluding onset seizures, is between 3% and 5% in the first year after a stroke and about 1–2% per year thereafter (Fig. 11.5). About 3% will have more than one seizure and could be regarded as having developed

100

This will obviously depend on the cause of the low blood pressure. In our experience, hypovolaemia is the most frequent problem and patients usually improve with intravenous fluids. Obviously, it is important to exclude cardiac failure before giving fluids in this way. There is little evidence on which to base a decision to increase the

Intracerebral haemorrhage

80 70

Subarachnoid haemorrhage

60 50 40 30 20 10 0

Treatment

Ischaemic stroke

90 % still seizure-free

Unfortunately, judgements about the optimal level of blood pressure after acute stroke are very difficult in routine practice since we have no easy and reliable techniques for assessing organ perfusion and more importantly function, other than the clinical state of the patient. Of course, when blood pressure is very low, patients may show signs of ‘shock’ (e.g. cold extremities, low urine output, worsening renal function, mental confusion, lactic acidosis) and actions to improve organ perfusion are required. In the International Stroke Trial, low blood pressure after acute stroke was associated with poor outcome, even after adjusting for stroke severity (Fig. 11.4).66 However, the low blood pressure might not be the cause of the poor outcome but rather the consequence of important comorbidity (e.g. heart failure, atrial fibrillation) or complications (e.g. dehydration, pulmonary embolism).

Time (years) Fig. 11.5 A Kaplan–Meier plot showing the proportion of patients remaining seizure-free at increasing intervals after a first-ever-in-a-lifetime stroke in the Oxfordshire Community Stroke Project. Separate plots are shown for patients with ischaemic stroke (n = 545), non-truamatic intracerebral haemorrhage (n = 66) and subarachnoid haemorrhage (n = 33). Adapted from Burn et al., 1997.80

551

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 552

552

Chapter 11 What are this patient’s problems?

epilepsy following a stroke.79 This represents a greatly increased relative risk of seizure (perhaps 20-fold) compared with stroke-free individuals of similar age. Patients with onset seizures, haemorrhagic strokes and infarcts involving the cerebral cortex have the highest overall risk of later seizures. Seizures may recur in about 50% of the patients, but are rarely troublesome if accurately diagnosed and appropriately treated. Patients who become functionally independent and who have not yet had a seizure are at very low risk of post-stroke seizures. The theoretical future risk of seizures is not great enough to prevent the patient driving an ordinary car in the UK where the acceptable risk of a seizure is < 20% per annum. The risk of having an epileptic seizure after a firstever-in-a-lifetime stroke is, on average, about 5% in the first year and 1–2% per year thereafter. However, the risk is higher in patients with haemorrhagic stroke or with large ischaemic strokes involving the cortex, and lower in patients with lacunar and posterior circulation strokes.

Assessment The diagnosis of seizures should, as usual, be based on a detailed description of the attack from the patient, and if possible a witness, and may very occasionally be confirmed by electroencephalography (EEG) during a seizure. Video-EEG monitoring is occasionally useful where patients are having frequent attacks of uncertain nature. If patients have seizures in the first few days after the stroke, and especially if they are partial (focal), they should be investigated. Non-stroke brain lesions complicated by post-seizure impairments (e.g. Todd’s paresis) may mimic strokes (section 3.4.2). Also, because the neurological deficits and conscious level may be temporarily much worse immediately after a seizure, ‘onset seizures’ can make the assessment of stroke severity unreliable (section 11.4). We have also seen occasional patients who have had non-convulsive status epilepticus, a diagnosis which can only be confirmed by EEG, who may, for example, be severely aphasic, and who have improved dramatically with anti-epileptic treatment. Furthermore, seizures should not automatically be attributed to the stroke since many other causes may be present coincidentally, or as a consequence of the stroke (Table 11.11); appropriate investigations should be performed to exclude these. Finally, seizures may mimic recurrent stroke if associated with worsening of the original focal deficit; this is particularly confusing if any seizures have been unwitnessed and the patient presents with worsening of their earlier stroke (section 11.5).

Table 11.11 Causes of epileptic seizures after ‘stroke’. Section General Alcohol withdrawal Anti-epileptic drug withdrawal Hypoglycaemia Hyperglycaemia, especially non-ketotic hyperglycaemia Hyper/hyponatraemia Hypocalcaemia or hypomagnesaemia Drugs: Baclofen given for spasticity Antibiotics for infections, e.g. ciprofloxacin Antidepressants given for emotionalism or depression Phenothiazines given for agitation or hiccups Anti-arrhythmics for associated atrial fibrillation Neurological Due to the primary stroke lesion Haemorrhagic transformation of infarction Underlying pathology: Arteriovenous malformation Intracranial venous thrombosis Mitochondrial cytopathy Hypertensive encephalopathy Wrong diagnosis Herpes simplex encephalitis Cerebral abscess Intracranial tumour Subdural empyema

11.18.4 11.18.3 11.18.1 & 11.18.2

11.20 11.12

11.31 11.10 & 11.29.1 11.2.3

5.7

8.2.4 7.21 7.19 3.4.5 3.4

Treatment Any precipitating cause should be treated. Status epilepticus, although rare, should be managed in the normal way.81 There is no evidence to support the routine use of prophylactic anti-epileptic drugs in stroke patients (including those with subarachnoid haemorrhage, section 14.3.3) who have not yet had a seizure but who are thought to be at high risk. Usually, an isolated seizure does not require treatment since the risk of further seizures is only about 50% over the next few years. However, if seizures recur, or if the patient wishes to minimize the risk of recurrence because of the implications for driving, employment or leisure activities, treatment after the first seizure may be warranted. We are not aware of any studies which have compared the efficacy of different anti-epileptic drugs in preventing seizures

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 553

11.9 Headache, nausea and vomiting

after stroke. Indeed, until recently there has been little evidence to suggest that any one of the most commonly used first-line drugs (i.e. phenytoin, sodium valproate, carbamazepine and lamotrigine) is more effective than any other in preventing partial (focal) or generalized seizures in adults.82,83,461,462 The choice will depend on availability, efficacy the risk of adverse effects and cost. Recent evidence suggests that sodium valproate is the most effective agent in generalized seizures and lamotrigine in partial seizures.461,462 Patients who have had a seizure should be advised not to drive (for a period which varies depending on national regulations), and to inform the necessary authorities (www.dvla.gov.uk/at_a_glance/ ch1_neurological.htm, www.epilepsy.com/epilepsy/ social_driving.html, http://www.ltsa.govt.nz/factsheets/ 17.html). However, many stroke patients who have seizures are too disabled to drive anyway (section 11.32.2). The risk of seizures after acute stroke, including subarachnoid haemorrhage, is not high enough to justify routine prophylactic anti-epileptic drugs. These may have significant adverse effects and are expensive.

vertebrobasilar strokes. It may not respond to antiemetics, indeed some believe that these may prevent tolerance developing. The place of so-called vestibular rehabilitation in such cases is unclear.

11.10 Hiccups

Hiccups are due to involuntary diaphragmatic contractions against a closed glottis. The precise cause is unclear. Although uncommon after stroke they may be persistent and troublesome in patients with strokes affecting the medulla.89 If they persist, other causes should be considered (e.g. uraemia, diaphragmatic irritation). Numerous folk cures (e.g. sudden frights), acupuncture and drugs (Table 11.12) have been suggested as effective treatments.90,91 It may be worth trying a short trial of each of the drugs in Table 11.12 when hiccups are persistent and distressing to the patient since any response is likely to be rapid and may not require ongoing drug treatment. Chlorpromazine, baclofen and gabapentin are probably the most frequently used. The choice should take into account the likely adverse effects.

11.9 Headache, nausea and vomiting

Headache is quite a common symptom at the time of stroke and may provide clues to the pathological type (e.g. haemorrhagic) and cause (e.g. giant-cell arteritis) (section 6.7.6). It is often associated with nausea or vomiting, most commonly in haemorrhagic and cerebellar strokes and less often in lacunar infarcts.84–88 It is more common in those with prior migraine. Headache which occurs after the onset of stroke may be caused by treatment, for example due to intracranial bleeding secondary to thrombolysis or more commonly dipyridamole started for secondary prevention (section 16.5.5). Nausea or vomiting without headache is generally secondary to vertigo. These symptoms, which can be severe initially, usually improve within a few days. Having established that there is no important underlying cause (e.g. venous sinus thrombosis, giantcell arteritis, arterial dissection) most patients simply require reassurance that the symptoms will improve, adequate analgesia and/or anti-emetics. However, these drugs may have adverse effects (see case history in section 11.5). Persistent vertigo which may be positional and associated with nausea or vomiting can be a particularly troublesome symptom, most commonly after

11.11 Immobility and poor positioning

Immobility is a major consequence of impaired conscious level; severe motor deficits including weakness, ataxia and apraxia; and less commonly of sensory (i.e. proprioceptive) and visuospatial deficits. Immobility makes the patient vulnerable to a number of complications such as infections (section 11.12), deep venous Table 11.12 Some of the drugs used to treat hiccups (most have important adverse effects). Chlorpromazine Haloperidol Gabapentin Baclofen Metoclopramide Sodium valproate Phenytoin Carbamazepine Nifedipine Amitriptyline

553

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 554

Section Maintenance of a clear airway Cerebral perfusion Cerebral oedema Muscle tone Swallowing Limb oedema Stimulation

11.2.1 11.7.2 11.3 11.20 11.19 11.25 11.31.3

28.0 19.7

P = 0.0001

20.8

11.4 14.1

CBP (mmHg)

Table 11.13 Physiological factors to take into account when deciding in which position a patient should be nursed.

ICP (mmHg)

Chapter 11 What are this patient’s problems?

104.1 7.4 89.5

P = 0.006 98.8

74.9

84.3 69.8

CPP (mmHg)

thrombosis and pulmonary embolism (section 11.13), pressure ulcers (section 11.16), contractures (section 11.20) and falls and resulting injuries (section 11.26). Immobile patients are unable to position themselves to maintain comfort, to facilitate activities such as drinking and passing urine, and to relieve pressure over bony prominences. They are in the ignominious situation of always having to ask others to position them. There has been little formal research of positioning after stroke.92,93 For example, should the patient be nursed sitting or lying, and if lying, on which side? There is some evidence for the optimum positioning of mechanically ventilated patients: nursing semirecumbent is associated with a lower risk of pneumonia.94 In the self-ventilating stroke patient the optimal position(s) are unclear but decisions should take into account the following physiological factors (Table 11.13): • Maintenance of a clear airway: In unconscious patients correct positioning is vital to maintain a clear airway and so reduce the risk of aspiration (section 11.2.1). • Oxygenation: Position may influence the patient’s ability to breathe, ventilate their lungs and oxygenate their blood (section 11.2.2). Oxygenation is usually optimal in the sitting position.92,95 • Cerebral perfusion: Lying flat may increase cerebral perfusion.96 Also, patients who are hypovolaemic may have reduced systemic blood pressure when sitting, which may reduce cerebral perfusion and possibly increase cerebral ischaemia.92,97 However, if intracranial pressure falls further on sitting, cerebral perfusion pressure may increase. • Cerebral oedema: Intracranial pressure is highly dependent on posture, being higher in supine patients and lower when sitting.92,98 The relationships between oxygenation, cerebral perfusion, intracranial pressure and cerebral blood flow are so complex that it is difficult to predict the optimum position for nursing an individual patient (Fig. 11.6). Whether our own anecdotal observation that some patients with severe strokes are more alert when sitting up than when lying

CBF (mL/100 g brain per min)

554

88.4

88.3 P = 0.8

69.7

70.2

51.0

52.1

69.3 48.9

64.7 P = 0.66

28.5

47.8 30.9

0° Position

30° head up

Fig. 11.6 Graph showing the interrelationships between intracranial pressure (ICP), carotid blood pressure (CBP), cerebral perfusion pressure (CPP) and cerebral blood flow (CBF) with changes in posture in patients with head injury. Data from Feldman et al., 1992.98

is explained by reduced intracranial pressure, reduced cerebral oedema, improved oxygenation, and/or increased sensory or social stimulation is not at all clear. • Tone: The position of a patient influences the tone in their trunk and limbs.99 Positioning of the patient is used by physiotherapists and nurses to promote higher or lower tone, whichever seems most appropriate for that particular patient.100 Spasticity and the tendency to develop contractures may be reduced by careful positioning to reduce tone (section 11.20), while patients who have low tone in their truncal muscles may benefit from positioning which promotes increased tone and leads to better truncal control. The positioning charts which are so often strategically placed by the patient’s bed to guide nursing care give the

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 555

11.12 Fever and infection

(Fig. 11.23) are, in our experience, valuable tools in positioning immobile stroke patients and they potentially allow some patients to reposition themselves to maintain comfort.

impression that we know what we are doing (see Fig. 11.22). However, although there is some agreement about the optimum positioning of patients with hemiplegia (e.g. fingers extended, spine straight), there are many areas of uncertainty (e.g. optimum position of the head, foot and unaffected limbs).100 There is clearly a need for further research into this area although this will be hampered by difficulties in prescribing specific positioning regimens, in ensuring compliance, and in reliably assessing outcomes (section 11.20). • Swallowing: Swallowing is easier and safer if the patient is sitting up with their neck flexed (section 11.17). • Pressure area care: For immobile patients who cannot shift their own weight, their position must be changed frequently enough to avoid developing ischaemia of the skin and subcutaneous tissues over bony prominences at weight-bearing points (section 11.16). • Limb oedema: The ankles and the paralysed arm of immobile hemiparetic patients frequently become oedematous and painful. This may increase muscle tone and further reduce function (section 11.25). • Stimulation: It is difficult for patients to see what is going on around them when lying flat. This lack of sensory and social stimulation and contact with daily events will encourage too much sleep during the day and may lead to boredom, reduced morale and sometimes confusion (section 11.31.3). The team should assess (and regularly reassess) each patient and decide which of these potential problems are most important. For example, is the patient hypoxic or dysphagic or at particular risk of pressure ulcers? Depending on this assessment, a positioning regimen which sets out what are thought to be the best positions, and the frequency of repositioning, should be prescribed and re-evaluated. However, educational programmes for nursing staff are probably only moderately effective in increasing the proportion of time that patients are positioned according to the prescribed regime.101 Equipment such as electrically operated multi-sectional profiling beds (Fig. 11.7) and specialist rehabilitation chairs

Fever is quite common after stroke although its frequency obviously depends on the population of stroke patients studied, the definitions used, and the method, timing and duration of monitoring. All published studies have been of hospital-referred patients, most have defined fever as an axillary or rectal temperature of > 37.5°C, and have monitored temperature for 2–7 days. Patients with fever during the first few days after stroke have a worse outcome than those without.102,103 It is unclear whether this adverse prognosis is simply because fever is a marker of a severe stroke (i.e. due to loss of central temperature control or resorption of subarachnoid blood), an indicator of an infective complication (e.g. pneumonia or urinary infection), or increases cerebral damage. The last is an attractive, although unproven concept, since it is consistent with research in animal models which has shown that hyperthermia increases and hypothermia decreases ischaemic cerebral damage.104 Table 11.14 lists some of the potential causes of fever after stroke, of which infection is probably the

Fig. 11.7 A multi-sectional profiling bed – a useful tool for positioning stroke patients. It even helps assess patient

hydration by facilitating examination of the jugular venous pressure, and testing for postural hypotension.

Electrically operated multi-sectional profiling beds are helpful in positioning patients and may provide some immobile patients with the opportunity to reposition themselves. We also find them helpful for examining the jugular venous pressure and testing for postural hypotension, both essential to assessing patient fluid status.

11.12 Fever and infection

555

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 556

556

Chapter 11 What are this patient’s problems? Table 11.14 Causes of fever after stroke. Causes Infective/inflammatory complications of the stroke Urinary infection Pneumonia Deep venous thrombosis Pulmonary embolism Pressure ulcers Infected intravenous access site Vascular problems, e.g. infarction of myocardium, bowel or limb Inflammatory causes of the stroke itself Infective endocarditis Arteritis Coincidental conditions Upper respiratory tract infection Drug allergy

accumulation of secretions, and care to avoid aspiration seem sensible measures to take (section 11.17). Section

11.12.2 Urinary infections 11.14 11.12.1 11.13 11.13 11.16

6.5.9 7.3

most common. Fever and infection may pre-date the stroke onset, and occasionally may actually cause or at least precipitate the stroke (sections 6.6.17 and 7.11). Of course, immobile stroke patients are prone to infections, the most common sites being the chest and urinary tract.105 There is some evidence that patients with stroke and other types of brain damage acquire an immune deficiency, mediated via the hypothalamic pituitary axis, which may contribute to their risk of infection.106,107 However, the clinical relevance of these findings has yet to be established. Infections are (in general) associated with worse outcomes, even having adjusted for other important prognostic factors as far as it is possible to do so, and they cause many deaths, interrupt rehabilitation and slow recovery.108

11.12.1 Chest infections Chest infections are much more common in the acute stage than later, occurring in about 20% of patients within the first month or two. Chest infections may be due to aspiration, failure to clear secretions, the patient’s immobility, reduced chest wall or diaphragmatic movement on the hemiparetic side, or comorbidities including chronic airway disease.109 However, there is no clear relationship between the side of any hemiparesis and side of pneumonia. Alterations in the bacterial flora in the mouth which have been observed after stroke110 and in tube-fed individuals111,112 may be associated with an increased risk of chest infection. Chest infections may be minimized by efforts to improve oral hygiene or alter the mouth flora113 but further randomized trials are required to evaluate this approach. In the meantime careful positioning, physiotherapy and suction to avoid the

About a quarter of hospitalized stroke patients develop a urinary infection within the first 2 months and this remains common over the subsequent months.105 Urinary infections can be avoided by maintaining adequate hydration and thus urine output, and by avoiding unnecessary bladder catheterization (section 11.14). Given that incomplete bladder emptying is associated with an increased risk of infection it seems sensible to avoid constipation and drugs with anticholinergic effects if possible.114 It is uncertain whether intermittent or continuous catheterization and/or prophylactic antibiotics are of benefit for patients with a persisting increased post-void residual volume.115–117 Assessment The patient’s temperature should be monitored at least 6-hourly during the first few days after the stroke and thereafter if there are any signs of infection or functional deterioration. However, fever may not accompany infection, especially in elderly and immunocompromised patients. Therefore, any functional deterioration or failure to attain a rehabilitation goal should prompt a search for occult infection, even if there is no fever. Obviously, the cause of fever should be identified using clinical assessment supplemented with appropriate investigations (e.g. blood neutrophil count, C-reactive protein, cultures of urine and blood, chest X-ray). Dipstick testing of a midstream urine to detect leucocytes and nitrates is insufficiently accurate to be relied on to exclude or confirm a urinary infection (section 11.14).118 Prevention and treatment There is currently little evidence to support the use of prophylactic antipyretics to reduce fever after stroke (section 11.19). In fact, the routine use of paracetamol to lower body temperature and attempts to induce more profound hypothermia after stroke are being evaluated in ongoing RCTs119 (http://www.strokecenter.org/ trials/index.htm). A single-centre RCT of levofloxacin (a fluoroquinolone antibiotic) in acute stroke failed to demonstrate a reduction in infections but further trials will be required to establish whether alternative antibiotic regimes might be more effective.120,121 The treatment of fever will depend on the cause (Table 11.14). We quite often start a broad-spectrum antibiotic, once specimens for microbiological testing have been taken,

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 557

11.13 Venous thromboembolism

(a)

(b) Fig. 11.8 (a) An MR scan of the legs using direct thrombus imaging to demonstrate extensive deep venous thrombosis (arrows). (b) A colour flow Doppler image of the femoral vein demonstrating thrombus within the lumen (arrow); this is

usually associated with an inability to compress the vein and if the vein is occluded a loss of augmentation of flow with respiration or squeezing the calf. (Also reproduced in colour in the plate section.)

since delays in treating infections may impede patients’ progress. However, with an increasing incidence of Clostridium difficile toxin-associated diarrhoea in many hospitals, the risks of early use of broad-spectrum antibiotics must be carefully weighed against the potential benefits.122 Appropriate antibiotics and supportive treatment (e.g. physiotherapy, oxygen) should be given in established infection. Also, it seems reasonable, whatever the cause of the fever, to use a fan and prescribe paracetamol (acetaminophen) since fever may worsen outcome (see above). Even if such interventions appear to be without risk, one must remember that they take up a nurse’s time which might be spent to greater effect in some other activity.

of the frequency depend on the types of patients included, the use of preventive measures, and the timing and method of detection. The most sensitive techniques, such as MRI direct thrombus imaging, detect DVTs in about 40% of hospitalized patients treated with aspirin and graduated compression stockings and about half of these are above-knee123 (Fig. 11.8a). Less sensitive, but more widely available, tests such as compression Doppler ultrasound (Fig. 11.8b) and plethysmography identify DVTs in about 20% of immobile hospitalized patients, above-knee in about half of these. Although DVT is said to be less common among Chinese a recent study has shown a similar incidence to that seen in white patients.124 Clinically apparent DVT confirmed on investigation is less common, occurring in under 5%. DVTs are most often asymptomatic, or unrecognized, but may still lead on to important complications.123 Pulmonary embolism (PE) occurs in about 10% of hospitalized patients but like DVT is frequently not recognized.123 It is an important cause of preventable death after stroke and is a frequent finding at autopsy.125

Any functional deterioration or failure to attain a rehabilitation goal should prompt a search for occult infection.

11.13 Venous thromboembolism

Deep venous thrombosis (DVT) of the legs is common in patients with a recent stroke, particularly older patients with a severe hemiplegia who are immobile. Estimates

Assessment DVT should be suspected if a patient’s leg becomes swollen, hot or painful or if the patient develops a fever. Unfortunately, the clinical diagnosis can be difficult because many paretic legs become swollen, mostly

557

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 558

558

Chapter 11 What are this patient’s problems?

because of the effects of gravity and lack of movement. If a paretic leg swells while a patient is still being nursed in bed, DVT is a likely cause, but where a patient is sitting out or mobilizing the clinical diagnosis is much less certain. Stroke patients who have communication difficulties, sensory loss or neglect may well not complain of discomfort or swelling, so that clinical detection will depend on the vigilance of members of the multidisciplinary team. Stroke patients who have communication difficulties, sensory loss or neglect may well not complain of discomfort or swelling associated with deep venous thrombosis, so that clinical detection will depend on the vigilance of members of the multidisciplinary team. If a patient develops a swollen leg on a stroke unit, deep venous thrombosis has to be actively excluded. Where the patient develops clinical evidence of a DVT or pulmonary embolism, confirmatory investigations must be carried out if treatment with anticoagulants is being considered. We would normally use compression Doppler ultrasound in the first instance to confirm the diagnosis since this is non-invasive, widely available and reasonably sensitive (>90%) and specific (>90%) in detecting at least above-knee DVT in symptomatic patients.126 However, it is operator dependent and if there is doubt about the result, or if one wishes to exclude thrombosis in the calf veins, contrast X-ray venography should be performed. The clinical diagnosis of deep venous thrombosis in stroke patients is particularly difficult because, on the one hand, a swollen leg may be due to paralysis and dependency while, on the other hand, a patient may not complain about pain and swelling because of language and perceptual problems. The value of screening asymptomatic stroke patients for DVT has not been established. In considering such a policy one has to remember that the sensitivity and specificity of non-invasive tests, such as D-dimers and compression Doppler ultrasound, are lower in patients who do not have symptoms of DVT than in those with symptoms, and more ‘positives’ will be ‘falsepositives’.127–129 Given the high frequency of PE in hospitalized stroke patients, if a patient has clinical features compatible with PE – breathlessness and/or tachypnoea, with or without pleuritic chest pain and/or haemoptysis for which there is no other reasonable clinical explanation – the probability of PE is high. In such patients a CT

pulmonary angiogram can be used to confirm or exclude a PE130 because it is difficult to distinguish clinically between breathlessness due to pulmonary embolism and other causes without further investigation. Prevention Manoeuvres which may reduce the risk of DVT and pulmonary embolism include: • Early mobilization of the patient and avoidance of prolonged bed rest, although the effectiveness of this regimen after stroke has never been tested in RCTs. • Hydration/fluids may influence the risk of venous thromboembolism. A raised urea, probably indicating dehydration, is associated with a higher risk of DVT.131 Also a systematic review of RCTs testing haemodilution in stroke indicated that this probably reduces the risk of DVT and PE (odds ratio 0.54, 95% confidence interval 0.30–0.99).132 It is unclear whether this is a specific effect of haemodilution or a non-specific effect of improved hydration. We give intravenous saline to most of our patients with acute stroke and immobility, in part because they are often unable to take adequate fluids orally (sections 11.17 and 11.18.1). • Full-length graduated compression stockings. Systematic reviews of RCTs in patients undergoing surgery have shown about a 60% reduction in the odds of developing DVT.133,134 Most of these RCTs tested full-length stockings or did not specify the length. There is very little evidence that short (below-knee) stockings effectively prevent DVT.134 But, in stroke, unlike surgery, stockings cannot be applied before the onset of the insult (i.e. the surgery itself), so DVTs may develop before stockings can be applied, and patients may be immobile for weeks and have prolonged leg paralysis. Moreover, compression stockings have potential hazards: they occasionally cause pressure ulcers (Fig. 11.9a) or acute limb ischaemia, the latter particularly in those with diabetes, peripheral neuropathy or peripheral vascular disease (Fig. 11.9b), they are uncomfortable and unpopular with patients and nurses; and considerable nursing resources are consumed in their application and monitoring. If used, stockings should be fitted in accordance with the manufacturer’s instructions and removed daily to check for skin problems. The evidence for their effectiveness in stroke patients is inconclusive but a large randomized trial is in progress (www.clotstrial.com;135 http:// www.strokecenter.org/trials). • Aspirin, started within 48 h of a presumed ischaemic stroke, reduces the relative risk of PE by about 30% and improves the patients’ overall long-term outcome136 (section 12.3). We start aspirin routinely (first dose

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 559

11.13 Venous thromboembolism

(a)

300 mg and 75 mg per day thereafter) as soon as we have excluded intracranial haemorrhage or other contraindications. • Heparin has been shown to reduce the risk of DVT in patients with ischaemic stroke, but this benefit is offset by the risk of haemorrhagic complications so that routine use of heparin does not improve overall outcome137 (section 12.4). We reserve unfractionated (5000 u twice per day) or low-molecular-weight heparin for patients we judge to be at particularly high risk of DVT and PE and low risk of haemorrhagic complications, accepting that these judgements are based on inadequate evidence. Such patients might include those with an ischaemic stroke with severe leg weakness and immobility, and cancer, thrombophilia or previous venous thromboembolism (section 12.4.3). • Other methods of prophylaxis, e.g. external pneumatic compression and functional electrical stimulation, have been suggested but not evaluated adequately in stroke.135,138 Treatment If a patient with a confirmed ischaemic stroke has a proven above-knee DVT or pulmonary embolism, we use subcutaneous low-molecular-weight heparin (LMWH) despite the lack of direct evidence that this is more effective than intravenous unfractionated heparin in stroke patients. We prefer LMWH because it is easier to use and does not need monitoring, and there is evidence in patients other than those with stroke that it is just as, or more, effective.139–141 If patients are symptomatic we also usually treat DVT restricted to the calf veins although this depends on the presence of any relative contraindications to treatment.142 In patients with DVT restricted to calf veins who have no symptoms we would also anticoagulate if a repeat Doppler ultrasound showed propagation of the thrombus in the popliteal or femoral veins. We normally continue heparin for a few days while starting oral anticoagulants which we continue for about 6 months, depending on our judgement of the patient’s risk of recurrence, taking into account their

(b)

Fig. 11.9 (a) A pressure ulcer behind the knee due to graduated compression stockings and inadequate monitoring by nursing staff. (b) A patient with peripheral vascular disease and diabetes mellitus who was fitted with graduated compression stockings. Note the necrotic skin over the anterior border of the tibia (white arrow) and where the stockings were creased at the ankle (black arrow). There were also necrotic areas over both heels which failed to heal and led to a right above-knee amputation following an unsuccessful revascularization procedure.

559

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 560

560

Chapter 11 What are this patient’s problems?

immobility and risk of bleeding.143 We do not believe that intracranial haemorrhage is an absolute contraindication to anticoagulation since, if a patient has had a life-threatening pulmonary embolus, the risk of anticoagulation may be worth taking. Alternatively one might consider insertion of a caval filter but there are no reliable studies which help us decide between these two treatments.144 In a small number of patients, thrombolysis may need to be considered for massive PE or DVT but this has not been evaluated in stroke patients and would inevitably be associated with a risk of intracranial bleeding.

11.14 Urinary incontinence and retention

Between one- and two-thirds of acute stroke patients admitted to hospital are incontinent of urine in the first few days.145,146 Urinary incontinence is more common in older patients, those with severe strokes, other disabling conditions and diabetes.145,147 Urinary incontinence may be caused by the stroke itself, but perhaps onefifth of patients have been incontinent before the stroke. Although detrusor instability is the most common single cause of urinary incontinence after the first 4 weeks, many other factors may contribute in the acute stage (Table 11.15). Urinary incontinence is an important cause of distress to patients and carers, increases the risk of pressure ulcers (section 11.16), often interferes with rehabilitation (e.g. by interrupting physiotherapy sessions or increasing spasticity; section 11.20) and influences the patient’s requirements for ongoing nursing care.147,148 Urinary incontinence may also cause patients to become dehydrated because patients tend to restrict their own fluid intake to limit urinary incontinence without telling the nurses or doctors. Urinary incontinence may cause patients to become dehydrated. This is because many patients restrict their own fluid intake to limit urinary incontinence – without telling the nurses or doctors. Assessment To identify patients with urinary incontinence one simply has to ask the carer or nursing staff. These are the people most aware of urinary problems since they have to deal with the consequences. It is important to ask, since many people consider incontinence an inevitable consequence of stroke and thus not worthy of mention.

Table 11.15 Factors which may contribute to urinary incontinence. Section Reduced conscious level Immobility (cannot get to the toilet in time) Communication problems (cannot ask to go to the toilet) Impaired upper limb function (cannot manipulate clothes or the urinal) Dyspraxia Loss of inhibition of bladder contraction (detrusor instability so cannot wait to go) Urinary infection (often without any other symptoms) Urinary overflow due to outflow obstruction (e.g. prostatism) Faecal impaction Excess urinary flow due to high fluid intake, diuretics and poorly controlled diabetes Too few carers/nurses (cannot attend to patients in time) Importance of maintaining continence underestimated by carers/nurses

11.3 11.11 11.30 11.21 11.28

11.12.2

11.15 11.18.3

Routine use of a measure such as the Barthel Index to monitor patient progress on the stroke unit should identify all patients with urinary incontinence. It is often useful, but frequently overlooked, to ask the patients themselves what they think is causing their incontinence. This may, for instance, help distinguish true incontinence from accidental spillage of urine from a urinal because of the patient’s poor manual dexterity (Fig. 11.10). Many of our patients with severe stroke, often with some cognitive dysfunction, seem unaware of either voiding or being wet. They usually deny incontinence when asked. More detailed information, including urinary volumes, frequency and times of voiding, which can be collated with a micturition chart, may be useful in identifying the causes of incontinence (e.g. diuretics, communication difficulties) and in formulating a management plan. Many of our patients with severe stroke, often with some cognitive dysfunction, seem unaware of either voiding or being wet. They usually deny incontinence when asked. Although we exclude specific causes and try both behavioural and pharmacological interventions we usually resort to incontinence pads or an indwelling catheter. Where the cause of urinary incontinence is unclear, and if it persists for more than a few days, the patient

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 561

11.14 Urinary incontinence and retention

(a)

(b) Fig. 11.10 Some patients have difficulty manipulating a urinal which leads to spills and, effectively, incontinence. (a) This urinal can be inverted without leaking; (b) a simple

one-way valve inserted in the neck of the urinal prevents the contents spilling.

should be investigated. Urine dipstick testing for leucocytes and nitrates may help to exclude an infection but, given the high frequency of urinary infection in stroke patients (section 11.12.2), it is insufficiently accurate to be used as a single test; microscopy and culture should also be performed to exclude or confirm the diagnosis.118 Measurement of postmicturition bladder volumes (by bladder ultrasound or catheterization) may be useful in assessing bladder sensation, contractility and outflow. We reserve formal urodynamic studies, which can better identify detrusor hyper- and hyporeflexia and bladder outflow problems, for the few patients with unexplained, troublesome incontinence which persists for weeks after the stroke and where the incontinence is out of proportion to the stroke severity.

among survivors, incontinence resolves in the majority, unless it pre-dated the stroke. Younger age, less severe stroke and lacunar stroke have all been associated with better rates of resolution.147,149 Persisting urinary incontinence is associated with worse functional outcome and higher rates of institutionalization.147,149,150

A negative urine dipstick test for leucocytes and nitrates does not reliably exclude a urinary infection; a midstream urine collection should be obtained if possible and sent for microscopy and culture.

Prognosis Urinary incontinence is an important predictor of poor survival and functional outcome after acute stroke (section 10.2.7): between 30% and 60% of incontinent patients die within the next few months.145,149 However,

Treatment There is very little reliable evidence about the treatment of post-stroke incontinence.151 If a patient is incontinent, but able to understand, then a careful explanation of the cause and likely prognosis should be given. Carers often benefit from such information too. Because urgency of micturition is such a common cause of incontinence in stroke patients, simple steps such as regular toileting, offering aphasic patients some means to alert the nurses to their needs, improving their mobility, or providing a commode by the bed, can all be effective. Provision of suitable clothing, such as trousers with a flap fastened with Velcro, may allow patients to use a urinal or commode independently and thus promote continence. Obviously, one should strive to treat the underlying cause (e.g. infection, outflow obstruction) and where possible remove exacerbating factors (e.g. excessive fluids, uncontrolled hyperglycaemia or diuretics). Having excluded easily treatable causes, we first employ ‘bladder retraining’ where patients are prompted to void

561

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 562

562

Chapter 11 What are this patient’s problems? Table 11.16 Drugs used to inhibit bladder contractility, and their adverse effects. Anticholinergic drugs Flavoxate hydrochloride Oxybutynin hydrochloride Tolterodine tartrate Propiverine hydrochloride Propantheline bromide Tricyclic antidepressants Imipramine Amitriptyline Nortriptyline Common adverse effects Dry mouth Blurred vision Nausea/vomiting Constipation/diarrhoea Confusion in the elderly Retention where there is bladder neck obstruction Precipitation of acute glaucoma

regularly. If this does not achieve continence, we use a bedside ultrasound machine to exclude a postmicturition residual of 100 mL or more and then introduce an anticholinergic drug (e.g. oxybutinin or tolteridine) assuming there are no contraindications, e.g. closed angle glaucoma.152 One must obviously be alert to the possible adverse effects of these drugs (Table 11.16). This approach achieves reasonable results with relatively few patients requiring formal urodynamic studies.153 An indwelling catheter should be avoided if at all possible because it makes resolution of urinary incontinence impossible to detect, and may lead to a number of complications (Fig. 11.11, Table 11.17). Intermittent catheterization facilitates detection of the resolution of incontinence and probably reduces the rate of bacturia but it is certainly more labour intensive and therefore may not be practical.114 Other aids and appliances may be useful (Table 11.18) in avoiding unnecessary catheterization. However, when patients are at high risk of pressure ulcers (section 11.16), and other means have failed to keep them dry, or if accurate monitoring of fluid balance is required for some reason, an indwelling catheter is the best option. Catheterization may also be required to relieve urinary retention (see below) until any cause or precipitating factor can be removed, e.g. enlarged prostate, urinary infection, severe constipation, anticholinergic drugs. There is limited evidence that prophylactic antibiotics reduce the rate of bacturia in patients with indwelling catheters and this policy may increase bacterial resistance and rates of Clostridium difficile-associated diarrhoea.112,114

Fig 11.11 Traumatic hypospadias caused by an indwelling catheter. This is most likely to occur in a man with cognitive or communication problems who may not indicate he is uncomfortable, and who is being looked after by inadequately trained nursing staff.

Occasionally, if urinary incontinence is a bar to discharge into the community, long-term catheterization with a silastic catheter may be the preferred option. This issue should be discussed with the patient and, where appropriate, their carer. Continence advisers, nurses with specialist training in the management of incontinence and access to information, aids and appliances, can often help other professionals, patients and their carers.151 The choice of continence aids and catheters is huge but there are few rigorous studies to establish which are the most cost-effective, and most acceptable to patients and carers.154,155 In many parts of the UK, laundry services run by health authorities or social services provide invaluable assistance to families having to cope with incontinence. An indwelling catheter should be avoided if possible because this makes resolution of urinary incontinence impossible to detect and may lead to a number of complications.

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 563

11.14 Urinary incontinence and retention Table 11.17 Problems (and solutions) of indwelling urinary catheters.434 Problem

Solution

Pain on insertion

Explain to the patient what is going to happen Use plenty of anaesthetic gel and allow time for it to work Ensure foreskin is not left retracted after insertion Avoid traction and monitor carefully; if it occurs replace with suprapubic catheter Explain why catheter is needed, how it works and how long it will be in place Provide a discreet drainage bag Use well-supported leg bag for mobile patients Use appropriate size of catheter Inhibit any involuntary bladder contraction which causes bypassing with an anticholinergic drug (Table 11.16) Change catheter if blocked Ensure adequate urine flow Remove encrusted catheter Avoid unnecessary catheterization since no proven method of preventing infection Ensure balloon inflated to correct volume or use larger volume balloon

Paraphimosis Traumatic hypospadius (Fig. 11.11) Poor self-esteem Immobility because of drainage bag Leakage

Blockage Infection Catheter falls out due to urethral dilatation or pelvic floor laxity Catheter rejection due to bladder contraction Catheter pulled out by patient Pain on catheter removal

Failure of balloon deflation

Avoid large volume balloon Inhibit with anticholinergic drug (Table 11.16) Manage without a catheter to avoid further trauma Avoid routine changes Explain procedure to patient Allow adequate time for balloon deflation Introduce ureteric catheter stylet along inflation channel

Urinary retention Table 11.18 Aids and appliances which may be useful in patients with urinary incontinence.435 Absorbent pads and pants These vary in the volume of urine they can absorb, their shape, and the method of holding them in position. Urinals Useful for men who are immobile or have urgency which gives them insufficient time to reach the toilet. They can be fitted with a non-spill valve for patients who have poor manual dexterity (Fig. 11.10), or fluid absorbing granules, to reduce spillage. Bedside commode Useful where urgency is associated with poor mobility so the patient has insufficient time to get to the toilet (Fig. 11.33b). Penile sheath Often viewed as an alternative to an indwelling catheter in men without bladder outflow obstruction, but they easily fall off and are therefore unsuitable for agitated or confused patients. Other problems include skin erosions due to urinary stasis or the adhesive strip, and twisting of the sheath and penile retraction during voiding which causes leakage.

Urinary retention, which may be acute or chronic, is common in stroke patients, more so in men. The main cause is pre-existing bladder outflow obstruction which may be exacerbated by constipation, immobility and drugs such as tricyclic antidepressants which have antimuscarinic effects.156 Urinary retention may present with dribbling incontinence, agitation or confusion and is easily missed in patients with a reduced conscious level, communication difficulties or other cognitive problems. It is important to palpate the patient’s abdomen or if in doubt perform a bladder ultrasound, on admission and later, if urinary problems or agitation develop, to exclude a distended bladder. Chronic retention with a postmicturition bladder volume of greater than 150 mL increases the risk of infection. A urethral catheter provides prompt relief. Removal of any exacerbating factors may avoid the need for a catheter or at least allow it to be removed quickly. In men with benign prostatic hypertrophy, alpha-blocking drugs (e.g. prazosin) or finasteride (which inhibits the metabolism of testosterone to dihydrotestosterone in the prostate) may enable one to remove the catheter without recurrence of

563

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 564

564

Chapter 11 What are this patient’s problems?

retention. Surgeons and anaesthetists are often unwilling to consider transurethral resection of an enlarged prostate until several months have passed following a stroke. We are not aware of any evidence to indicate how long we should delay. It is important to palpate the patient’s abdomen on admission and, later, if urinary problems or agitation develop, to exclude a distended bladder.

11.15 Faecal incontinence and constipation

Constipation is very common after stroke and may lead to abdominal discomfort, anxiety and faecal smearing or incontinence. Immobility, poor fluid and food intake, constipating analgesics and anticholinergic drugs are common causes. Faecal incontinence affects about onethird of patients after stroke and has been associated with increasing age, diabetes, other disabling conditions, stroke severity, immobility and size of brain lesion.145,157

Faecal incontinence which is not associated with severe cognitive problems is almost always remediable by dealing with constipation or diarrhoea. The rectum is often full of soft faeces which the patient cannot evacuate effectively. Treatment There is little reliable research to guide treatment.158 Avoidance of constipation by ensuring an adequate intake of fluid and fibre is the best approach, but laxatives, suppositories and, occasionally, enemas are sometimes required. We find that stimulating laxatives (e.g. senna) are generally more effective than osmotic ones (e.g. lactulose) in elderly patients although the choice will be influenced by whether the stool is hard or soft. It is important to remember that laxatives may cause incontinence in immobile patients. Where patients are unable to toilet themselves, and a carer is not constantly available, it may be necessary to induce constipation, with for example codeine phosphate, and then relieve this with regular enemas to coincide with visits from a carer. Simple interventions including advice on diet, fluids and use of laxatives are helpful in reducing problems in the longer term.

Assessment The frequency of bowel movements should be monitored. Simple monitoring will detect constipation and diarrhoea and may help establish the pattern of any faecal incontinence. Review of the drug chart will often identify possible causes. Abdominal and rectal examination will usually identify faecal impaction and indicate whether the constipated stool is hard or soft. Occasionally, if the patient has faecal incontinence associated with diarrhoea, it may be useful to culture the stool, and test for Clostridium difficile toxin to exclude infection or to X-ray the abdomen to exclude high faecal impaction. More detailed investigation is not required unless there are persistent unexplained problems. Prognosis Faecal incontinence early after stroke onset often resolves in survivors but may develop among those who were initially continent.145,157 Later development of faecal incontinence is usually due to preventable causes such as drugs and inadequate staffing levels. Achieving continence can be a crucial step in discharging a patient home since faecal incontinence is so practically and socially difficult to cope with, and is invariably a considerable strain for the carer.

11.16 Pressure ulcers

Pressure ulcers occur when local pressure on skin and subcutaneous tissues exceeds the capillary opening pressure for long enough to cause ischaemia. In addition, friction may cause blistering and tears in the skin. Pressure ulcers usually occur over weight-bearing bony prominences (Fig. 11.12). Ulcers occur in patients who are immobile and unable to redistribute their own weight when lying or sitting. The frequency of pressure ulcers in hospitalized stroke patients is bound to vary depending on the population studied, the prophylactic methods used and the diagnostic criteria. Pressure ulcers are more common in patients who are malnourished, infected, incontinent or have serious underlying disease.159 They cause pain, increase spasticity, slow the recovery process, and may be fatal if complicated by severe sepsis. They prolong length of stay in hospital, often require intensive treatment and can therefore be extremely expensive to health services.160 They can and should be prevented, although they may develop in the interval between the onset of the stroke and admission to hospital.

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 565

11.16 Pressure ulcers

(a)

Assessment Immobile patients should be examined regularly (sometimes several times in a day) to identify early signs of pressure damage, i.e. skin redness. It is important that patients who are at particular risk of developing pressure ulcers are identified as early as possible so that preventive measures can be taken. The Waterlow Scale (Table 11.19) is one of many clinical scoring systems developed to indicate an individual patient’s risk. Most scales include some measure of mobility, continence, cognitive function and nutritional status and none, based on less than ideal evaluations, is clearly superior to the others.161

Skin Subcutaneous tissue Bony prominence

(b)

Shoulders (3%) Spine (2%)

Sacrum (36%) Trochanter (3%)

Natal cleft (3%)

Heels (25%)

Buttocks (21%)

Malleoli (2%) Other (4%)

Fig. 11.12 (a) The distortions of tissues over a bony prominence due to compression or shear which may lead to pressure sores and (b) the anatomical distribution of established pressure sores based on data from a cross-sectional UK survey of pressure sores in patients being nursed on the Pegasus Airwave System. Although patients with cerebrovascular disease were the largest group in this sample, they made up only 14% of the whole group. (From St Clair, 1992.459)

Each patient’s risk of pressure ulcers should be assessed and documented; actions, appropriate to the level of risk should be taken to prevent pressure ulcers developing. Pressure ulcers can be prevented by good nursing. The clinician should be alert to behaviour which, in patients with communication and cognitive problems, may indicate a painful pressure area. Patients may repeatedly move themselves out of a desired position. For example, patients, colloquially known as ‘thrusters’, with a painful sacrum may force themselves out of a chair by extending their bodies at the waist. This may become a major problem for nursing staff. If patients develop pressure ulcers it is useful to have some objective measure of their severity so that healing or lack of healing can be monitored. Photographs incorporating a centimetre scale are a convenient and reliable method to demonstrate change, but where this facility is not available, tracing the limits of the ulcer or simply measuring it in several planes is useful.162 Patients who are at risk, or who have established pressure ulcers, should be investigated to exclude malnutrition, hypoalbuminaemia, anaemia and infection (in the pressure ulcer or elsewhere), all of which can slow healing.162 Prevention The most important way to prevent ulcers is to relieve the pressure on the tissues for long enough, and at frequent enough intervals, to allow the tissues to receive an adequate blood supply. This can usually be achieved by regular turning of patients (2- to 4-hourly depending on the assessment of risk) but this takes up a lot of skilled nursing resources. Although the introduction of a variety of special mattresses and cushions (Table 11.20) may reduce the need for regular turning, most patients do still need to be turned. Some beds are designed to turn the patient automatically (e.g. the net suspension bed).

565

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 566

566

Chapter 11 What are this patient’s problems? Table 11.19 The Waterlow Scale. Build/weight

Visual skin

Average 0 Above average 2 Below average 3

Healthy Tissue paper Dry Oedematous Clammy Discoloured Broken

Continence 0 1 1 1 1 2 3

Mobility

Complete Occasionally incontinent Catheter/incontinent of faeces Doubly incontinent

0 1 2 3

Fully mobile Restricted/difficult Restless/fidgety Apathetic Inert/in traction

Sex/age 0 1 2 3 4

Male Female 14–49 50–64 65–74 75–80 81+

Appetite 1 2 1 2 3 4 5

Average 0 Poor 1 Anorexic 2

A total score of 10 indicates a patient is at risk of pressure ulcers, one of 15 indicates a high risk and a score of 20 a very high risk. In addition to the basic scale in which the scores for each of the six domains (i.e. weight, skin, continence, mobility, age/sex and appetite) are summed, additional points are added for special risk factors: poor nutrition (8 points); sensory deprivation including stroke (5 points); high-dose anti-inflammatory drugs, steroids (3 points); smoking > 10/day (1 point); orthopaedic surgery or fracture below waist (3 points).

Table 11.20 Specialized mattresses, cushions and beds. Passive systems Sheepskin fleeces and bootees which reduce skin shear and moisture; natural fleeces are better than man-made ones but they are rendered ineffective by poor cleaning and being covered by sheets. Padded mattresses containing polyester fibres, e.g. Spenco. Polystyrene bead system. Foam mattresses (e.g. Vaperm) vary in their pressure-relieving properties. Gel pads can be used under heels and sacrum. Roho cushions are effective but very expensive (Fig. 11.13a). Active systems Ripple mattresses and airwave systems provide alternating pressure; the larger the cells the better but they tend to break down and leak. Low air loss systems (e.g. Mediscus) providing constant low pressure; although effective, tend to be noisy, expensive and complex, needing regular maintenance and training (Fig. 11.13b). Flotation beds and deep water beds are difficult to nurse patients on, are very heavy and some patients get motion sickness. Dry flotation providing constant low pressure produced by glass microspheres blown by air. Air can be turned off when the patient needs to be repositioned. Effective but bulky and expensive. Mechanical beds turn the patient, e.g. net suspension bed. Effectiveness uncertain, patients may not like being suspended (in full view of people around them). Useful for turning patients who are in pain.

Pressure-relieving mattresses and cushions (Fig. 11.13a) are divided into ‘passive’ and ‘active’ systems (Table 11.20). The ‘passive’ systems distribute the patient’s weight through a larger area and make it easier for them to reposition themselves. High-specification foam mattresses are more effective than standard foam mattresses.160 ‘Active’ systems (Fig. 11.13b) usually work by inflating and deflating air cells to relieve pressure on each point at regular intervals, and are more effective than the ‘passive’ systems in intensive care patients.160 However, they are expensive and can make certain nursing tasks more difficult, e.g. positioning a patient to reduce the risk of contractures, to help breathing and to facilitate swallowing, because they offer a less firm base.

Other interventions such as staff education, nutritional support and local treatments (e.g. creams, lotions) applied to unbroken areas of skin may have a role in prevention but their effectiveness has not yet been demonstrated.163,164 The ultimate choice of preventive method will depend on an assessment of the individual patient’s risk of pressure ulcer, the availability of nurses, the patient’s other needs, e.g. positioning, and available resources. Further research is required to identify the most cost-effective strategy for preventing pressure ulcers.160 Studies will have to take into account patients’ absolute risk of developing ulcers, the reduction in risk with each intervention and the cost of the intervention, as well as the cost

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 567

11.17 Swallowing problems

(a)

(b)

Fig. 11.13 Pressure-relieving cushion and bed: (a) Roho cushion; (b) low air-loss bed.

of treating any pressure ulcers which develop. However, we believe that, whatever technology is employed, adequate numbers of skilled nursing staff will still be essential if pressure ulcers are to be prevented. Finally, it is important to recognize that certain management strategies (e.g. positioning in a chair, use of graduated compression stockings) may actually cause pressure ulcers unless they are properly applied and monitored (Fig. 11.9). Treatment For patients with established pressure ulcers, the relief of pressure probably remains the most important factor in promoting healing. ‘Active’ pressure-relieving systems are more effective than ‘passive’ ones.165 In addition, it is important to optimize the patient’s general condition by providing a good diet with adequate protein intake and by treating concurrent illness aggressively (e.g. infections, cardiac failure). This may need intensive nursing input in frail, elderly, anorexic patients who are drowsy or have swallowing problems. The pain associated with pressure ulcers may increase tone and lead to contractures (section 11.20) which may well hinder rehabilitation, while the discomfort may also affect the patient’s morale and even further worsen their outcome. Adequate analgesia, with opiates if necessary, should be given to patients, especially before renewal of dressings. Antibiotics may be required if there is local or systemic infection (spreading cellulitis, osteomyelitis). Debridement to remove necrotic tissue, and skin grafting to achieve skin coverage, may sometimes be necessary. A bewildering variety of local dressings and treatments (e.g. vitamin C, zinc, ultrasound, electrical stimulation, ultraviolet light) which aim to promote healing and

reduce infection are available. Some small RCTs evaluating these interventions have been reported, but need to be systematically reviewed.166,167

11.17 Swallowing problems

Up to one-half of conscious patients admitted to hospital with an acute stroke cannot swallow safely on bedside testing.168 However, estimates of the frequency of swallowing difficulties vary because of differences in definitions, timing and methods of detecting dysphagia and in selecting patients for study. Swallowing difficulties have been associated with a high case fatality and poor functional outcome and certainly put patients at risk of aspiration, pneumonia, dehydration and malnutrition.168 However, much of the excess mortality and morbidity is probably due to the severity of the stroke itself rather than to swallowing difficulties.

11.17.1 Mechanisms of dysphagia The following patterns of dysphagia after stroke have been identified with videofluoroscopy:169,170 • Poor oral control (oral preparatory phase) and delayed triggering of the swallow leads to aspiration before the swallow, i.e. liquid trickles over the back of the tongue before the swallow starts. In addition, patients with weakness or incoordination of the face or tongue often have difficulty keeping fluids in their mouth, and in chewing and manipulation of food to produce a wellformed food bolus.

567

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 568

568

Chapter 11 What are this patient’s problems?

• Failure of laryngeal adduction leads to aspiration during the swallow itself. • Reduced pharyngeal ‘peristalsis’, or cricopharyngeal dysfunction, may allow food to collect in the pharynx and spill over, past the vocal cords and into the trachea. Thus, aspiration occurs after the swallow. Poor oral control and delayed triggering of the swallow are the most common mechanisms causing dysphagia after stroke but more than one abnormality can usually be identified in any individual patient.

11.17.2 Detection of dysphagia Despite their frequency, and the serious consequences of failing to detect them, swallowing problems are often not sought systematically in patients admitted to hospital with an acute stroke.60 The current recommended approach is for a suitably trained nurse or other healthcare worker to perform a simple bedside swallow screen (Fig. 11.14), to distinguish patients able to take oral food and fluids safely from those who require a more detailed

Addressograph, or Name

SWALLOWING SCREENING TEST – an interdisciplinary tool Pre-assessment criteria: Site: If the patient is drowsy and unable to sit upright, than it is NOT SAFE to complete this assessment. They should remain Nil By Mouth (NBM). Monitor conscious level: attempt to screen daily until completed.

DoB Unit number

NB: Nutritional screening required 6.7 mmol/L, or 120 mg/dL) occurs in about one-third of non-diabetic acute stroke patients and two-thirds of those with diabetes.194 About one-quarter of those with hyperglycaemia are known to have diabetes mellitus already and another quarter have a raised HbA1C which suggests that their blood glucose has been high for some time before the stroke, referred to as ‘latent diabetes’.195 Whether the hyperglycaemia in non-diabetics is due to release of catecholamines and corticosteroids as part of the stress response is controversial.196 Hyperglycaemia, after a stroke, at least in non-diabetic patients, is associated with increased case fatality and poor functional outcome.194,197 This could be explained by more severe strokes producing a greater stress response and, thus, hyperglycaemia so that hyperglycaemia is simply a marker of severe stroke. However, some studies have demonstrated that hyperglycaemia is associated with a poor outcome having adjusted for stroke severity and other baseline prognostic factors.194 This finding, along with some (but not all) animal work showing that hyperglycaemia can exacerbate ischaemic neuronal damage, has led many to believe there is a causal relationship between hyperglycaemia and poor outcome.197,198

stroke. This will keep the patient free from thirst and avoid excessive diuresis which may cause dehydration (section 11.18.1). Whether more aggressive control of blood sugar, which has at least theoretical benefits for the ischaemic penumbra, is sensible will depend on any benefits, and the risks of hypoglycaemia, which will almost certainly depend on the intensity of monitoring available.198 One small and one moderate sized, but underpowered RCT have shown that glucose-potassiuminsulin (GKI) infusions can be used with reasonable safety to control hyperglycaemia in acute stroke but have not demonstrated any improvement in survival or functional outcomes.199,200

11.18.4 Hypoglycaemia Hypoglycaemia may occasionally mimic stroke or transient ischaemic attack (sections 3.4.5 and 7.16). Ideally, it should be excluded on first contact with medical services (e.g. paramedics, primary care physician) by measuring the glucose on a capillary sample in all patients taking hypoglycaemic medication.201 Hypoglycaemia also occurs after stroke because of efforts to normalize blood sugar in those with hyperglycaemia (section 11.18.3) and in diabetic patients receiving hypoglycaemic medication but a reduced food intake because of dysphagia (section 11.17) or other post-stroke problems (section 11.19.1). Since hypoglycaemia may, if severe or undetected, cause worsening of the neurological deficit, the blood sugar should be monitored particularly carefully in diabetic patients on hypoglycaemic medication.

Assessment A random blood glucose should be measured in all patients with stroke. In those with hyperglycaemia, a fasting blood glucose and an HbA1C will help distinguish latent diabetes from hyperglycaemia due to the stroke itself. Blood sugar levels are likely to fall spontaneously in the first few days after stroke onset.199 If necessary, a glucose tolerance test after the acute stage of the illness (i.e. when the patient is medically stable) may be helpful in sorting out which patients have diabetes or impaired glucose tolerance, and which simply have hyperglycaemia related to the acute stroke. Patients with established diabetes and latent diabetes should be assessed to exclude vascular (both micro and macro) and neurological complications. Treatment We currently aim to keep the blood sugar less than 11 mmol/L (200 mg/dL) in the first few days after an acute

11.19 Nutritional problems

11.19.1 Malnutrition Malnutrition is a common and often unrecognized problem in patients admitted to hospital, especially the elderly.202 Inevitably, the reported frequency of malnutrition after stroke has varied depending on patient selection, the definitions of malnutrition and the method and timing of assessments. Estimates vary from 8% to 34%.203,204 It is not clear which factors are associated with malnutrition on admission, but the non-stroke literature suggests that malnutrition is more frequent in older patients, those living in institutions and poor social circ*mstances, and those with prior cognitive impairment, physical disability or gastrointestinal

575

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 576

576

Chapter 11 What are this patient’s problems?

disease. In a proportion of stroke patients, nutritional status may worsen during hospital admission.205,206 Like any acute illness, stroke may cause a negative energy balance and greater nutritional demands but, at the same time, stroke patients may be less able to adapt to these.207,208 They often have swallowing difficulties (section 11.17) and even those who can swallow may have a poor appetite because of lack of taste associated with the stroke itself209 or medication, intercurrent illness, depression, or apathy; and they may eat slowly because of facial weakness, lack of dentures or poor arm function. Poor nutrition has been associated with reduced muscle strength, reduced resistance to infection and impaired wound healing (although not specifically in stroke patients). Among patients with stroke, muscle weakness, infections and pressure ulcers are common and account for many deaths and much morbidity.105 Malnutrition on admission is associated with increased case fatality and poorer functional outcomes even having adjusted for other factors, but it is unclear to what extent this association is causal.210,211

11.19.2 Obesity Obesity is common among hospitalized stroke patients.36,212 As well as being a risk factor for stroke (section 6.6.10), and presumably recurrent stroke obesity can also be a problem during recovery from stroke. Where patients have restricted mobility, especially where they rely on others for help with transfers, obesity can be a crucial factor in how long they remain in hospital and how much support they require. It is also a problem in the long term in achieving adequate control of vascular risk factors such as hypertension (section 16.7.3) and diabetes. Patients quite often gain weight after stroke, presumably because of decreased energy expenditure and excessive calorie intake.

11.19.3 Assessment In routine clinical practice there are practical difficulties in assessing stroke patients’ nutritional status. A dietary and weight history may not be available because of patients’ communication problems and an alternative source of this information may not be available if, as is common, the patient lives alone. Simple assessments of weight and height to estimate the body mass index (BMI) pose problems in immobile stroke patients. Specialized equipment, of limited availability, such as weighing beds or hoists and scales which accommodate wheelchairs, may be required and height may have to be estimated from the patient’s demi-span or heel–knee

Fig. 11.19 A hoist which incorporates an electronic scale (inset) for weighing immobile stroke patients who are unable to stand or sit unaided.

length (Fig. 11.19). Laboratory parameters such as haemoglobin, serum protein, albumin and transferrin may not necessarily reflect nutritional status. More complex anthropometric measures, vitamin estimations, antigen skin testing and bioelectric impedance are all used in research but are not widely available. Awareness of the possibility of malnutrition is a key factor in identifying malnourished patients. It is probably worthwhile using a standardized screening tool for nutritional problems in the stroke unit which can highlight those patients who would benefit from a dietician’s input, but none have been adequately evaluated for stroke patients.213,214 A simple and informal end-of-thebed assessment (i.e. is the patient undernourished, normal or overweight?) is better than no assessment at all since it reliably identifies most stroke patients with a low BMI and abnormal anthropometry, as well as predicting outcome.210,215 Estimation of the BMI, serial weights to identify weight loss, and monitoring of dietary intake should be used to screen patients on admission and monitor them while on the stroke unit. Simple laboratory tests including a serum albumin may be worth monitoring where there is clinical evidence of poor or worsening nutrition. All patients admitted to hospital with stroke should have an early assessment of their nutritional status. In the absence of a formal assessment tool the team’s judgement of whether the patient is undernourished, normal or overweight is better than nothing. At least it focuses the team on this important aspect of care.

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 577

11.19 Nutrition

11.19.4 Prevention and treatment of malnutrition General approach It is useful to involve the speech and language therapist and dietician in the assessment and care of patients who have swallowing problems, or when nutritional intake may be inadequate for other reasons, e.g. confusion. Patients often eat very slowly after stroke and need supervision to ensure safe swallowing. Simple measures such as providing appetizing food of an appropriate consistency (section 11.17), placing the patient’s meal in their intact visual field, and ensuring that the patient has well-fitting dentures should not be overlooked. Labelling a patient’s dentures with their name while in hospital helps prevent losses and unfortunate mix-ups. Staff shortages may mean that patients receive insufficient food, which will eventually cause malnutrition, or have food forced into them hurriedly by unthinking staff, which is very demeaning and adversely affects morale. Where trained staff are unable to cope, the patient’s family or even volunteers can be easily trained to help with feeding. Dentures are often lost during hospital admission. Labelling a patient’s dentures with their name helps prevent losses and unfortunate mix-ups. Most people do not like to be fed; they prefer to feed themselves. By providing patients with suitable feeding utensils or foods which they can pick up may allow them to feed themselves (section 11.32.6). If not, adequate staff or family members, who have been instructed in how to do it, need to be employed. Oral nutritional supplements Routine oral nutritional supplementation, which provide both proteins and calories, in hospitalized stroke patients is probably not worthwhile since they have not been shown to have a clinically useful impact on outcome. The FOOD trial randomly allocated 4023 patients with an acute stroke to routine oral nutritional supplements or not, for the duration of their hospital admission. There was no significant difference in survival or functional outcomes overall, but there were insufficient patients enrolled who were undernourished to determine if this subgroup might benefit.212 Therefore, the results of the FOOD trial, taken along with a large number of much smaller single-centre RCTs of nutritional support in elderly patients, suggest that oral supplements are probably useful, but just in those who are identified as malnourished or who have inadequate food intake.212

Enteral tube feeding Where swallowing is impaired, early introduction of nasogastric tube feeding probably reduces the risk of death although maybe at the expense of keeping more patients with severe disability alive (Fig. 11.20).36 This also facilitates administration of important medications. Early tube feeding does not appear to influence the risk of chest infections but does seem to increase the risk of gastrointestinal haemorrhage (section 11.2.3) although it is unclear why this occurs and whether any preventive treatment such as a proton pump inhibitor would be effective (Table 11.26). Inserting nasogastric tubes is often difficult and ensuring that they are in the stomach, rather than in the lungs, at the beginning of each feed is not straightforward.216 Also, patients frequently find the nasogastric tube uncomfortable, and feeding is often interrupted by the patient repeatedly removing the tube. By restraining the patient (Fig. 11.21a,b) one can probably improve the continuity of nasogastric feeding but such restraints are not acceptable to some patients, their families and healthcare staff. Although enteral feeding via a percutaneous endoscopic gastrostomy (PEG) is perceived as more comfortable and reliable in providing the prescribed feed, its routine early use is not associated with better survival or functional outcomes.36 Indeed, the FOOD trial which included 321 patients allocated to initial tube feeding via nasogastric or PEG showed that patients allocated to PEG feeding tended to have worse outcomes (7.8% increased risk of death or severe disability (95% CI –0.08% to 15.5%). However, where a patient is likely to need prolonged tube feeding, or where nasogastric feeding is not practical, a PEG is usually the best option (Table 11.26). But one does have to take account of the associated complications. About one-fifth will develop aspiration pneumonia shortly after insertion, about 10% will develop a wound infection, and potentially life-threatening complications such as peritonitis and major bleeding occur in about 1%.217–222 Patients, or more frequently their relatives, need to be counselled about these risks and the real possibility that their frail relative may die in the days following the procedure.

The early introduction of PEG tube feeding, rather than persisting with a nasogastric tube, in dysphagic patients is unlikely to be associated with better outcomes. However, if nasogastric tube feeding becomes impractical there may be no alternative to PEG insertion.

577

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 578

578

Chapter 11 What are this patient’s problems?

Point estimate from FOOD trial Feed 100 patients early – prevent about six deaths

There was a 5.8% absolute (95% CI − 0.8 –12.5) reduction in the proportion of patients who had died at 6months follow-up in the early enteral tube feeding group compared with the avoid early feed group. The reduction in death or dependency was only 1.2% (95% CI − 4.2– 6.6).

Lower 95% confidence interval: feed 100 patients early and cause about one additional death

Upper 95% confidence interval: feed 100 patients early and prevent about 13 deaths

Fig. 11.20 Results from the FOOD trial comparing outcomes (deaths by 6 months per 100 patients treated) following enteral tube feeding (mostly via nasogastric tube) started early after hospital admission versus avoiding enteral tube feeding for at least 1 week.

11.19.5 Treatment of obesity

11.19.6 Other considerations

Patients who are obese, in particular if this causes problems with mobility or control of diabetes or blood pressure, should be encouraged to lose weight and be offered dietary advice (section 16.7).

It is important to remember that eating plays an important role in our social lives and is a source of pleasure. Eating with other people during the recovery phase of stroke encourages communication and social

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 579

11.20 Spasticity and contractures Table 11.26 Advantages and disadvantages of nasogastric and percutaneous endoscopic gastrostomy feeding.

Advantages

Disadvantages

Nasogastric tube

Widely available Cheap

Percutaneous endoscopic gastrostomy

Tube rarely displaced Cosmetically acceptable Long-term use practical

Easily inserted into lungs Often pulled out Unsightly Uncomfortable May lead to aspiration Nasal irritation/ulcer Gastric bleeding Interference with swallow Limited availability Expensive Aspiration when sedated for insertion Wound infection Bleeding Peritonitis

(a)

(b) Fig. 11.21 (a) A mitten used to prevent a patient from removing a nasogastric tube. (b) An American football helmet to prevent a patient from pulling out a nasogastric tube.

interaction. Patients who are concerned by their appearance after a stroke (e.g. due to facial weakness) may gain confidence from this opportunity to socialize. On the other hand, dribbling or ending up with the food down their shirt, or on the table or floor, may distress patients and inhibit social interaction. Advice from the occupational therapist regarding equipment to allow the patient to eat independently should be sought to maximize the positive, and minimize the negative, aspects of eating in a communal setting (section 11.32.6).

(Reproduced from Levine and Morris, 1995460 with permission.)

11.20 Spasticity and contractures

Spasticity has been defined as a motor disorder characterized by velocity-dependent increase in muscle tone with exaggerated tendon reflexes. It develops in between one-fifth and one-third of hospital-admitted stroke patients over the first year and is more common in those with hemiparesis and after severe strokes, and is more

579

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 580

580

Chapter 11 What are this patient’s problems?

common in the arm than the leg.223,224 However, spasticity contributes to disability in only a minority of patients with hemiplegia.224,225 Spasticity is usually accompanied by muscle weakness and clumsiness and sometimes by flexor or extensor spasms. Immediately after a stroke the muscular tone in the limbs and trunk may be lower, the same as, or higher than normal. The reasons for this variation are unclear. Although tone may be lower than normal in the acute phase, in most patients who do not recover it tends to increase over the first few weeks. In patients with a hemiparesis, the tone in the arm is usually greater in the flexors than extensors, while in the leg it is greater in the extensors than the flexors. This explains the typical hemiplegic posture (i.e. elbow, wrist and fingers flexed and arm adducted and internally rotated with the leg extended at the hip and knee and the foot plantar flexed and inverted). Tone in the truncal muscles may also be abnormally high or low. Tone may increase in any muscle group so much that it restricts the active movement which the residual muscle strength can produce. Imbalance in muscle tone can eventually result in shortening of muscles and permanent deformity and so restrict the full range of movement, i.e. contractures. Associated reactions are involuntary movements of the affected side (most typically flexion of the arm) elicited by a variety of stimuli including the use of unaffected limbs (e.g. self-propelling a wheelchair, yawning or coughing) and the upright posture.99,226 Associated reactions become more obvious with increases in tone. Associated reactions may be misinterpreted as voluntary movements by family and ill-informed staff who should be educated about their significance to prevent them becoming unduly optimistic about the patient’s motor recovery. Spasticity and contractures may cause pain, deformity, disability and, if severe, secondary complications such as pressure ulcers at points of contact between soft tissues (e.g. on the inner aspect of the knees in a patient with contractures of the thigh adductors). Associated reactions may be misinterpreted as voluntary movements by family and ill-informed staff who should be educated about their significance to prevent them becoming unduly optimistic about the patient’s motor recovery. Assessment Tone and spasticity: Physicians are trained to assess the tone in a limb by asking the patient to relax (which is almost guaranteed to have the opposite effect) and then moving the limb through its range of movement at each

Table 11.27 The modified Ashworth scale for the clinical assessment of muscle tone.438 0 1

No increase in muscle tone Slight increase in muscle tone, manifested by a catch and release, or by a minimal resistance at the end of the range of motion when the affected part(s) is moved in flexion or extension 1+ Slight increase in muscle tone, manifested by a catch, followed by minimal resistance throughout the remainder (less than half) of the range of movement 2 More marked increase in muscle tone through most of the range of movement but affected parts easily moved 3 Considerable increase in muscle tone, passive movement difficult 4 Affected parts rigid in flexion or extension

joint at different speeds and noting any resistance to these movements. Unfortunately, many physicians do not appreciate how much tone is influenced by factors such as the patient’s position, anxiety, fatigue, pain and medication. Tone may change from minute to minute. This makes it difficult to assess objectively with good inter-observer reliability. Physiotherapists, who spend far more time than physicians handling patients, are more aware of changes in tone and try to take advantage of them in their therapy. Formal measurement of tone can be attempted using clinical scales (e.g. the modified Ashworth Scale, Table 11.27) or techniques such as electrogoniometry or quantitative neurophysiology. Unfortunately, the last two are not widely available or practical in routine clinical practice. The inter-observer reliability of the modified Ashworth Scale is generally good for assessing tone in the arm and around the knee but poor for assessing tone around the ankle.227 When assessing the effectiveness of treatment in an individual patient, or in a group of patients in RCTs, it may be better to measure function (e.g. walking speed or dressing), or the achievement of a specific goal (to allow the palm of the hand to be accessed to maintain hygiene) than relying on measurement of tone itself. Indeed, the management of abnormal tone, like all other aspects of rehabilitation, should be directed at achieving realistic, relevant and measurable goals (section 10.3.3). Contractures are easier to assess because, by definition, the deformity is fixed. Thus, one can objectively measure the range of movement around a joint and repeat the measure to determine whether or not an intervention has improved the range. However, occasionally an apparent contracture responds to injection of botulinum toxin (see below) indicating that the shortening is due to

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 581

11.20 Spasticity and contractures

muscle contraction rather than permanent shortening of muscles or tendons. Prevention and treatment In rehabilitation our aim is to modulate changes in muscle tone to the patient’s advantage. For instance, to increase tone in a flaccid leg and so provide the patient with a more secure base on which to walk, or reduce tone in the arm to facilitate more active movement. We use several complementary approaches to prevent the development of unwanted patterns of tone and to alleviate existing problems due to spasticity and contractures. Some are applicable to any patient while others are only required to deal with exceptional problems: • Avoidance of exacerbating factors: Pain (section 11.23), urinary retention (section 11.14), severe constipation (section 11.15), skin irritation, pressure ulcers (section 11.16), anxiety (11.31.1) and any other unpleasant stimulus may cause an unwanted increase in tone. These factors must be avoided or alleviated. • Positioning and seating: Poor positioning, especially in immobile patients, can lead to detrimental changes in tone. For example, long periods spent lying supine will increase extensor spasms, presumably by facilitation of basic reflexes. Regular positioning, such as external rotation of the shoulder, appears to reduce unwanted tonal changes.228 Unfortunately, there seems to be quite a lot of uncertainty about the optimum positions for patients with hemiplegia and there are also practical difficulties in keeping them in the required position, i.e. patients tend to move.229 The positioning charts which can be found on most stroke units (Fig. 11.22) can only be used as a general guide. Finding the optimum position for a patient is often a matter of trial and error and relies on the experience of the therapists and nurses. Appropriate seating which can be tailored to the individual patient is essential (Fig. 11.23). Ideally, patients should be seated in a balanced, symmetrical and stable position which they find comfortable and which enables them to function, e.g. eat, drink, etc. • Passive movements and physiotherapy: It is important that muscles are not allowed to remain in a relaxed and shortened position for too long. For example, if the arm is left in a flexed posture at the elbow, or the foot plantar flexed in bed, this can lead to permanent shortening and contractures. Patients’ limbs should be moved passively to stretch the muscles and maintain the range of movement, even in the very acute stages. However, it is important that carers are taught to do this without damaging vulnerable structures such as the shoulder (section 11.24). Also, handling methods, during transfers for example, can influence tone, at

least in the short term.230 Physiotherapists have an important role in teaching nurses and informal carers the correct positioning and handling techniques to minimize risk of injury (to patient and handler) and unwanted spasticity. Patients with a hemiparesis often attempt to ‘overuse’ their sound side to achieve mobility but, as a consequence, the tone may increase in the affected side. Using the unaffected leg to self-propel a wheelchair is said to lead to increased spasticity in the affected arm and leg.231 The impact of any changes in tone associated with such activities on functional recovery has not been established.232 The aim of many of the facilitation and inhibition techniques used by therapists is to use basic reflexes and postural changes in tone to promote function. However, these techniques, although widely accepted and practised, have not been evaluated adequately in RCTs (section 11.21). Various physical techniques including the application of cold, of heat, splinting and electrical stimulation can, at least for a short time, reduce spasticity.230,233,234 This may relieve discomfort, allow improved hygiene, and plaster casts or splints to be applied. Whether these physical techniques have direct longer-term benefits is unclear:233 • Splinting and casting: Occasionally these may be necessary to prevent or treat contractures. Progressive splinting and application of casts can improve range of movement but the optimum duration and best methods are unclear. Also, badly fitted casts and splints can cause pain, pressure ulcers and tendon damage which may exacerbate rather than relieve spasticity and contractures. • Oral antispastic drugs: Where spasticity is not adequately controlled by physical techniques, or where the patient is suffering from painful muscle spasms, certain drugs have been advocated: baclofen, dantrolene, tizanidine and diazepam which all reduce tone by altering neurotransmitter function or ion flux in the spinal cord, or in muscles. But there is little reliable evidence from RCTs to indicate that they usefully reduce spasticity or improve function in stroke patients.235 Indeed, in our experience they rarely make much difference and they have a number of adverse effects (Table 11.28). Adverse effects are much more common in older patients but can be minimized by starting with low doses and increasing the dose slowly until the desired effect is achieved, or adverse effects necessitate withdrawal. We normally only use baclofen or tizanidine for a trial period in patients who our physiotherapists feel might benefit. • Local and regional treatments: Injection of botulinum toxin directly into muscles can reduce troublesome spasticity. Small RCTs have confirmed that this does

581

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 582

582

Chapter 11 What are this patient’s problems?

Lying on affected side No backrest. One or two pillows for head. Affected shoulder pulled well forward. Place good leg forward on a pillow. Pillow placed behind back.

Sitting up Sitting well back and in centre of chair. Affected arm placed well forward on table or pillow. Feet flat on floor. Knees directly above the feet.

Lying on unaffected side No backrest. One or two pillows for head. Affected shoulder forward with arm on pillow. Place affected leg backward on a pillow. Pillow placed behind back.

Sitting in bed Sitting in bed is not desirable. Affected arm placed on pillow. Legs are straight. Sitting upright and well supported.

indeed reduce tone, which may reduce clawing of the fingers and allow better skin hygiene.236–238 This treatment appears to be safe, at least in the short term, but because its effects wear off over several months, repeated injections may be required and treatment costs can be very high. Sometimes, a single injection can allow simpler measures to be introduced which avoids the need for further injections. Further large trials are in progress to establish whether function really is improved (www.strokecentre.org). Very occasionally, and usually where simple measures (see above) have been inadequate, spasticity can be so troublesome as to warrant more invasive procedures: • local nerve blocks with ethanol or phenol although occasionally useful to solve specifc problems, can

Fig. 11.22 The typical chart used to guide the positioning of stroke patients with hemiplegia (shown in black) whilst in bed.

lead to unwanted muscle weakness and painful dysaesthesias;239 • intrathecal infusion of baclofen using implantable pumps;240 • surgical procedures such as anterior and posterior rhizotomy and more recently lesioning of the dorsal root entry zone (so-called drez-otomy);233 and • tendon lengthening and transfers which can, for example, help to reduce equinovarus deformity.233 Anecdotally, the prevalence of severe contractures appears to have declined dramatically over the last 30 years which is probably a result of improvements in the standards of general care. We now virtually never have to resort to the more invasive procedures described here.

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 583

11.21 Limb weakness, poor truncal control and unsteady gait

changes in tone, is associated with the development of a painful shoulder (section 11.24) and swelling of the hand (section 11.25). Poor hand and arm function is a major cause of dependency in activities of daily living. Weakness of the leg, which affects about 45% of patients (Table 11.2), may be severe enough to immobilize the patient and thus predispose to the complications of immobility (section 11.11). Leg weakness, making it difficult to stand, transfer or walk independently, is one of the most important factors prolonging hospital stay in stroke patients. In patients with hemiparesis, which affects about 55% of patients, the arm is usually weaker than the leg (Table 11.2). Assessment

Fig. 11.23 A chair which can be tailored to the individual needs of the patient. This model can be raised, or lowered, the back rest and seat can be adjusted, arms can be altered and extra supports can be inserted.

11.21 Limb weakness, poor truncal control and unsteady gait

These three aspects of the patient’s condition are impossible to separate and will therefore be discussed together. Weakness of an arm, leg or both, sometimes with unilateral facial weakness, is probably the most common and widely recognized impairment caused by stroke. However, there are often associated but less obvious problems with the axial muscles which impair truncal control and walking. Facial weakness, which affects about 40% of patients (Table 11.2), apart from its cosmetic effects, may contribute to dysarthria (section 11.30.2) and cause problems with the oral preparatory phase of swallowing (section 11.17). Weakness of the upper limb, which affects about 50% of patients (Table 11.2), along with

In assessing function in a patient with stroke, unlike making the diagnosis, it is more useful to observe the range and control of voluntary movements of the limbs than assessing power in individual muscle groups. For instance, do the patients have only coarse control of movement around the hip or shoulder, or have they retained movement at the more distal joints? In stroke patients, distal movements are usually more severely impaired than proximal ones. The assessment of motor function and truncal control has already been described (section 3.3.4). Recovery from a hemiplegic stroke has been likened to early infant development, in that the recovery of truncal control follows the same general pattern as that of a growing child. Head control returns first, followed by rolling over, sitting balance and then standing balance, and lastly the patient can walk with increasing steadiness and speed (Fig. 10.16). After a stroke it is useful to know where the patient is on this ‘developmental ladder’ when assessing prognosis and setting goals for rehabilitation.241 It is also important to assess truncal control and gait since truncal ataxia can occur without limb incoordination in patients with midline cerebellar lesions. Indeed, it is not unknown for patients to undergo full gastrointestinal and metabolic investigation to elucidate the cause of vomiting before their truncal ataxia is noted and a cerebellar stroke diagnosed. It also seems absurd that, although immobility is the main reason for a stroke patient needing to stay in hospital for rehabilitation, mobility and balance are very often not assessed properly by doctors admitting stroke patients.59 Having stressed the importance of testing truncal control and gait, it is important that in doing so neither the patient nor the doctor are put at risk of injury. Poor handling and lifting technique may dislocate a patient’s flaccid shoulder (section 11.24), result in a fracture from a fall (section 11.26) and may even injure the doctor’s

583

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 584

584

Chapter 11 What are this patient’s problems? Table 11.28 Adverse effects of antispasticity drugs. Adverse effect

Baclofen

Diazepam

Dantrolene

Tizanidine

Sedation/central nervous system depression Confusion Hypotonia/weakness Unsteadiness/ataxia Exacerbation of epileptic seizures Psychosis/hallucinations Insomnia Headache Urinary retention Dry mouth Hypotension Nausea/vomiting Diarrhoea/constipation Abnormal liver function Hyperglycaemia Visual disturbance Skin rashes Pericarditis/pleural effusions Blood dyscrasia Withdrawal symptoms Drug interactions

++ + + + ++ ++ + + +

+++ + + ++

++ +

+ +

+ + + + + + +

+ + + +

+ ++

+ + + +

+ + + ++ +

+ ++

+ +++

+ + + +

+ +

+ reported occasionally; ++ quite frequent; +++ potentially fatal.

back. Physiotherapists should ideally provide appropriate training to all staff, and informal carers, who are involved in handling patients. The severity of weakness of individual muscle groups is often graded with the Medical Research Council (MRC) Scale242 (Table 11.29). This was originally designed to assess motor weakness arising from injuries to single peripheral nerves, not stroke. Unfortunately, although the MRC Scale has good inter-observer reliability if applied rigorously,227 it is often misused and the optional expansion to include extra grades (4 plus and 5 minus) makes it even less reliable in the routine recording of motor weakness. The motricity index, a modification of the MRC Scale (Table 11.29), for use in patients with stroke, allows the observer to grade the severity of the hemiparesis rather than each separate muscle group.243 This can be useful in charting patients’ progress for research purposes but is of limited value in routine clinical practice because it is difficult to remember the weights applied to individual movements and it requires a small block to assess grip strength. Several other tools are available for objectively measuring and recording motor function (Table 11.30).

Treatment The amount of spontaneous recovery of motor function is highly variable. The more severe the initial impairment, the less likely is full recovery. One study suggested that motor and sensory function 5 days after stroke onset explained 74% of the variance in motor function at 6 months with the Fugl-Meyer Scale.244 The pattern of recovery of motor function parallels that of other strokerelated deficits, with the most rapid recovery occurring in the first few weeks and then the pace of improvement slows over subsequent months (section 10.2.5). In patients with hemiparesis it is generally thought that motor function in the leg improves more than that in the arm, although this has been questioned.245 Also, unless the patient has some return of grip within 1 month of the stroke, useful return of function is unlikely, although not impossible.246 Physiotherapy is the main therapeutic option in hemiparesis although techniques vary. The two broad approaches most commonly employed are the ‘facilitation and inhibition’ technique and the ‘functional’ approach.

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 585

11.21 Limb weakness, poor truncal control and unsteady gait Table 11.29 The MRC Scale of Weakness, and the motricity index which was developed from the MRC Scale for use in stroke patients.242,243 MRC Scale 0 No contraction 1 Flicker or trace of contraction 2 Active movement with gravity eliminated 3 Active movement against gravity 4 Active movement against resistance 5 Full strength The motricity index Arm Pinch grip Elbow flexion (from 90°) Shoulder abduction (from chest) Total arm score Leg Ankle dorsiflexion (from plantar flexed) Knee extension (from 90°) Hip flexion Total leg score Scoring system for pinch grip 0 Pinch grip, no movement 11 Beginnings of prehension 19 Grips block (not against gravity) 22 Grips block against gravity 26 Against pull but weak 33 Normal Scoring system for movements other than pinch grip 0 No movement 9 Palpable contraction only 14 Movement but not against gravity/limited range 19 Movement against gravity/full range 25 Weaker than other side 33 Normal

— — — — — — — —

Table 11.30 Measurements of motor function.439 Impairments MRC Scale (Table 11.29) Motricity index (Table 11.29) Trunk control test Motor club assessment Rivermead motor assessment Dynamometry Disability Arm Nine-hole peg test Frenchay arm test/battery Action research arm test Truncal control/mobility Standing balance Functional ambulation category Timed 10-m walk Truncal control Rivermead mobility index Subsection of Office of Population Censuses and Surveys (OPCS) scale440

Table 11.31 Common features of facilitation/inhibition-based therapy.249 Recognition of the intimate relationship between sensation and movement Recognition of the importance of basic reflex activity Use of sensory input and different postures to facilitate or inhibit reflex activity and movement Motor relearning based on repetition of activity and frequency of stimulation Treatment of the body as an integrated unit rather than focusing on one part Close personal interaction between the therapist and patient

• The facilitation and inhibition technique is based on the premise that posture and sensory stimuli can modify basic reflex patterns which emerge after cerebral damage. Several workers have developed different treatments based on this concept, the best-known being those of Bobath247 and Brunnstrom.248 Although these techniques differ, certain features are common to all249 (Table 11.31): to achieve as normal a posture and pattern of movement on the affected side as possible. • On the other hand, the functional approach simply aims, through training and strengthening of the unaffected side, to compensate for the impairment to achieve maximum function. For example, patients may be encouraged to transfer and walk as soon as possible after the stroke. Supporters of the facilitation and inhibition approach claim that, although the functional approach might achieve earlier independence, it results in more abnormal patterns of tone and movement which in the long term may lead to contractures and loss of function. Vigorous activity involving the unaffected side may increase the tone in the affected limbs during the activity. This does not seem to occur with all activities (e.g. pedalling) and any long-term effects on tone are unclear.250 We believe there is a place for both approaches in clinical practice. Quite often, where a patient has not achieved useful independent mobility despite a prolonged period of physiotherapy (with the facilitation and inhibition approach), we switch to a functional approach to maximize the patient’s autonomy. For example, we will train patients to transfer independently and self-propel a wheelchair even though this may be associated with, at least in the short term, unwanted changes in tone. There has been very little formal evaluation of the physiotherapy techniques. Although several small RCTs

585

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 586

586

Chapter 11 What are this patient’s problems?

have been reported, no definite conclusions about their relative merits can be drawn.251 In any case, comparisons of different techniques may have limited relevance to current clinical practice as many therapists adopt an eclectic approach, using selected aspects of each technique where appropriate for individual patients. There are, therefore, some important questions about physiotherapy after stroke which need to be answered in properly designed RCTs: • When should physiotherapy start ? • How long should it continue? • What is the optimum intensity of physiotherapy?252 • Which specific therapeutic interventions are the most effective? • Is therapy provided by relatively unskilled therapists as effective as that provided by skilled therapists? • Which patients gain most from physiotherapy and can we prospectively identify them? The results of small RCTs support the hypothesis that physiotherapy, especially focused on achieving particular tasks, improves function even when started late after stroke.253 The trials generally indicate that therapy has a greater impact on specific motor impairments than the resulting disability. This may be because the resulting disabilities are the consequence of sensory and cognitive as well as motor problems. The size of any treatment effect is probably influenced by the intensity of treatment.252 However, many older, sicker patients may not be able to tolerate intensive regimes, which emphasizes the need for research to identify the optimum physiotherapy regime for particular subgroups of patients.254 Given the methodological difficulties in systematically reviewing and performing RCTs of physiotherapy techniques (and those of other therapists), this is a daunting challenge (Table 11.1). Other interventions A large number of physical techniques have been developed with the aim of improving motor function or gait. Some have been evaluated in small RCTs which have included highly selected patients and focused more on impairment than disability as outcome measures. None are supported by enough evidence to recommend their routine use.253 These techniques include: • electromyographic, visual and auditory feedback;255–257 • functional electrical stimulation, which is effective as an orthosis; when applied it can reduce foot drop to facilitate gait, and probably increases muscle strength, but it is unclear whether it improves functional outcome and many patients find it uncomfortable;253,258 • acupuncture259 and transcutaneous electrical nerve stimulation (TENS);260

• treadmill gait retraining with or without bodyweight support;261 • ‘forced use’ or ‘constraint induced therapy’ where the unaffected arm is immobilized for a major part of the day and during physiotherapy sessions;262 • drugs to enhance motor recovery.263,264 Other approaches In this section we have dealt with interventions that aim to decrease disability by improving impairments. The complementary strategy of providing appropriate mobility aids, e.g. walking sticks, splints and wheelchairs is dealt with in section 11.32.1.

11.22 Sensory impairments

In about one-fifth of patients with acute stroke it is impossible to assess sensation adequately because of reduced conscious level, confusion or communication problems, but about one-third of the remainder have impairment of at least one sensory modality (Table 11.2). Sensory problems are more easily identified in patients with right rather than left hemisphere stroke, probably because they have fewer communication difficulties. Severe sensory loss may be as disabling as paralysis, especially when it affects proprioception. Furthermore, loss of pain and temperature sensation in a limb, or sensory loss with neglect, may put a patient at risk of injury from hot water, etc. And disordered sensation with numbness or paraesthesia, even without functional difficulties may, if persistent, be as distressing to some patients as central post-stroke pain (section 11.23). We have discussed some of the difficulties in assessing sensory function earlier (section 3.3.5). Patients often complain bitterly about what appears to the doctor to be a minor change in sensation. Do not underestimate the effect which facial numbness, or tingling in a hand, can have on the morale of a patient. Little is known specifically about the recovery of sensation after stroke, although it probably follows a similar pattern seen in most other impairments (section 10.2.5). However, sensory symptoms may evolve and even become more distressing with time and they may worsen during intercurrent illness (e.g. infections) leading to concerns about recurrent stroke. Under

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 587

11.23 Pain (excluding headache)

these circ*mstances it is important to give appropriate explanation and reassurance to the patient and any carer. Although patients may be given sensory stimulation as part of their therapy, the effect this or any other intervention has on sensation is unknown.265 Where patients have lost temperature or pain sensation in a limb, especially if there is associated neglect, it is important to counsel them about commonsense strategies to avoid injury to the limb.

11.23 Pain (excluding headache)

Pain is a common complaint among stroke patients. Perhaps one-third require analgesia for pain (excluding shoulder pain) during hospital admission after an acute stroke and, although it becomes less of a problem with time, persisting severe pain may affect about one-fifth of patients.105,266 There are many potential causes, some of which are coincidental and others which are in some way due to the stroke (Table 11.32). Usually, the cause becomes obvious after one has asked the patient about the distribution, nature and onset of the pain and has examined the relevant area. However, some pains (e.g. due to spasticity, axial arthritis and central post-stroke pain) may be difficult for patients to describe and localize. Diagnosis and assessment of analgesic requirements are particularly difficult in patients with communication and cognitive problems.

Table 11.32 Causes of pain after stroke. Section Headache due to vascular pathology Painful shoulder Deep venous thrombosis and pulmonary embolism Pressure ulcers Limb spasticity Fractures Arterial occlusion with ischaemia of limb, bowel or myocardium Coexisting arthritis exacerbated by immobility or therapy Instrumentation, e.g. catheter, intravenous cannulae, nasogastric tube Central post-stroke pain (thalamic pain/ Dejerine–Roussy syndrome)

3.7.1, 11.9 11.24 11.13 11.16 11.20 11.26

11.23

The treatment of pain depends on the cause. Simple analgesics (e.g. paracetamol) along with some reassurance that the pain does not indicate any serious problem may be all that is needed. Other interventions such as local application of heat or cold, and transcutaneous electrical nerve stimulation (TENS) or acupuncture, may relieve symptoms with a low risk of adverse effects. Pain due to spasticity should initially be treated by alleviating exacerbating factors and by carefully positioning the patient (section 11.20). Antispasticity drugs are occasionally required, seldom work and have significant adverse effects (Table 11.28). Musculoskeletal pain can be treated with simple analgesics and, if these are ineffective and there are no contraindications, a non-steroidal anti-inflammatory drug. If a joint becomes acutely painful it may be necessary to rest it and investigate to exclude more serious causes, e.g. septic arthritis, fracture, or gout exacerbated by diuretics or aspirin. We find it useful to discuss the patient’s pain at the multidisciplinary team meeting where one can establish the pattern, severity and control of pain in different settings, e.g. on the ward, at night, during therapy sessions. This provides important clues to its cause and the best approach to treatment. Central post-stroke pain This is variably described as a superficial burning, lacerating or pricking sensation often exacerbated by factors including touch, movement, cold and anxiety.267 It usually affects one-half of the body but may be more localized, affecting a quadrant, one limb or just the face. It affects up to 10% of hospitalized patients surviving at 6 months in total, and 5% severely268 though this estimate is certainly higher than we would expect in our clinical practice. A far less common, but related, problem is post-stroke pruritus.269 Central post-stroke pain is usually associated with some abnormality of pinprick or temperature sensation and may be accompanied by autonomic changes, for example sweating or cold.270 It may start immediately after the stroke but more frequently after a delay of weeks or months.267 Although the term ‘thalamic pain’ is commonly used synonymously with central poststroke pain, this is misleading since the pain occurs in patients with stroke lesions affecting any part of the sensory pathways.271 Central post-stroke pain is often resistant to therapy. Avoidance of those factors which exacerbate the symptoms is an important first step. Tricyclic antidepressants (e.g. amitryptiline) may alleviate the symptoms and also lift any associated depression.272 A small dose should be used initially, usually at bedtime, and be increased

587

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 588

588

Chapter 11 What are this patient’s problems?

slowly until adequate symptom control is achieved or adverse effects become troublesome. If tricyclic antidepressants are ineffective, addition of lamotrigine or gabapentin are reasonable second-line treatments. A range of alternative drugs including other anti-epileptics (e.g. carbamazepine, phenytoin, valproate, clonazepam), anti-arrhythmics (e.g. mexiletine), anaesthetics (e.g. ketamine), opiates and intrathecal baclofen have all been advocated but none adequately evaluated in placebo-controlled randomized trials.272–274 Physical methods such as acupuncture and TENS are worth trying since they may occasionally provide relief and are at least fairly safe without any lasting adverse effects. Psychological interventions may help but have not been formally evaluated. More invasive and destructive techniques such as stereotactic mesencephalic tractotomy, or deep-brain stimulation, are occasionally used in severe cases which are resistant to other treatment modalities, but are not necessarily effective.

11.24 Painful shoulder

Shoulder pain is reported by at least one-fifth of patients in the first 6 months after stroke. Its frequency increases over the first few months. It is more common in those with severe sensorimotor deficits, and thus hospitaladmitted patients, and usually affects the shoulder on the hemiparetic side.275–277 Although many factors (Table 11.33) have been associated with painful shoulder, their role in its development is unclear.277 A small proportion of patients who complain of a painful shoulder after a stroke have the other clinical features which comprise the syndrome of reflex sympathetic dystrophy or shoulder–hand syndrome (Table 11.34). It is unclear whether this represents a distinct entity or simply the severe end of a spectrum. Our ignorance of the causes and prognosis of shoulder pain is in part due to major problems of definition and the lack of well-validated and reliable assessment tools.278 But, although our understanding of the epidemiology and causes of painful shoulder is incomplete, no one involved in stroke rehabilitation can doubt its importance. It causes patients great discomfort, it may seriously affect morale, and can inhibit recovery. In some patients it persists for months, even years.

Table 11.33 Factors associated with painful shoulder after stroke. Section Associated features Shoulder pain before the stroke Low tone allowing glenohumeral subluxation/malalignment Spasticity Severe weakness of the arm Sensory loss Neglect Visual field deficits Neurological mechanisms Reflex sympathetic dystrophy (shoulder–hand syndrome) Central post-stroke pain Brachial plexus injury Orthopaedic problems Adhesive capsulitis (frozen shoulder) Rotator cuff tears due to improper handling or positioning Acromioclavicular arthritis Glenohumeral arthritis Subdeltoid tendinitis

Fig. 11.25 11.20 11.21 11.22 11.28 11.27.1 11.24 11.23

Table 11.34 Features of reflex sympathetic dystrophy. Pain and tenderness on abduction, flexion and external rotation of the arm at the shoulder Pain and swelling over the carpal bones Swelling of metacarpophalangeal and proximal interphalangeal joints Changes in temperature, colour and dryness of the skin of the hand Loss of dorsal hand skin creases, and nail changes Osteoporosis Note: there is probably considerable overlap between reflex sympathetic dystrophy and the cold arms described by Wanklyn286,441 (section 11.25).

Prevention and treatment Treatment of an established painful shoulder is often ineffective so that any measures to prevent its development are important. It is probably useful to introduce policies on the stroke unit which have been identified as being associated with a lower frequency of the problem in non-randomized studies (Table 11.35). It may also be useful to identify individuals who are at particularly high risk, based on the factors in Table 11.33, and make all staff aware of the potential problem. More specific

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 589

11.24 Painful shoulder

(a)

(b)

Fig. 11.24 By supporting the weight of the arm, glenohumeral subluxation can be reduced. This may be achieved when the patient is sitting using an arm support (a) which attaches to the chair or wheelchair or, alternatively, a perspex tray (b) which,

because it is transparent, allows the patient to check on the position of their feet. Both are better than pillows which invariably end up on the floor. Several designs of sling (c) are available to reduce subluxation when patients are upright.

Table 11.35 General measures which may reduce the frequency of painful shoulder after stroke.

unproven.279–281 Many therapists worry that the use of slings and cuffs will inhibit recovery of the arm because it is held in a position which promotes spasticity. When a patient complains of shoulder pain it is important to exclude glenohumeral dislocation (Fig. 11.25), fracture or specific shoulder syndromes. For example, painful arc (supraspinatus tendinitis) may respond better to specific measures (e.g. local steroid injection) although even the evidence supporting treatments for these specific syndromes is poor.282,283 In established painful shoulder many treatments have been suggested, some have been evaluated in small RCTs, but further studies are needed to define the best treatments (Table 11.36). Some interventions are probably harmless (e.g. application of cold, heat, taping, TENS) and if they produce even short-term relief are worth trying. Others have potentially important adverse effects and costs (e.g. oral corticosteroids) and therefore need to be evaluated further before being adopted into routine clinical practice.277

Instruct all staff and carers to: Support the flaccid arm to reduce subluxation Teach patients not to allow the affected arm to hang unsupported when sitting or standing. While sitting they might use one of several arm supports which attach to the chair or wheelchair. All are more effective than a pillow which spends most of the time on the floor. Shoulder/arm orthoses may, depending on their design, prevent subluxation but have not been shown to reduce the frequency of painful shoulder (Fig. 11.24c). Avoid pulling on the affected arm when handling the patient Staff and carers should be trained in methods of handling and lifting patients to avoid traction injuries. Avoid any activity which causes shoulder discomfort Therapy sessions sometimes do more harm than good. Maintain range of passive shoulder movements

interventions including slings, cuffs and taping to support the flaccid arm (Fig. 11.24), and functional electrical stimulation may reduce subluxation but their effect on the frequency of painful shoulder and arm function is

(c)

Shoulder pain after stroke is common, ill understood, difficult to prevent and none of the suggested treatments are supported by reliable evidence.

589

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 590

590

Chapter 11 What are this patient’s problems? Table 11.36 Treatments used for painful shoulders.

Fig. 11.25 X-ray of shoulder showing glenohumeral subluxation, a common finding in stroke patients, but does it matter? Note the widened joint space (double-headed arrow) and the increased distance between the lower border of the glenoid cavity (short arrow) and the lower border of the humeral head (long arrow). Photograph provided by Dr Allan Stephenson.

11.25 Swollen and cold limbs

Swelling with pitting oedema, and sometimes pain, quite often occurs in the paralysed or neglected hand, arm or leg, usually within the first few weeks.284,285 The swelling may limit the movement of the affected part and the pain not only further restricts movement but also exacerbates spasticity and associated reactions (section 11.20). Some patients complain of coldness of the limb, more often of the arm than the leg.286 Swelling more often occurs in the legs of patients who sit for prolonged periods (section 11.13). Gravity and lack of muscle contraction, which reduce venous and lymphatic return, presumably play a part but autonomic changes and control of regional blood flow may be relevant.287 An isolated painful, swollen hand may be a mild form of the shoulder–hand syndrome. There are a number of other causes which need to be considered (Table 11.37). People with fractures or acute ischaemia of the limb usually complain of severe pain, but stroke patients with sensory loss, visuospatial dysfunction or communication

Physiotherapy Positioning and mobilization Exercises Heat or cold Support Strapping Shoulder/arm orthoses (Fig. 11.24c) Bobath sling Rood shoulder support Arm supports for bed or chairs Lapboard (Fig. 11.24b) Forearm support Wheelchair outrigger (Fig. 11.24a) Medication Systemic Analgesics Non-steroidal anti-inflammatory drugs Corticosteroids442 Antispastic drugs (Table 11.28) Phenoxybenzamine Antidepressants Local Corticosteroid injection of shoulder Local anaesthetic Botulinum toxin to periarticular muscles Stellate ganglion block Other physical Ultrasound Acupuncture Biofeedback Transcutaneous electrical nerve stimulation (TENS) Surgery Sympathectomy Humeral head suspension Relief of contractures

Table 11.37 Causes of swollen limb after stroke. Section Gravity in a dependent limb Lack of muscle contraction Deep venous thrombosis Compression of veins or lymphatics by tumour, etc. Cardiac failure Hypoalbuminaemia Occult injury Acute ischaemia Gout Reflex sympathetic dystrophy

11.13

11.19 11.26

11.24 (Table 11.34)

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 591

11.26 Falls and fractures

difficulties may not, which can lead to these diagnoses being overlooked.

Table 11.38 Factors likely to contribute to falls after stroke. Section

Assessment The other causes of a swollen limb should be excluded by clinical examination and appropriate investigation before attributing the swelling simply to immobility or dependency. Investigation is not usually necessary but Doppler ultrasound or venography to exclude deep venous thrombosis of the leg (section 11.13), a simple X-ray of a swollen wrist or ankle to exclude a fracture, and a serum albumin may be useful. One can monitor the effect of any intervention by simply measuring the circumference of the limb although plethysmography has been used in research studies.288 Treatment The treatment obviously depends on the cause. Where the swelling appears to be due to immobility we try the following: • elevation of the affected limb when at rest; • encourage active movement (this may be impossible with severe weakness but is important if neglect is the main cause); • graduated compression stocking on the leg (full length or below-knee depending on extent of swelling) or a bandage on the arm, although the latter may worsen swelling of the hand; and • intermittent compression of the limb although this was shown to be ineffective in reducing arm swelling in a randomized trial.289 In addition, where the limb is painful, simple analgesia may ameliorate the secondary effects of pain on tone which leads to spasticity and contractures. Diuretics should be avoided unless there is evidence of heart failure, since immobile stroke patients have a tendency to become dehydrated (section 11.18.1) and incontinent (section 11.14).

11.26 Falls and fractures

Falls are common after stroke. About one-third of patients are expected to fall during their in-hospital care.105,290–292 Falls are also frequent following discharge from hospital.293,294 Many factors contribute to this tendency; some are listed in Table 11.38. Combinations of such factors are associated with higher risks, though

The patient Muscle weakness (especially of 11.21 quadriceps) Sensory loss (especially visual 11.22 & 11.27 impairments) Impaired balance, righting reflexes 11.21 and ataxia Confusion 11.29 Visuospatial neglect, e.g. denial of 11.28 hemiparesis Deformity, e.g. plantarflexion causing 11.20 toe catching Epileptic seizures 11.8 Postural hypotension due to drugs or 11.2.3 & 11.18.1 dehydration The environment Inappropriate footwear, e.g. slippers Slippery floors, deep pile carpets and loose rugs Excess furniture Poorly positioned rails and inappropriate aids Lack of supervision Fire doors (these may close automatically and hit slow-moving stroke patients) Drugs: Sedatives and hypnotics Hypotensive drugs Antispastic drugs Table 11.28 Anti-epileptic drugs

patients with severe post-stroke impairments are often at lower risk because they are unable to mobilize at all.294 Several ‘falls risk scores’ have been developed which incorporate these (and other) factors but none have yet been shown to be practical and sufficiently accurate to gain widespread use.291,295,296 A small proportion of falls occurring in hospital ( 70 years 16.6 s.399 Treatment Where the patient has residual problems, a physiotherapist or occupational therapist can improve mobility.400 Training aimed at improving cardiorespiratory function probably increases walking speed.401 Patients’ mobility

quite often deteriorates, after a period of stability, in the months or years after a stroke. This may be attributed to further strokes, progression of coexisting pathology (e.g. arthritis) or attenuation of the benefits of regular physiotherapy. It is important that whoever is responsible for monitoring the patient’s progress, most often the general (family) practitioner, is alert to this and refers the patient back to the physiotherapist. Although not backed up by RCTs, we have little doubt of the value of so-called ‘top-up’ physiotherapy in patients such as this, although its effectiveness will inevitably depend on the cause of any deterioration. Therapy in the patient’s own home may be more helpful than in a hospital outpatient department since the therapist can then deal with the real, everyday problems that the patient is experiencing. For example, patients can be taught to climb their own stairs and to overcome the particular problems associated with the layout

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 605

11.32 Dependency in activities of daily living Table 11.47 Walking aids.451 Shoes Should be comfortable, supportive and have non-slip soles Occasionally, a smooth sole facilitates a smooth swing phase in patients with foot drop when walking on carpets Walking sticks May improve standing stability and walking speed but must be tailored to the patient (i.e. length, handle shape, weight and appearance)452,453 but not evidence based;454 a long stick can be used to encourage patients to put weight through their affected leg (Fig. 11.26d) Replace worn rubber ferrule (i.e. rubber bit on end of stick) Should be held in the unaffected hand in patients with hemiparesis but this means the patient’s functioning hand is no longer available for other tasks Other uses include extending the patient’s reach, and ‘self-defence’ Tripods and quadrapods Provide a broader base of support than a walking stick Their weight may cause associated reactions and may worsen the quality of gait452 Not useful on uneven ground Walking frames (Fig. 11.26) Available in many sizes and shapes with or without wheels Useful where balance is poor, especially if the patient tends to lean backwards Often give patients confidence Needs good upper limb function and, of course, prevents patient from using hands or carrying things; many patients attach a basket or use a wheeled trolley to overcome this problem Patients often try to pull up on their frame to rise from a chair, which can cause them to fall Frames may trip up patients and be difficult to manoeuvre with lots of furniture and thick carpets Cannot be used on stairs, therefore patients may require several frames, one on each level Beware bent frames, protruding screws, loose hand-grips and worn ferrules Other uses, drying underclothes Ankle–foot orthosis or brace (Fig. 11.27) Usually of thermoplastic construction and tailored to the individual Occasionally useful in spastic foot drop to improve gait pattern,455 but does not improve spasticity456 and may actually exacerbate the problem Beware in patients with oedema and tendency to leg ulcers Knee brace Occasionally required to prevent hyperextension of the knee Functional electrical stimulation An experimental treatment which has been used to correct foot drop after stroke (section 11.21) Hand rails Short grab-rails can be useful to provide support, especially at thresholds and doorways Full-length rails on staircases are invaluable to those with poor balance, confidence or vision Advice on positioning hand rails should be obtained from an occupational therapist; poorly placed hand rails can cause accidents or may not be used

of their own home. Two RCTs which compared the effectiveness of home vs hospital-based physiotherapy showed that improvement in ADL function was slightly greater in those treated at home, but other outcomes were no different.402 Although the therapist may be able to reduce the patient’s degree of impairment, even when intervention starts very late after a stroke, the therapist’s main tool at this point is the provision of mobility aids (Tables 11.47 and 11.48). When one prescribes an aid for patients, it is important to provide them and any carers with training in its use, and to ensure that it is maintained in good

order (Figs 11.26–11.36). Many patients do not use the aids provided, which is both a waste of resources and an indication of the lack of value of that aid. It is important to follow up patients who have been given aids to ensure they fit the patient’s needs and are actually being used, and to retrieve any unused equipment. If one prescribes an aid, one should teach the patient or carer how to use it, ensure that it solves the problem, check that it is maintained in good working order, and provide follow-up to ensure its continued use or appropriate removal.

605

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 606

606

Chapter 11 What are this patient’s problems? Table 11.48 Other aids to mobility. Chairs (Fig. 11.28) Higher seats (but not so high as to cause the feet to dangle) without too much backward angle, and firm arm rests (of the correct height, shape and length) facilitate transfers It is easier to stand from a chair if the patient can place his/her heels under it when preparing to stand Chairs should be tailored to the individual’s requirements, but existing chairs may be raised or lowered Chairs on castors can be dangerous Ejector chairs (Fig. 11.29) The seat or chair may be sprung to provide an initial lift to allow the patient to rise These must be tailored to the individual’s weight to avoid the patient being catapulted across the room Electric chairs which slowly bring the patient to a semi-erect position are occasionally useful but are expensive, bulky and require training to use properly Beds A firm mattress at the correct height will facilitate transfers from a bed Blocks or leg extensions can be used to modify an existing bed A grab-rail attached to the bed frame or floor can facilitate independent transfers (Fig. 11.30a) It may be impossible to attach aids to a divan-style bed Mechanical devices which help the patient to sit in bed can occasionally be helpful (Fig. 11.30b) Wheelchairs An invaluable means to improve mobility Many hundreds of designs including self-propelled, carer-propelled and electric chairs Need to be tailored to individual’s requirements Must be kept in good working order (Fig. 11.31) Ramps Useful to those in wheelchairs and their carers Many designs available for different situations Lifts Many types, including: Domestic stair-lifts (Fig. 11.32) for those who cannot manage stairs Through-the-floor lifts for accessing upper floors in wheelchairs Short-rise lifts for accessing vehicles in wheelchairs or where there is insufficient space for a ramp Lifts are expensive and may require structural alterations to accommodate them

Patients who are unable to walk outside, or who can walk only short distances, or who have difficulty using their own or public transport, may be helped by putting them in contact with any local transport schemes. For example, in the UK special financial allowances are available to help with the added cost of being immobile, specially adapted taxis can be provided and some shopping centres provide electric wheelchairs. In the last few years there has been a major campaign to improve access to public places for patients with disabilities.

11.32.2 Driving Many people who have had a stroke never return to driving.403,404 There are several reasons: severe residual motor, sensory, visual or cognitive impairments are the most common. In many countries the authorities place restrictions on driving after stroke, and particularly if the stroke is complicated by epileptic seizures (section 11.8). The variation in regulations probably reflects the dearth

of relevant research in this area. Special, and usually more stringent, regulations apply to those who drive commercial vehicles, taxis and heavy goods vehicles. The main issue for driving authorities is not so much risk of recurrence – well under the 20% per annum for normal drivers in the UK – but physical disability and cognitive difficulties which can be so difficult to judge. Patients are often not asked by their doctors whether they drive, thus many continue to drive and put lives at risk in contravention of their local regulations.404–406 It is the doctor’s responsibility to ask patients whether or not they drive so they can be given relevant information. Failure to do so might have serious consequences, although it is unclear whether patients who have had a stroke and return to driving are at increased risk of road traffic accidents.407 There appears to be little agreement about the optimal method of assessing a disabled patient’s fitness to drive. Informal judgements are often made by patients or their family doctors.404 Bedside testing of neurological

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 607

11.32 Dependency in activities of daily living

Fig. 11.26 Walking aids. (a) A typical light-weight walking frame of adjustable height which may be useful if a patient has poor balance but reasonable arm and hand function. (b) A walking frame with wheels which may be better than (a) in patients who tend to fall backwards or who have Parkinson’s disease. (c) A tea trolley which will provide patients with support as well as allow them to take things from place to place. (d) A long walking stick held in the unaffected hand can be used to encourage the patient to put more of their weight through their affected leg.

(a)

(b)

(c)

(d)

impairments including cognitive function, assessments in driving simulators and road tests have all been advocated.408 One might start with a bedside assessment to demonstrate any physical, and probably more importantly visual, visuospatial and cognitive impairments, which would make safe driving very unlikely.409,410

Patients who pass on the bedside testing could then be assessed on a driving simulator or an ‘off-road’ test which would identify those who are clearly unsafe to drive.411 The remainder could then be given an on-road test in a dual-control car with an appropriately trained instructor,412,413 which can be performed with good inter-rater

607

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 608

608

Chapter 11 What are this patient’s problems?

(a)

(b)

(a)

Fig. 11.27 Ankle supports. (a) A thermoplastic ankle–foot orthosis: the famous AFO splint. Although this is very useful in patients with foot drop secondary to lower motor neurone lesions, in stroke patients it can sometimes increase the tendency to plantarflexion by stimulating the sole of the foot. (b) A lateral ankle support may be useful in patients who have a tendency to invert their foot when walking.

(b) Fig. 11.28 Choosing an appropriate chair for a stroke patient who has difficulties getting from sitting to standing. (a) A bad chair to ensure the user never escapes! The seat is too low and slopes backwards, the arms are soft and offer little resistance to allow the person to push up on them, the base is solid so the patient cannot tuck their feet under them, and this chair is on castors so that if the patient eventually manages to stand the chair slides away and causes them to fall backwards. Note the Velcro pads on the arms to allow a tray to be fixed across the user’s escape route. (b) A good chair to facilitate easy transfers. This chair is of reasonable height, is upright, has firm but padded arms and allows the user to tuck their feet underneath them which makes it easier to get their weight over their feet. This chair does not have wings, which may provide support for the user’s head when sleeping but can stop interaction with other people sitting to either side. Fig. 11.29 An ejector chair. The number of springs (arrow) can be altered depending on the weight of the user. Some patients find that this type of chair pushes them off balance.

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 609

11.32 Dependency in activities of daily living

(a) Fig. 11.30 Aids to getting out of bed. Although many people can get out of a bed independently during a therapy session when they are ‘warmed up’, they may need aids to get out in the middle of the night or first thing in the morning. (a) A grabrail attached to the bed can facilitate transfers from lying to

(b) standing and may therefore prevent incontinence at night. The height of this model can be adjusted to suit the individual. (b) A pneumatic mattress elevator can be helpful in getting patients from lying to sitting.

Fig. 11.31 (opposite) Things to look for when examining a wheelchair. This wheelchair has many hazardous and uncomfortable features. Note the white sticky tape covering the name of the hospital on the side of this wheelchair, which wished, very reasonably, to remain anonymous. 1: Check that the tyres are properly inflated (arrow). Flat tyres make the wheelchair hard work to propel and the brakes inoperative. 2: Check the brakes work (arrow). 3: Check that the sides and arms (arrow) are removable to facilitate transfers onto toilets, into cars, etc. 4: Check that the foot plates are not fixed (arrow) or flapping about. Either problem may cause injury. The foot plates should fold neatly and securely out of the way when not in use. 5: Check that the arms (arrow) are properly padded to avoid discomfort and compression neuropathy of the ulnar nerve. 6: Check that the back (arrow) and seat (arrow) are not sagging and that the appropriate cushion is being used to reduce the risk of pressure ulcers. 7: Check that the grips on the handles are in good order (arrow).

reliability. This stepwise approach should minimize the risk of injury to the patient, instructor and other road users. Many countries provide specialist centres for the assessment of disabled drivers. These centres not only assess driving skills, but also provide advice on vehicle adaptations which allow patients with severe physical disability to drive. Training on a driving simulator has been evaluated in a small RCT, and while definite

conclusions are limited by methodological issues, this approach is promising.414

11.32.3 Toileting The problems of urinary and faecal incontinence have been discussed in sections 11.14 and 11.15, respectively. About 10% of surviving patients are still dependent in

609

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 610

610

Chapter 11 What are this patient’s problems? Table 11.49 Toileting problems and solutions. Cannot get to toilet because of mobility or access problems Solutions: Urinal (Fig. 11.10) Commode, many designs to suit individual needs (Fig. 11.33) Chemical toilet, useful where regular emptying of a commode is not possible Cannot transfer on and off the toilet, or poor balance making manipulation of clothes difficult Solutions: Grab-rails Toilet frame (Fig. 11.33) Raised or adjustable toilet seat (Fig. 11.33) Cannot clean myself Solutions: Toilet paper holder on the unaffected side which can be used with one hand Large aperture toilet seat Self-cleaning toilet, like a bidet and toilet in one

11.32.4 Washing and bathing

Fig. 11.32 A typical stair-lift for domestic use. The arm-rests fold out of the way to facilitate transfers. The foot-rest and seat fold up when not in use. The controls are on the arm-rests.

toileting 1 year after their first stroke due to inability to transfer independently, walk, or dress and undress (Table 11.45). For the cognitively intact patient, this is an embarrassing disability which severely damages their self-esteem. An assessment by an occupation therapist should define the severity and cause of the problem. The patient’s ability to toilet themselves is very dependent on the environment so that a home assessment may be particularly useful. Simple factors such as the width of the door to the toilet or bathroom, the position and height of the toilet, and the position of the toilet roll holder can make a crucial difference to whether or not the patient can use the toilet independently. Therapy aimed at improving performance in mobility, transfers and dressing will all facilitate independence in toileting. Table 11.49 lists some common problems and simple solutions, and Fig. 11.33 shows some simple toilet aids.

About 10% and 30% are dependent in washing and bathing respectively 1 year after a first-ever-in-a-lifetime stroke (Table 11.45). Motor, sensory, visuospatial and cognitive impairments all contribute to these disabilities. Although poor arm function makes washing and grooming more difficult, most patients can perform these tasks with their unaffected arm, but those with visuospatial and cognitive deficits, even when arm function seems quite good, may still be unable to wash and groom themselves independently. Independence in bathing obviously requires some independence in mobility and transfers. Other disabling conditions in the elderly (e.g. painful arthritis) often contribute. A skilled assessment by an occupational therapist, preferably in the patient’s own home, will delineate the problems. This assessment ensures there is adequate access to the bathroom (e.g. for a wheelchair). The size and layout of the bathroom will determine which aids can be used. Therapy aimed at improving those impairments, especially motor impairments, which are contributing to the disability in bathing and washing may be of some help but, again, at this stage the provision of the wide range of available aids to allow the patient to compensate for their impairments is more effective (Fig. 11.34 and Table 11.50). Specific training in the use of bathing aids, both in hospital and in the patient’s home, does appear to increase use of the bath.415 Of course, in many countries patients prefer to take showers rather than a bath, which generally reduces the difficulty. One general point is that a patient who has trouble bathing should ensure that somebody else is in the house and that the bathroom door is left unlocked, so that if they fall or are unable to get out of the bath somebody can easily reach them.

9781405127660_4_011.qxd 10/13/07 10:55 AM Page 611

11.32 Dependency in activities of daily living

(a)

(b)

Fig. 11.33 Simple toilet aids. (a) A raised toilet seat with a frame. (b) A commode; this one attaches securely to the bed (arrow) for use at night. Table 11.50 Problems with grooming or bathing/showering and solutions. Cannot brush dentures with one hand Solutions: Half fill sink with water, attach brush with a suction pad to the sink and brush dentures with functioning hand Cannot get into, and even more important, out of the bath Solutions: Grab-rails appropriately mounted around bath Non-slip bath mat Bath stool Bath board Inflatable or hydraulic bath seats which can lower patients into or raise them out of the bath (Fig. 11.34) Hoists Replace bath with shower Fig. 11.34 A pneumatic bath aid to lower patients into and to lift them out of a bath. However, the humble bath rail (arrow) with some other simple aids is very commonly all that is needed to allow the person to use a bath independently.

Patients who have difficulties bathing independently should not be left in their house alone while bathing. Some form of alarm or even a cordless or mobile telephone will increase safety.

Some bathroom adaptations are expensive and cause considerable disruption to the patie