Algorithms in Differential Diagnosis How to Approach Common Presenting Complaints in Adult Patients, for Medical Students and Junior Doctors Dr. Nigel Fong Singapore General Hospital, Singapore Ve World Scientific NEW JERSEY - LONDON - SINGAPORE - BEIJING - SHANGHAI - HONG KONG - TAIPEI + CHENNAI - TOKYOPublished by World Scientific Publishing Co. Pte. Ltd. 5 Toh Tuck Link, Singapore 596224 USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601 UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE Library of Congress Cataloging-in-Publication Data ‘Names: Fong, Nigel (Nigel Jie Ming), author Title: Algorithms in differential diagnosis : how to approach common presenting complaints in adult patients, for medical students and junior doctors / Dr. Nigel Fong. Description: New Jersey : World Scientific, 2019. | Includes index. Identifiers: LCCN 2018046261 ISBN 9789813232921 (hardcover : alk. paper) | ISBN 9813232927 (hardcover : alk. paper) | ISBN 9789811200557 (pbk. : alk. paper) | ISBN 9811200556 (pbk. : alk. papes) Subjects: | MESH: Diagnosis, Differential | Algorithms | Adult | Case Reports Classification: LCC RC71 | NLM WB 141.5 | DDC 616.07/5—de23 LC record available at hitps://lecn.loc.gow/2018046261 British Library Cataloguing-in-Publication Data ‘A catalogue record for this book is available from the British Library. Cover design by Jonathan Yong-Emn Lim Copyright © 2019 by World Scientific Publishing Co. Pte. Ltd All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means, electronic ormechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the publisher. For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher. Disclaimer This book is intended only for a professional audience, and should be used in conjunction with other reference texts and clinical experience. Accurate diagnosis and safe clinical practice is the physician's responsibility. Readers are urged to keep up to date with changes in medical knowledge and best practice. Additionally, junior physicians are advised to consult senior staff if unsure how to manage a patient, While the Author and Publisher have made every effort to ensure that the content of this book is accurate and up to date, neither the Author nor Publisher assumes, liability for any error, omission, or harm resulting from the material in this book. For any available supplementary material, please visit bttps:/;www.worldscientific.com/worldscibooks!1 0.1 142/10787#t~suppl Printed in SingaporePreface This is a book for medical students and first-year doctors who wish to learn how to approach a patient's symptoms, and sharpen their skills of clinical reasoning and diagnosis. Clinical medicine begins with the patient. Many of us learn one disease condition at a time, yet, patients present with symptoms and not a known disease. Many students know the correct evidence-based treatment of heart failure, yet are stumped when given a breathless patient on the ward, and told to ‘figure out what is wrong’. Many can list the differentials for a symptom, but struggle to separate important diagnostic fea- tures from the irrelevant details that many patients throw at us. It is tempting to rely ever more heavily on our growing armamentarium of diagnostic tools. But testing without thinking not only confounds and misleads, it also costs our patients dear. Having been a student myself not too long ago, I have experienced first-hand the struggles that a budding clinician faces in synthesising vast amounts of new informa- tion and applying it to real patients. In this book, I try to offer a toolkit to tackle this challenge. Each chapter tackles one presenting complaint, identifying key differentials, pro- viding a strategy to distinguish each differential from the other and setting out the thought process behind history, examination and initial investigations. Each approach cuts across different specialties, integrating approaches and conditions from various medical and surgical fields. This book uses algorithms to aid diagnosis. This is a method of clinical reasoning that uses critical pieces of information as branch points to distinguish between groups of diagnoses. It complements (not replaces) clinical skills and knowledge of individual disease conditions. Junior diagnosticians particularly benefit from learning algorithms, because they not only provide a systematic and functional way to approach patients, but also serve as a scaffold to organise knowledge and learn the skills of clinical rea- soning. I have written these algorithms to be usable for those who are just starting out, rather than theoretically complete but too complex to use. Finally, always remember that if one hopes to develop clinical acumen, there is no substitute to seeing patients—they are our best teachers, inspiration and sources of creative inquiry. Nigel Fong 2018vii Subspecialty Reviewers CARDIOLOGY DERMATOLOGY ENDOCRINOLOGY GASTROENTEROLOGY GERIATRIC MEDICINE GyNaEcoLocy HAEMATOLOGY INFECTIOUS DISEASE NEPHROLOGY NEUROLOGY Dr Wang Yue Senior Resident, Cardiology National Heart Centre, Singapore Dr Oh Choon Chiat Associate Consultant, Dermatology Singapore General Hospital Dr Tan Zhen-Wei, Matthew Dr Matthew Tan Diabetes & Endocrine Care Farrer Park Medical Centre, Singapore Dr Ong Ming Liang, Andrew Associate Consultant, Gastroenterology & Hepatology Singapore General Hospital Dr Lim Kim Hwa Jim Senior Consultant, Geriatric Medicine Changi General Hospital, Singapore A/Prof Tan Lay Kok Senior Consultant, Obstetrics & Gynaecology Singapore General Hospital Dr Dixon Grant Associate Consultant, Haematology Singapore General Hospital Dr Benjamin Cherng Pei Zhi Consultant, Infectious Diseases Singapore General Hospital Dr Jason Choo Chon Jun Senior Consultant & Program Director, Renal Medicine Singapore General Hospital Dr Wee Chee Keong Consultant, Neurology, National Neuroscience Institute Singapore Dr Gosavi Tushar Divakar Consultant, Neurology, National Neuroscience Institute Singaporeviii ALGORITHMS IN DIFFERENTIAL DIAGNOSIS OPHTHALMOLOGY OrrHopaepic SURGERY OTOLARYNGOLOGY PsycuiaTRY RESPIRATORY MEDICINE RHEUMATOLOGY SURGERY EDUCATION ADVISORS Dr Loo Jing Liang Senior Consultant Singapore National Eye Centre Dr Jerry Delphi Chen Yonggiang Senior Resident, Orthopaedic Surgery Singapore Health Services Dr Toh Song Tar Senior Consultant, Otolaryngology Singapore General Hospital A/Prof Ng Beng Yeong Ng Beng Yeong Psych Med Clinic Mount Elizabeth Medical Centre, Singapore Dr Ong Thun How Senior Consultant & Program Director Respiratory & Critical Care Medicine Singapore General Hospital A/Prof Koh Siyue, Mariko Senior Consultant, Respiratory & Critical Care Medicine Singapore General Hospital Dr Fong Weng Seng, Warren Consultant, Rheumatology and Immunology Singapore General Hospital Dr Lee Lip Seng Consultant, General Surgery Changi General Hospital, Singapore A/Prof Tan Choon Kiat, Nigel Senior Consultant, Neurology National Neuroscience Institute, Singapore Deputy Group Director, Education (Undergrad) Singapore Health Services A/Prof Tay Sook Muay Senior Consultant, Anaesthesia ‘Singapore General Hospital Associate Dean, Yong Loo Lin School of Medicine National University of SingaporeAcknowledgements Ipraise God for his work in my life, which I scarcely deserve. He has blessed me greatly, and His grace sustains me each day. Each day in the wards, I marvel at how each of us are the work of His fingers, yet am reminded that to Him we will one day return. To Him be all glory. I thank Marianne, my wife and closest companion, for being a wonderful blessing and support. [ am also very grateful to my parents; to whom I owe so much. 1 am indebted to the many senior clinicians from the SingHealth family (listed on page vii) who have taken personal time to review my writing. They have provided invaluable subspecialty input, refined the algorithms and corrected numerous errors and omissions. I am particularly thankful for Dr Ong Thun How and A/Prof Tay Sook ‘Muay, both of whom first encouraged me to write this book, and who have mentored me in many ways. Many peers and juniors have also contributed to this book (and its previous versions) as reviewers: Dr Amanda Chin Dr Hester Lau Dr Mohd Alimi Dr A. Sayampanathan Dr How Guo Yuan Dr Nicholas Ngiam Dr Benjamin Nah Dr Huang Wenjie Dr Ng Qin Xiang Dr Calida Chua DrJerold Loh Dr Peng Siyu Dr Chang Zi Yun Dr Jiang Bochao Dr R. Chockelingam Jr Dr Chaung Jia Quan Dr Joel Goh Dr Ryan Lee Dr Chng Wei Qiang Dr Kera-Anne Tan Dr Samuel Ee Dr Chua Si Min. Dr Ke Yuhe Dr Sharon Goh Dr Daniel Lim Dr Kevin Kwek Dr Stephen Sumarli Dr Darius Pan Dr Khoo Bo Yan Dr Tan Pei Zheng Dr David Zhao Dr Krystal Koh Dr Tang Wanchu Dr Davin Ryanputra Dr Lee Bing Howe Dr Tony Li Dr Dorothy Huang Dr Leong Jia Li Dr Tryphena Ng Dr Edwin Chow Dr Leong Yun Hao Dr Wang Hankun Dr Eugene Gan Dr Liang Sai Dr Yii Zheng Wei Dr Gabriel Tan Dr Marianne Tsang Dr Zhao Yang Dr Gob Jiaying Dr Michael Chee Dr Graham Goh Dr Michelle Sim I must thank my teachers. Over the years, I have had the benefit of many excellent clinical tutors, who have not only imparted medical knowledge, but more importantly, taught me how to think, and how to care for patients. I am also greatly indebted to my alma mater, the Yong Loo Lin School of Medicine, National University of Singapore.x ALGORITHMS IN DIFFERENTIAL DIAGNOSIS ‘Most of all, I thank my patients, who time and again have proven to be my best tutors, and who have given medicine its meaning. Finally, I thank the publishing team at World Scientific for putting this book together. My editor, Lim Sook Cheng, has been particularly approachable and helpful. T also thank Jonathan Lim, a long-time friend and artist, who provided the cover illustration.Contents Preface Subspecialty Reviewers Acknowledgements Introduction 1 Learning How to Diagnose Heart and Lungs 2 An Approach to Hypotension An Approach to Chest Pain ‘An Approach to Dyspnoea and Hypoxaemia ‘An Approach to Palpitations and Arrhythmias An Approach to Hypertension. An Approach to Cough od ane w An Approach to Haemoptysis, Gut and Abdomen 9 An Approach to Abdominal Pain 10 An Approach to Jaundice and Abnormal Liver Function Tests 11 An Approach to Gastrointestinal Bleeding 12 An Approach to Nausea and Vomiting 13 An Approach to Diarrhoea 14 An Approach to Constipation 15 An Approach to Dysphagia Kidneys and Urinary Tract 16 An Approach to Acid-Base Disorders 17 An Approach to Electrolyte Imbalances 18 An Approach to Elevated Creatinine and Oliguria 17 19 27 39 57 65 70 77 83 85 102 112 120 128 135 139 145 147 156 177xii ALGORITHMS IN DIFFERENTIAL DIAGNOSIS 19 An Approach to Polyuria 20 An Approach to Haematuria and Proteinuria 21 An Approach to Difficulty Urinating Brain, Nerves and Senses 22 An Approach to Headache 23 An Approach to Blackouts: Syncope and Seizure 24 An Approach to Giddiness 25 An Approach to Weakness 26 An Approach to Cranial Nerve Palsy 27 An Approach to Red Eye and Loss of Vision 28 An Approach to Hearing Loss 29 An Approach to Ataxia 30 An Approach to Abnormal Movement 31 An Approach to Altered Mental State and Cognitive Decline Blood 32 An Approach to Anaemia 33 An Approach to Abnormal Bleeding 34 An Approach to High Cell Counts Endocrine and General Physiological Disturbances 35 An Approach to Fever 36 An Approach to Thyroid Abnormalities 37 An Approach to Hyper- and Hypoglycaemia 38 An Approach to Obesity and the Cushing's Syndrome 39 An Approach to Falls in the Elderly Skin and Subcutaneous Tissues 40 An Approach to Rash 41 An Approach to Skin Growths and Lumps 42 An Approach to Groin Lumps 43 An Approach to Breast Lumps and Complaints 44 An Approach to Neck Lumps 45 An Approach to Lymphadenopathy 185 189 201 209 21 220 232 239 259 269 276 282 287 293 303 305 320 328 339 341 351 360 366 371 377 379 392 397 402 406 411Contents xiii 46 An Approach to Lower Limb Pain and Ulcers 416 47 An Approach to Lower Limb Swelling 423 Joints and Muscles 429 48 An Approach to Joint Pain (General) 431 49 An Approach to Shoulder Pain 444 50 An Approach to Knee Pain 449 51 An Approach to Back and Neck Pain 454 Female Genital Tract 163 52 An Approach to Vaginal Bleeding (Non-Pregnant) 465 53 An Approach to Vaginal Bleeding (Pregnant) 4a 54 An Approach to Amenorrhoea and Vi 475 Index to Conditions 483 Erratum to Algorithms in Differential Diagnosis can be accessed at the link: https://www.worldscientific.com/worldscibooks/10.1142/10787#t=supplIntroductionChapter Learning How to 1 Diagnose Clinical Case A final-year medical student has difficulty making clinical diagnoses. Although she is able to regurgitate a great deal about many conditions (having read the textbooks cover-to- cover), she has great difficulty pinpointing the correct diagnosis whenever she sees a patient on the ward. She says that she ‘does not know where to begin’, and ‘just can’t think of the right diagnosis. What would you advise her? ‘Welcome! We must begin by learning how to diagnose. These skills are central to a physician's healing art—making an accurate and timely diagnosis allows the physician to administer appropriate treatment, which gets the patient better. Conversely, diagnostic errors harm patients. This chapter opens with an exposition of how clinicians make diagnoses, and dis- cusses the process of diagnostic reasoning, along with some other issues. Next, I offer some advice on how to use this book, and suggestions on how to learn the science and art of clinical diagnosis. We will use the clinical scenario (Box 1.1) as a starting point for discussion. Box 1.1. Diagnostic Reasoning Dt 19 Ward Rounds Third-year student: Our next patient is a 70-year-old gentleman who comoleins of jaun- dice. Itstarted 1 month ago, and has been getting worse. He does not have fever, abdominal pain, nausea or vorniting. He has been feeling lethargic and has also lost 10 kg in the last 3 months. His bowel movements are normal. He is a non-smoker, and has no travel or con- tact history. A sexual history is unremarkable. He has no past medical history. On examina- tion, his abdomen is soft and there is no palpable liver. Liver function tests show elevated bilirubin, AST, ALTand ALP. Consultant: What do you thinkit could be? Third-year student: Hmm... does he have cirrhosis? He has jaundice and abnormal liver, enzymes. Hepatitis? Haemolysis? Gallstones? Final-year student: I've clerked him too. He has no stigmata of chronic liver disease, so cir rhosis is unlikely. | don’t think he has haemolysis either, as his haemoglobin was normal. He has no abdominal pain, history of eating raw shelifish, unprotected sexual Intercourse or hepatotoxic drug ingestion, so it's probably not hepatitis. His liver function test is actually (Cont'd)4 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS showing higher ALP than AST and ALT, suggesting biliary obstruction, maybe by gallstones or amass. Resident: Yes, this is a gentleman with obstructive jaundice. The causes of obstructive jaundice can be divided into painful and painless. He is painless, which makes gallstones less likely, and extrinsic compression of the bile ducts more likely. | would order an ultra- sound next, looking for dilated bile ducts. If ducts are dilated, then we need to find out if there is biliary tree compression by an extrinsic mass, or if he has a stricture. If ducts are not dilated, then we should send off antimitochondrial antibody for primary biliary cholangitis. Consultant: Alright. This is an elderly gentleman with painless progressive obstructive Jaundice. You are right that in the absence of pai duct compression is more likely than gallstones. Strictures and primary biliary cholangitis are possible, but in this patient with dlinically significant weight loss, am most worried about a pancreatic tumour with bile duct compression. Lets skip ultrasound and go straight to CT. That will tell us whether there is any pancreatic mass. How Do Cli EME Ts ery Observe how each person in the ward round (Box 1.1) employed a different method of dinical problem solving. The third-year student was guessing, and had little chance of arriving at the correct diagnosis, except by fluke. The final-year student was testing one hypothesis at a time until he could find one to match the available information (hypothetical-deductive reasoning). The resident was using an algorithm, which divided the causes of jaundice based on key predictors (e.g., pre-hepatic, hepatic and obstructive jaundice; painful and painless). The expert clinician effortlessly recognised this pattern of symptoms, and synthesised his intuition with analytical decision rules (painful vs. painless). Understanding the diagnostic thought process is fruitful for all who wish to learn clinical diagnosis. Hypothetical-Deductive Reasoning Hypothetical-deductive reasoning is about generating and evaluating differential diag- noses. Every medical consult is an iterative process of information gathering, data interpretation, hypothesis generation and testing. Hypotheses so generated prompt a search for additional information (further history, examination or investigation); in turn, addition information may confirm or refute existing hypotheses, or trigger the generation of new hypotheses. (a) Hypothesis generation: Do not simply take a history and examine the patient. Think! Think of: (i) what are the most likely diagnoses, and (ii) what are the most dangerous/important diagnoses that must not be missed (even if they are less likely). (b) Evaluating differentials: Critically examine the differentials generated (Table 1.1). For every hypothesis, ask: (i) what information fits? (ii) what does not fit? (iii) what did I expect but could not find? Then rank the diagnoses in order of likelihood. Next, ask: (iv) what additional information do | need to confirm or refute thisLearning How to Diagnose Table 1.1. Hypothetical-Deductive Reasoning in a 60-year-old Presenting with Shortness of Breath and Wheeze Pa pring DO ceriad Sed Potengi PCRS enc en One tor eee foi Asthma Acute-onset Unusual to have | Thereisno Is there an atopic wheeze new-onset asthma | precipitating history? at 60years old trigger cord Smoker Should be chronic, Is the chest X-ray not acute hyperinflated? Heartfeilure | Pedal oedema Hehas no known | Are there q-waves can cause wheeze heart disease on his ECG? hypothesis? This prompts a search for additional information (whether further history, examination or investigation), which would further allow one to distinguish between the possible differentials. Howto use: The zbility to critically examine a list of differentials, rank them according to likelihood, and pick up ‘what does not fit’ is very important. Drawing this table for every patient seen is a useful exercise in clinical reasoning. Additionally, having a list of differentials guides physical examination (you can only find what you look for) and investigations. Caveats: Generating differentials based simply on ‘whet I can think of’ is a potential source of error, as the correct diagnosis may never be considered. Conversely, going through every possible differential for a certain symptom is inefficient and impractical. Therefore, while the hypothetical-deductive model is a helpful tool to evaluate differ- entials and identify ‘what does not fit, it is best to augment this approach with a method to consistently generate the most important differentials. Algorithms employ differentiating pieces of information (history, physical signs or investigations) as key branch points to distinguish between groups of diagnoses. The clinical consult begins by identifying the appropriate algorithm to use. Thereafter, information gathering is guided by the algorithm (Figure 1.1), with particular emphasis on deciding between diagnostic groups as made explicit by the branch points. After several branching points, only a small number of possible differentials are left, and hypothetical-deductive reasoning may be used to rank the remaining options. How to use: Algorithms systematically and rapidly identify likely differentials, helping to reduce cognitive load and the chance of not considering the correct diagnosis. They reflect an organised knowledge structure. This book is about helping you develop algorithms and organise your knowledge, so that you may be able to use algorithms in your diagnostic process.6 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS JAUNDICE y y Pre-hepatic Hepatic Post-hepatic Acute | Chronic) | Painful | Painless Figure 1.1. A simple algorithm for jaundice. See Chapter 10 for complete algorithm. Caveats: Algorithms are best used in conjunction with the critical lens of hypothetical- deductive reasoning. Algorithms identify the most likely diagnoses; the most dan- gerous diagnoses (that have to be excluded) may need to be considered separately via 2 hypothetical-deductive approach, even if deemed less likely by the algorithm. Allofus can instantaneously recognise people we know, without conscious or effortful thought. In the same way, clinicians may instantly recognise a patient's diagnosis,’ with little deliberate thought. This is an intuitive process involving matching the new patient's problem against similar ones solved previously. The ability to do so requires accumulation of ‘illness scripts’ (previous patients seen and mental prototypes) through clinical experience, as well as a well-organised knowledge structure so that correct matches can be retrieved rapidly based on salient cues. How to use: To use pattern recognition, you must have accumulated a database of illness scripts through clinical experience, and organised your knowledge structure via deliberate reflection on patients seen. Furthermore, illness scripts are condition- specific (having seen thousands of fractures does not make one better able to diagnose chest pain) and context-specific (having seen 10 sick pneumonia inpatients does not help one diagnose and manage community-acquired pneumonia in a general practice setting). The accuracy of pattern recognition depends on the clinician's breadth and volume of condition-specific illness scripts; simply ‘seeing many cases’ is insufficient— these cases must encompass the variety of conditions encountered, and accurately represent the range of ways in which each condition presents. This is why developing pattern recognition takes time, and it is not something that can be taught in a book. Caveats: Errors occur when the clinician is inexperienced, over-confident, distracted or fatigued. Patients who present atypically may be misdiagnosed, for example, acute coronary syndromes are more likely to be missed when the patient lacks chest pain. ‘Referred to as ‘pattern recognition’, ‘system one thinking’ or ‘non-analytical approach’ in literature.Learning How to Diagnose 7 There is good evidence that combining intuitive pattern-recognition with analytical strategies (algorithms and hypothetical-deductive reasoning) improves diagnostic outcomes. Most clinicians with some experience employ pattern recognition as default, as it is faster and less effortful. But when faced with an atypical patient, or novel clinical scenario, one is unfamiliar with, itis wise to abandon pattern recognition approaches in favour of conscious analytical thinking. Even when pattern recognition is used, safe diagnosis requires constant vigilance for information which ‘does not fit’ in which case analytical reasoning must take over. The Process of Diagnostic Reasoning Having understood the mental frameworks clinicians employ in making a diagnosis, this section delves further into the key elements of the diagnostic reasoning process: () information gathering, (ii) interpretation of the information to formulate a problem representation, lii) making a diagnosis, (iv) searching for aetiologies and complications and (v) dealing with patients who have multiple issues. Murua) Some students find that after an initial phase of allowing the patient to talk freely, they are not sure what questions to ask, or default to a memorised list of questions for each symptom. It is helpful to form an early impression of the patient's presenting complaint, and have some possible diagnoses in mind. Search for further information (history taking, examination and investigations) that would discriminate between the differen- tials. This is where an algorithm guides information gathering by specifying clinically important pieces of information that differentiate between diagnostic groups (ie., the branch points). It is a useful maxim that ‘the eye does not see what the mind does not think’—dlinically significant details often have to be explicitly solicited, or may be missed. Information is interpreted as it is gathered; this is an iterative process culminating in the promotion of one differential as the ‘most likely’ and the relegation of others (Table 1.1). (b) Problem Representation ‘Another early cognitive step is the creation of a one-sentence summary defining the case in specific terms. This may be explicitly articulated, or subconscious. Going back to our clinical scenario (Box 1.1), observes the consultant's problem representation—an elderly gentleman with painless progressive obstructive jaundice and clinically signifi- cant weight loss. Notice how the third-year student's description of ‘jaundice ... getting worse ... no abdominal pain’ has become ‘painless progressive jaundice, and the final-year student's comment ‘higher ALP than AST and ALT’ is now ‘obstructive jaundice’. These are8 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS abstract semantic qualifiers—painless, as opposed to painful; progressive, not time- limited. The consultant has interpreted the information gathered and recognised diagnostically significant details. ‘A sound problem representation using semantic qualifiers is associated with strong clinical reasoning. Conversely, failure to generate an accurate problem representation (whether explicit or subconscious) can lead to guessing possible diagnoses (as in the third-year student) based on isolated findings. The mechanics of making a diagnosis has been discussed in the previous section. Ideally, a combination of pattern-recognition and analytical methods should be used; for the novice clinician to whom pattem recognition is unavailable, analytical methods remain useful. Don’t stop at making a diagnosis. Many diseases have underlying aetiologies (although some are ‘idiopathic’); keep asking why, why, and why—until no further why can be asked. Similarly, every diseases causes complications—search for them (Table 1.2). Table 1.2. Asking ‘Why’ and ‘So What’ in Various Diseases 40-year-old with acute | Why does he havea stroke Is there mass effect? cerebral infarct (eg, cardioembolic, atherosderosis) | |s he able to swallow? Whyis getting a stroke this young? 21-yearoldwithan | Whydid he get an exacerbation? Ishein respiratory failure? asthma exacerbation | (e.g, infection, allergen exposure) 70-year-old with iron | Whyis there iron deficiency? Is he symptomatic? deficiency anaemia __ (eg, occult gastrointestinal bleeding) nts with Multiple Issues Patients can be complex. One disease leads to another, and multiple organ systems canbe affected by ongoing disease processes. Add to that the potential for drug-disease and drug—drug interactions and things can get really confusing. Try to tease out cause-effect relationships and formulate a problem list to make sense of the madness. For example, consider the scenario (Box 1.2).Learning How to Diagnose 9 Box 1.2. A Patient A75-year-old man suffers a fall while getting up from bed. He has been having diarrhoea and vomiting after eating some leftovers, and was given antibiotics by a general practitioner 2 days ago. When he got out of bed, he felt giddy and passed out momentarily. He recalls waking up on the floor, alert and in pain. Examination is normal except for marked postural hypotension and decreased skin turgor. ECG Is normal. Initial blood tests show a creatinine level above his baseline, and an intemational normalised ratio (INR) of 5. A small subdural haemorrhage is seen on CT brain. He has a past medical history of diabetes, atrial fibrillation and heart failure, and is on metformin, bisoprolol, furosemide and warfarin. A problem list might look as follows: 1. Subdural haemorrhage secondary to fall and over-anticoagulation. 2. Syncope secondary to postural hypotension. 3. Hypovolaemia due to gastroenteritis and diuretics, complicated by postural hypo- tension and acute kidney injury. 4, Over-anticoagulation? due to drug interaction (warfarin and antibiotio). Formulating a problem list clarifies the issues on hand, and facilitates the creation of a management plan, which addresses all the problems. Some other issues in diagnostic reasoning deserve brief mention—the use of diagnostic tests, diagnostic uncertainty and diagnostic error. Using Diagnostic Tests ‘A few words about using diagnostic tests is in order. Unfortunately, the scenario in Box 1.3 is not uncommon, This young man has a very low clinical probability of pros- tate cancer, and prostate specific antigen (PSA) testing was inappropriate. The test is likely to be a false positive. Even with a positive test, he remains unlikely to have pros- tate cancer, but most clinicians (and patients) would consider themselves obliged to investigate further ‘to rule out’ cancer. On the other hand, in an 80-year-old gentleman with a hard nodular prostate, bony pain and weight loss, a raised PSA would only increase the clinician’s confidence in a diagnosis of prostate cancer. Box 1.3. The Fallacy of Prostate Spe ic Antigen Te: A25-year-old man seeks a consult fora mildly elevated prostate specific antigen (PSA). He is asymptomatic. It tums out that his company offers ‘free’ yearly health screening, and PSA was included in the standard package. He is anxious about the raised PSA level and upset because his application for a new health insurance policy had been denied on the basis of elevated PSA.10 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS Diagnostic tests help refine a list of differential diagnoses (Table 1.1) by providing additional information to increase the probability of certain diagnoses while decreasing others. For example, amylase/lipase levels are useful in a patient complaining of epigas- tric pain radiating to the back; a positive test makes pancreatitis likely, but a negative test would prompt a search for alternative diagnoses. Diagnostic tests may also help to guide management (e.g, deciding ifa transfusion is necessary), and identify aetiologies (eg, if pancreatitis, are there gallstones?) and complications of disease. As a rule of thumb, before ordering a diagnostic test, know what you are looking for—be sure you know how to interpret the test, how it will help you with diagnosis, and how it will change your management of the patient. Every diagnostic test performs differently; each has a certain false negative and false positive rate. For example, a negative CT scan does not rule out subarachnoid haemorthage in a patient with high pre-test probability (false negative rate), and a moderately high troponin does not always imply myocardial infarction in a patient on chronic dialysis (false positive rate). As illustrated, interpretation of test results depends on the pre-test probability, and one should be familiar with the test characteristic before ordering it. Generally, a sensitive test (low false negative rate) helps to ‘rule out” disease, while a specific test (low false positive rate) helps to ‘rule in’ disease. Some awareness of the issues in diagnostic testing is important, but a detailed discussion of probabilistic reasoning is beyond the scope of this text (but see ‘GO FURTHER at the end of this chapter). Pe Rel tee UCL ae Diagnostic uncertainty is a reality in clinical medicine. Information is limited and conflicting. Treatment decisions often have to be made before the results of clarifying investigations return. This unnerves many but every clinician must learn to be com- fortable with diagnostic uncertainty. Some tools: Admit the uncertainty. Do not commit to one diagnosis and shut out all others, when this is not borne out by evidence. Rather, keep the differentials in mind. Look for a test that will increase the probability of one diagnosis andjor decrease the probability of another. When stakes are high and treatment has a favourable benefit/risk ratio, consider empirical treatment even if the diagnosis is uncertain. For example, patients with minor wounds often receive anti-tetanus toxoid even though the risk of tetanus is low, simply because the risk of treatmentiis smaller than the remote but catastrophic possibility of tetanus. Similarly, in acutely unwell patients with multiple competing differentials, consider treating both diagnoses in parallel if benefit/risk ratios are favourable. For example, in a patient presenting with fever, melena and hypotension: it will be prudent to treat for both septic shock and gastrointestinal bleeding. Reverse what is easily reversible. For example, if it is unclear whether a patient's weakness is due to hypoglycaemia or stroke, administer dextrose and re-examine in 10 min; if weakness persists then it is no longer due to hypoglycaemia. Similarly, if it is unclear whether a patient's breathlessness is due to pneumonia or progressionLearning How to Diagnose Wl of lung cancer, it is reasonable to treat pneumonia (even if the diagnosis is not ‘definite’ ), as the latter would be difficult to treat. Remember that diagnosis is a dynamic process. The patient's response to ongoing treatment is usually very informative. Reassessment over an appropriate time period allows us to use time as a diagnostic tool; be prepared to re-consider the initial diag- nosis if the patient fails to respond to treatment for the assumed diagnosis. Finally, realise that in some situations, it is not necessary to know the exact diag- nosis. These situations include when the treatment for both differentials is identical, when the risk of diagnostic testing outweighs the benefit of information gained (eg. subjecting patients with multiple comorbids to invasive investigations), or when there is no intention to initiate treatment even if a diagnosis is made (e.g., due to quality of life). PoE Te Drill todd Every clinician commits diagnostic errors. Wise clinicians learn from their errors, and learn the pitfalls so that errors are minimised. There are three sources of diag- nostic error: (i) knowledge deficits, (ii) attitude problems such as overconfidence and (ii) cognitive bias. Apart from a good dose of humility and diligent work to improve knowledge, being aware of cognitive biases (Table 1.3) helps one to avoid them. ACRES MT Ea ler) 4 This book suggests sample diagnostic algorithms for several common presentations of disease. I hope that learning from these algorithms will help you to become a better diagnostician, and ultimately help patients. For maximal benefit, some advice is in order. re Things This Book Is NOT |. This is NOT a guide to history taking and clinical examination. Basic history taking and examination skills are prerequisites before higher-order thought processes can be learnt . This is NOT a replacement for standard clinical textbooks. The focus is on differen- tial diagnosis, not on the presentation, diagnosis and management of individual Nv disease entities. Familiarity with individual diseases will help greatly in differential diagnosis—an algorithm makes little sense if the diseases it leads to are completely alien to you. 2 . This is NOT a protocol to rote-learn and follow rigidly, in liew of the fullness of clinical reasoning, Algorithms are but one tool in the diagnostic toolbox, and must be used in conjunction with the other tools discussed in this chapter. . This is NOT a substitute for seeing real patients. That would be a tragedy. Students with extensive textbook knowledge but inadequate patient contact often find S12 Table 1.3. Common Cogni ALGORITHMS IN DIFFERENTIAL DIAGNOSIS 1. Apatient is admitted from the emergency Diagnostic Accepting a previous diagnosis department for pneumonia’. Although he momentum without sufficient scepticism. has no fever, the inpatient team continues to treat‘pneumonia’. Two days later, he is intubated for respiratory failure. A repeat chest K-ray shows worsening fluid overload. 2. Apatient with diabetes and ischaemicheart Anchoring Locking onto a diagnosis too disease presents with acute crushing chest early, based on initial pain. He is given loading doses of aspirin and information. Ticagrelor, while awaiting troponin results. One hour later, he collapses and a post- Premature Accepting the first diagnosis mortem finds aortic dissection. closure without considering whether there is a better answer. Under- Failure to revise the diagnosis adjustment when new information becomes available. Ayoung lady presents with symmetrical small Blind Undue deference to a false- Joint polyarthritis and morning stiffness obedience negative diagnostic test lasting 2 hr. Her general practitioner declines (or other authority). 10 refer her to a rheumatologist because her theumatoid factor was normal. 4, After missing a pulmonary embolism, a Availability Diagnosis influenced by ease of physician orders CT pulmonary bias recollection of possible angiography for the next five patients diagnoses. with breathlessness. 5. A’frequent flyer’ with borderline personality Framing Decision making unduly biased disorder, who is known towaste emergency _effect by the way itis presented. department resources, presents with hip pain and difficulty walking. Sheis sent home Affective Personal emotions affecting with analgesia, with no X-ray done. She is bias judgement. later found to have a hip fracture. themselves unable to apply their knowledge to a clinical situation. Rather, seeing patients goes hand-in-hand with learning knowledge and knowledge structure— remember, patients are a clinician's best teachers. . This is NOT comprehensive or absolute. My goal is simply to help beginner diag- nosticians develop a functional diagnostic framework. Therefore, I focus on common presentations of common conditions, downplaying rare diseases and uncommon exceptions to general rules. This book’ algorithms are generally written for adult patients, and will lead to error if applied in paediatric contexts without modification.Learning How to Diagnose 13 Five Learning Aids to Help You Apart from an algorithm to every symptom, and a discussion of the content of the algorithm, this book includes several elements to aid learning: Clinical case and case discussion: Every chapter opens with a clinical vignette, and closes by applying the content of the chapter onto the clinical case. You are encouraged to pen down how you would approach the clinical vignette before reading the discussion. S . End-chapter summaries of key learning points. Common pitfalls highlighted at the end of each chapter. Illustrative ECGs and radiographs are included in some chapters. For the sake of brevity, only a limited number of images can be included—looking up more exam- ples can be helpful. ae Discussion questions in several flavours are found at the end of each chapter. (a) REFLECT! questions encourage reflection on clinical cases previously encoun- tered, so as to reinforce ‘illness scripts’ and encourage reflective practice. (b) EXPLORE! questions prompt reading up on topics related to the chapter, or prompt reinforcement and application of concepts discussed. (© DISCUSS! questions should be discussed in small groups. They include differing opinions up for debate, as well as mini-exercises to synthesise and apply content learnt in the chapter. (d) GO FURTHER! questions suggest slightly more advanced concepts interested learners might wish to explore. e Ways to Use This Book |. Asa primer. Before beginning a new clinical rotation, flip quickly through relevant chapters. Use the approach discussed as a roadmap to read up on what you are unfamiliar with. Try to hit the ground ranning and learn as much as possible from the first patient you see. . As a just-in-time aid. Bring this book to the bedside. If you ever get ‘stuck’ while clerking a patient, whip out the relevant chapter and see f it can give you some ideas. . Self-practice. Read the clinical case at the beginning of each chapter. Jot down how you will approach the patient. Then read the chapter; as you read, continually edit your draft approach. . Deliberate practice with patients. Clerk patients (without referring to this book or the case notes); list and rank the top differentials. Next, refer to the relevant book chapter, and see if there is anything else it might prompt you to consider (other dif- ferentials to consider, important differentiating questions you did not ask, etc). Go back to the patient and refine your differential list. Finally, open the case notes and investigations, and check if your diagnosis is similar to that of the expert clinician's (if it differs, ask why—the expert is not always right). Repeated deliberate practice is essential in learning clinical diagnosis. S & S14 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS 5, Critical reflection. Examine the algorithms critically. Identify the key differentia- tors used (usually at branch points), and ask: @ What is the pathophysiologic basis of these key differentiators? Understanding the pathophysiologic mechanisms for the presentation of diseases aids under- standing. (ii) To what extent are they reliable? It is helpful to understand what are ‘hard signs’ with high specificity (e.g., upgoing plantars makes an upper motor neuron lesion very likely) and what are ‘soft signs’ with low specificity (eg, a positive anti- nuclear antibody does not always mean that a patient has autoimmune disease). (HH _Using what you have leamt, pen down your approach to the Clinical Case at the start of the chapter BEFORE reading the discussion below. Case Discussion This student is not alone in having good content knowledge but struggling with clinical reasoning and the diagnostic process. Very often the root of the problem lies in (i) poor study habits—rote learning without processing to improve knowledge structure, and (ii) inadequate patient contact. She should immediately: 1. Put down the textbook and hit the wards. Try to clerk as many typical patients as possible, and for every patient, list three top differentials and complete Table 1.1. ny Use algorithms to jump-start her journey in clinical reasoning. This book can help but it is not another textbook to be learnt by rote; rather, it is a teach-book to be used alongside seeing patients (page 13). a « Diagnosis requires content knowledge and knowledge structure. Com- Lessons bining analytical reasoning (via hypothetical-deductive methods and algorithms) with intuitive pattem-recognition increases the chance of diagnostic success. XN An appropriate search for underlying aetiologies, appropriate use of diag- nostic tests, and problem list formulation can assist in diagnostic reasoning, w Hard work, humility and an awareness of cognitive biases are helpful in minimising diagnostic error. Common 1. Trying to memorise this book is not going to work. Pitfalls N Trying to learn medicine without seeing patients is doomed to failure. Questions 1. REFLECT! Recall one diagnostic error that you have made (or witnessed someone make). What were the factors that led to the error? What could have been done differently to prevent the error from being made? EXPLORE! What is the difference between surface learning and deep learning? When might one favour either approach? xLearning How to Diagnose 15 3. DISCUSS! The explosion of medical knowledge in the past 10-20 years has made it difficult for any clinician (must less a beginner) to keep up. What are some strategies to cope with this volume of information and pace of progress? 4, GO FURTHER! Read up on Bayesian conditional probability. How does the pre-test probability influence interpretation of a given diagnostic test?Heart and LungsChapter An Approach to 2 Hypotension Clinical Case You are called to review a 64-year-old gentleman on the Renal ward for a blood pressure of 75/43. He has a history of end-stage renal failure on haemodialysis, diabetes, hypertension, Ischaemic heart disease, atrial fibrillation and peripheral vascular disease with a left below-knee amputation. He was admitted 4 days ago for a thrombosed arteriovenous fistula. As thrombectomy of the fistula was unsuccessful, a tunnelled temporary dialysis catheter (‘Permcath’) was inserted in his right internal jugular vein 3 days ago. He is due for discharge tomorrow. How will you approach his hypotension? Hypotension is a medical emergency. It indicates that the patient is severely ill, with a disease process causing physiologic compromise exceeding the body’ compensatory mechanisms. Pathophysiologically, hypotension reflects a failure in preload (cannot fill left ventricle due to volume loss, inflow obstruction or excessive heart rate), contractility (heart muscle dysfunction or outflow obstruction) or afterload (decrease in systemic vascular resistance due to peripheral vasodilation). In addition to hypotension, look out for tachycardia, oliguria, abnormal mental status and clammy peripheries, which are likewise indicators of shock—a state of tissue hypoperfusion that leads to organ dysfunction. High lactate indicates anaerobic metab- olism due to insufficient tissue perfusion; in sepsis, levels above 4 mmol/L, even with normal blood pressure, has been shown to have a mortality rate comparable to overt septic shock. The causes of hypotension can be classified as hypovolaemic, cardiogenic, obstruc- tive and distributive (Table 2.1). nitial Approach Obtain a rapid clinical impression from history, examination and bedside investiga- tions (including ECG and arterial blood gases). Focus initially on common causes of shock, and any cause that the patient may be particularly predisposed to. The temper- ature of the peripheries is a helpful differentiator—Warm peripheries indicate vasodila- tion, as mediated by bacterial endotoxin (septic shock), histamine (anaphylaxis) or loss of sympathetic tone (neurogenic shock and Addisonian crisis). Conversely, clammy peripheries indicate vasoconstriction, which is an appropriate response to hypovola- emic, cardiogenic and obstructive shock. Figure 2.1 provides an initial approach.20 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS Table 2.1. Causes of Shock Orne Haemorhage— Acute myocardial Massive PE Septic shock obvious or occult. infarct Tension (common) Dehydration Archythmias pneumothorax Anaphylaxis Polyuria or Mechanical Cardiac Addisonian crisis over-diuresis dysfunctioneg, | tamponade Neurogenic acute valve shock rupture Myxoedema crisis Bleeding Chest pain Respiratory Fever Abdominal pain Diaphoresis distress, pleuritic Angioedema, Diarrhoea, Dyspnoea chest pain wheeze, new vomiting Palpitations Unilateral swollen medication leg taken Ea Cold Cold Cold Warm Under-filled Over-filled (usually) | Over-filled Under-filed or Under-filled (tamponade) euvolaemic (very acute) Euvolaemic (PE, pneumothorax) Ce) Pallor Archythmia Calf swelling (PE) —_Localising source Corin Per-rectal bleeding Displaced apex _-Hyper-resonant of sepsis ‘Abdominal beat lung Angioedema, tendemess in Murmur (pneumothorax) stridor occult bleeding Beck's triad (anaphylaxis) (tamponade) FT) Haemogiobin+ ECG, cardiac ECG Septic workup, oxm enzymes Bedside echo lactate Urine pregnancy Chest X-ray CT pulmonary test Consider echo angiogram Consider abdominal imaging GXM, group and cross-match; PE, pulmonary embolism. Institute emergent management based on the bedside clinical impression; do not wait for blood tests to return. Return to review the patient, reviewing if the patient has responded to initial therapy, and if blood results support the initial diagnosis, Failure to respond to initial treatment should prompt reconsideration of the initial diagnosis, and escalation of initial treatment (eg., more fluids and starting inotropes). Adjuncts to bedside clinical diagnosis, in particular bedside ultrasound scanning, can be helpful. Bear in mind that there may be multiple concurrent causes of shock, for example, sepsis precipitating an acute myocardial infarction. The subsequent sections discuss individual classes of shock in greater detail.An Approach to Hypotension 21 Hypotension or raised lactate (>4 mmol/L) Caro) ‘Assess at bedside for a cause (history, examination, ECG) ¥ From initial assessment Is there any obvious cause of shock? e.g. Immediate resuscitation Fever, warm peripheries — Septic shock ~ Chest pain, ischaemic ECG changes Cardiogenic shock Arrhythmia consistent with hypotension (eg.VI.VF) ~ Cardiogenic shock = History of fluid loss, bleeding, melen — Hypovolaemic shock = Urticaria, stridor, and wheeze developing after ‘anew medication was just taken — Anaphylactic shock Pleurtic chest pain, hyper-resonant percussion (especially if on positive pressure ventilation) > Tension pneumothorex Y 1 No obvious cause found Presumptive cause suspected ~ Systemicaly consider the various Treat and monitor response ‘causes of shock (see toxt) >| Send investigations ~ Adjuncts, eg. ultrasound sean (e.g. septic workup, troponins, Hb) Re-evaluatediagnosis |< | Poor response | _|Good response Escalate treatment << | - Continue treatment Consider concomitant 2nd Hb, haemoglobin aetiology of shock Figure 2.1. Initial clinical approach to hypotension. PCI a4 In hypovolaemic states, examination reveals under-filled peripheries (decreased skin turgor, dry mucous membranes, flat jugular venous pressure (JVP) and clear lungs). Hypovolaemia can arise from: Workup © Fluid loss: There may be a clinical history of gas- trointestinal losses (diarrhoea and vomiting), renal losses (polyuria eg., in recovery phase of acute tubular necrosis, or hyperglycaemic emergencies). Victims of heat stroke and burns are also hypo- volaemic.22 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS © Haemorrhagic shock: This is easily suspected in the trauma victim, and if there is an obvious gastrointestinal bleed (eg. haematemesis and melena). © Occult haemorrhage: This requires a high index of suspicion. Be alert to any complaint of abdominal pain in a hypotensive patient which ‘may suggest a ruptured aortic aneurysm (examine for pulsatile mass anda rigid abdomen), a ruptured ectopic pregnancy (look for vaginal bleeding, examine for cervical motion tenderness), or a ruptured hepatocellular carcinoma. Do a digital rectal examination for melena. (eT POEL ra 4 The EGG is a critical bedside diagnostic tool. © Acute coronary syndromes: This diagnosis is easy in the classic presentation of chest pain, dyspnoea and diaphoresis. However, many patients, especially those with diabetic neu- ropathy, present atypically. ECG reveals acute S-T changes (see Chapter 3). © Unstable arrhythmias: Certain tachycardias (ventricular tachycardia and fast atrial fibrillation) and bradycardias (high-grade atrioventricular blocks, ice, type 2 second degree and third degree blocks) can cause hypotension (Chapter 5). ‘Hypotension is less common with other arrhyth- mias (e.g., ectopic beats and low-grade atrioven- tricular blocks) and another cause should be sought. © Poor cardiac contractility: In decompensated heart failure and in severe metabolic disturbances (eg., hyperkalaemia, acidosis and hypothermia). © Structural catastrophe: Acute valve rupture, ventricular septum or free wall rupture post myo- cardial infarction. Disti e Shock ~ Full blood count (FBC): ‘A drop in haemoglobin is suggestive, but it may remain normal in acute blood loss ~— Bedside ultrasound ~ CT abdomenipelvis, (if stable enough to be transported) ~ Urine pregnancy test Workup ~ Cardiac enzymes — Echocardiography ~ Cardiac enzymes — Electrolytes — Electrolytes — Echocardiogram ~ Echocardiogram In distributive shock, cardiac output is normal but peripheral vasodilation and de- creased systemic vascular resistance results in hypotension.An Approach to Hypotension 23 Workup © Septic shock: Immunocompetent patients may ~ Septic workup present with fever and localising symptoms. (inflammatory markers, Immunocompromised hosts (eg., neutropenia, blood cultures and look immunosuppressed and HIV/AIDS), on the other for a source, eg,, chest hand, may not mount a febrile response. Sepsis X-ray [CXR], urine should be considered in all immunocompromised formed element patients regardless of fever. Look carefully for a mictoscopy (UFEME] source (see Chapter 35), and urine cultures) — Lactate © Anaphylactic shock: A clinical picture of rash, — Mast cell tryptase (asa facial angioedema, flushing and stridor developing confirmatory test where soon after a precipitant, such as a medication diagnosis is unclear) error or new drug. © Neurogenic shock: High spinal cord injury — Ifsuspected, treat with (usually above T6) compromises sympathetic IV steroids. outflow to the heart and peripheries. This should — Confirmatory testing be apparent from the clinical history e.g., trauma. (e.g. Synacthen test) inappropriate in hypotension. * Addisonian crisis: Difficult to diagnose unless suspected. Consider especially in patients on chronic steroids (or traditional medicine with steroids) whose steroids are abruptly withdrawn, or who are suffering an intercurrent illness without receiving stress doses of steroids. © Myxoedema coma: Severe hypothyroidism may — Thyroid function tests present with hypotension. Findings may include hypothermia, bradycardia (in spite of hypoten- sion) and signs of chronic hypothyroid disease including coarse skin, facial oedema, eyelid oedema and thick tongue. Pe) aa Cites Workup © Massive pulmonary embolism: This patient ~ CT pulmonary would be hypoxic and in respiratory distress. angiogram (Figure 2.2) ‘There may be evidence of deep vein thrombosis (DVT) (swollen and tender calf) and risk factors for DVT/PE (eg,, prolonged immobility, recent surgery and active malignancy). ECG may reveal an S,Q;T; pattern (most pathognomonic; Figure 3.2(£)), sinus tachycardia (most common) or T24 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS inversions and signs of right heart strain (e.g., right bundle branch block). © Tension pneumothorax: A patient with pleuritic chest pain, dyspnoea and hypoxia. Examination will reveal tracheal shift, unilateral absence of breath sounds and hyper-resonance to percussion. ‘Mechanical ventilation of a tension pneumothorax rapidly worsens shock. © Cardiac tamponade: Classically, Beck’s triad describes a patient with distended jugular veins, muffled heart sounds and hypotension—but all three criteria are rarely fulfilled in practice. There may be a background of malignancy (lung or breast), myocarditis, recent myocardial infarction, autoimmune disease (e.g, lupus or rheumatoid arthritis) or uraemia. ECG may show small voltage complexes and electrical alternans. ~ This should be a clinical diagnosis and not a CXR diagnosis, — Bedside echocardiogram Figure 2.2. Pulmonary embolism—A premorbidly well 45-year-old gentleman presented with acute hypotension, chest pain and dyspnoea. There was no suggestion of infection or bleeding, ECG showed sinus tachycardia, CXR was normal and troponins were negative. A CT pulmonary angiogram revealed a large saddle embolus (arrow). He collapsed soon after this CT was taken and could not be resuscitated. (GP Using what you have learnt, pen down your approach to the Clinical Case at the start of the chapter BEFORE reading the discussion below. Case Discussion This patient is at risk for multiple types of shock. With a central vascular catheter, he is predisposed to central-line associated sepsis. He is a vasculopath and at high risk of acuteAn Approach to Hypotension 2s myocardial infarction. He may be hypovolaemic due to over-ultrafiltration during dialysis, and may have bled from procedural complications of dialysis catheter insertion. First assess airway, breathing and circulation. Learn to recognise the toxic patient whose collapse is imminent—an obtunded patient with agonal breathing is clearly sicker than a conversant patient with the same blood pressure. Resuscitation proceeds simultaneously alongside diagnosis. Review the vitals chart. History and examination should first target the most likely differentials. Ask for symptoms of myocardial ischaemia (e.g., chest pain), of sepsis (e.g. fever, dyspnoea and intradialytic rigours), bleeding and so on. Feel the peripheries (warm suggests sepsis), examine for sources of sepsis (e.g, pus coming from indwelling lines, lung crepitations, etc), and look for bleeding from the procedural site. An ECG may reveal ischaemic changes or new arrhythmias (his known atrial fibrillation is unlikely to cause hypotension). Review if he has been given any new medication that could cause hypo- tension. Initial investigations, at a minimum, should include haemoglobin, inflammatory markers, blood cultures, cardiac enzymes, electrolytes and lactate (look for metabolic acidosis). Identify a likely diagnosis. If there are two competing diagnoses, it may be necessary to pursue both simultaneously (e.g., start antibiotics while awaiting troponin results). Fluids are helpful in most types of shock but exercise caution with fluids in proven left ventricular infarction, aortic dissection (allow permissive hypotension), and where fluids would delay emergent treatment (i.e. giving adrenaline in anaphylaxis, decompressing tension pneumothorax and cardioverting an unstable arthythmia). Fluids should not be withheld from this patient just because he is on dialysis—hypotension kills before fluid overload does—and should be given unless he is already overtly fluid overloaded. Proceed with empiric treatment of the likely cause of shock, but re-evaluate the response to treatment frequently, and refine the working diagnosis as additional information becomes available. Key 1, Faced with a hypotensive patient, think through the four main classes of Lessons shock: Hypovolaemic, cardiogenic, distributive and obstructive. 2. Warm peripheries indicate vasodilation and hence distributive shock; cold peripheries suggest the other types of shock. 3. Resuscitation proceeds alongside diagnosis. Begin empiric treatment with frequent review to assess for response. Common 1. Occult bleeding is easy to miss. Be alert for any abdominal pain or tender- Pitfalls ness ina hypotensive patient. 2. Do not fixate on a working diagnosis—be ready to re-evaluate the diag- nosis ifthe patient fails to respond to empiric treatment, and as new infor- mation becomes available.26 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS . There can be multiple causes of shock; all need to be treated simultane- ously. Questions REFLECT! Obstructive shock can be difficult to recognise. Have you ever encountered a patient with obstructive shock? What was the initial impres- sion, and what clues made you/your team suspect the diagnosis? DISCUSS! ‘The management of shock depends little on its aetiology’. To what extent do you agree? . EXPLORE! Look up the Surviving Sepsis guidelines (wwwsurvivingsepsis. org). This provides a helpful overview of the approach to sepsis. Can you understand the rationale behind the recommendations given? GO FURTHER! Bedside ultrasound can be extremely valuable in shock. How does ultrasound distinguish between the causes of shock, and assist "te management?27 Chapter An Approach to 3 Chest Pain Clinical Case A 55-year-old indian gentleman presents with sudden-onset chest pain and diaphoresis. His past medical history includes hypertension, obesity and heavy smoking. He has had intermittent episodes of chest tightness in the last few years, which was attributed to gastrointestinal reflux disease, but they had been getting worse of late. On arrival at hospital 15 min later, his pain has improved markedly. His vital signs are: BP 170/94, HR 98, SpO2 98% on room air. ECG is as shown. How would you approach his chest pain? ae eran en tla at ee ee Patients with chest pain can be very sick. Learn to triage rapidly—the patient winced up in pain, panting in distress, and whose clothes are soaked with sweat is rather likely to have an acute myocardial infarction, and should be tended to immediately. At the same time, chest pain is a staple among the worried well—not all of whom require the customary two or three sets of cardiac enzymes. The causes of chest pain can be clas- sified by organ system and presentation (Table 3.1).28 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS Table 3.1. Major Causes of Chest Pain en AMIl unstable | Pneumo- esophageal eorus is angina thorax rupture Aortic PE dissection Cardiac tamponade CT EFETS | myocarditis | Pneumonia Cholecystitis Rib fracture Lebel Pericarditis Pancreatitis Herpes zoster Er Angina AS HOM ret GERD Musculoskeletal esophageal Anxiety spasm Referred pain AMI, acute myocardial infarction; AS, aortic stenosis; GERD, gastroesophageal reflux disease; HOCM, hypertrophic cardiomyopathy; PE, pulmonary embolism. The approach to chest pain rests on a well-taken history and examination. The ‘Socrates’ questions (site, onset, character, radiation, associated symptoms, fime course, exacerbating and relieving factors, severity) seem to have been written for chest pain—and work well here (Table 3.2). OT tear hy The presentation and causes of cardiac ischaemia are worth further discussion. Ischaemic chest pain is classically a crushing, poorly localised retrosternal pain radi- ating to either arm or jaw, associated with diaphoresis and dyspnoea. Some patients (especially the elderly) may describe their symptoms as discomfort or tightness rather than pain, or may not have much chest pain at all. Not every ischaemic pain is equal—it may represent myocardial ischaemia or infarction, thatiis, cell death (Table 3.3). Differentiating the Syndromes of Cardiac Ischaemia History differentiates stable angina—of ‘usual’ quality and duration—from the acute coronary syndromes (ie, unstable angina, non-ST-elevation myocardial infarction [NSTEMI] or $T-elevation myocardial infarction [STEMI]), which is anything new or more severe than before. ECG identifies a STEMI, and may show ischaemic changes in the other syndromes of cardiac ischaemia. Unstable angina vs. NSTEMI cannot be differentiated on history and ECG alone, and must wait for cardiac enzymes; troponins are elevated in NSTEML, representing cell death.An Approach to Chest Pain 29 Table 3.2. Differentiating the Causes of Chest Pain: The ‘Socrates’ Questions Onsetandtime Wt.) Acute Episodic Variable Episodic a pain (angina) ey Diffuse Tearing, Pleuritic Burning Localised crushing migratory fe Eitherarm, Interseapular Lateral, back Epigastrium Nil [rere Diphoresis Stroke, Dyspnoea _—_—Reflux - Pd syncope Heartbum Migratory pain TET) exertion Inspiration —_Meals,lying_-- Coughing rs down Palpation Sry Nitrates Time (pain Antacids tors max at onset) SEYEPETESI) vascular risk Hypertension Youngand = — - erat factor Connective thin tissue corp disease DVT tisk factors AMI, acute myocardial infarction; COPD, chronic obstructive pulmonary disease; DVT, deep GERD, gastrointestinal reflux disease; MSK, musculoskeletal. thrombosis; Table 3.3. Syndromes of Ischaemic Chest Pain NSTEMI, non ST-elevation myocardial infarction; STEMI, ST-elevation myocardial infarction, |__| rethophysictoy | tory eco troponin | Stable coronary _| Episodic exertional chest pain, Normal occlusion lasting 3-5 min, completely relieved by rest or nitrates Worsening Newonset angina, worsening STdepressionor | Normal coronary ‘angina (increasing severity, normal (if pain occlusion frequency, duration of had resolved) symptoms or triggered by less. exertion) and angina at rest [FSET] mroote death Elevated Myocytedeath Acute episode, more severe than sT-elevation Elevated usual angina, lasting >30min | with reciprocal changes30 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS I | Approach Resuscitate the patient—there may be hypotension (cardiogenic or obstructive shock), or hypoxia (pulmonary oedema from myocardial infarction, tension pneumothorax, massive pulmonary embolism), The immediate priority is to consider the life threat- ening causes of chest pain (Figure 3.1). —— Tetetfetiveatenng | CHESTPAN) Otheracate causes couse = Pericarditis = AML ~ Myocarditis ~ Aortic dissection = Pneumonia =Pneumetherax (~|_History & Examination ~Grolecysttis PE yrs Rib fracture ~Carciac tamponade Zoster Oesophageal rupture < Ea. eee leek Specie noo Toponnes-3 Spec wor on Bc od ts [reesutra insider cardiac conditions with serious complications Coronary artery disease ~> stresstesting ~ Aortic stenosis Thee jet Other conditions Gastrointestinal GERD ~ Oesophageal spasm = Referred pain Musculoskeletal Anxiety = Costochondritis ~ Cervical spondylosis ~ Shoulder arthritis ‘AMM, acute myocardial infarction; GERD, gastroesophageal reflux disease; HOCM, hypertrophic cardiomyopathy; PE, pulmonary embolism. Figure 3.1. Approach to chest pain. PU Pee Wu l) © Acute coronary syndrome: Crushing retrosternal pain radiating to arm or jaw, associated with diaphoresis and dyspnoea (syndromes of cardiac ischaemia, page 28). Cardiac and respiratory auscultation may be entirely normal. © Aortic dissection: A sharp tearing pain radiating to the back between the scapulae. Onset is sudden, pain is maximal at onset and improves with time. If dissection involves the carotid artery origin, there may be neurological symptoms (stroke, tran- sient ischaemic attack, syncope); if dissection involves the subclavian artery origin, there may be unequal blood pressure and pulses between the limbs (manifesting as radial-radial and radial-femoral delay on clinical examination). Retrograde spreadAn Approach to Chest Pain 31 of dissection causes aortic regurgitation (an early diastolic murmur) or myocardial infarction, and forward spread causes pain to migrate from chest to abdomen tolegs. © Pneumothorax: Acute-onset pleuritic (worse on breathing in/out) chest pain + dyspnoea + desaturation, with decreased chest expansion, hyper-resonance and decreased air entry on the affected side. A one-way valve may result in intra-thoracic pressure building up, obstructing venous return and causing hypotension— this is a tension pneumothorax, which manifests with hypotension, mediastinal shift and tracheal deviation, in addition to the signs of a simple pneumothorax. Pneumothorax may be primary (eg., in tall young men), or secondary (e.g., in patients with chronic obstructive pulmonary disease, on positive pressure ventilation, or post-thoracic procedures). © Pulmonary embolism: Classically pleuritic chest pain, dyspnoea and hypoxia with anormal lung examination; if massive, there may be hypotension. Presentations are often subtle with unexplained hypoxia, tachycardia and tachypnoea; absence of hypoxia makes pulmonary embolism less likely. Look for unilateral leg swelling (deep vein thrombosis) and risk factors (e.g., immobility, malignancy, pregnancy— see Chapter 47). © Cardiac tamponade: Chest pain and dyspnoea, associated with hypotension, ele- vated jugular venous pressure, pulsus paradoxus (large fall in systolic blood pressure [> 10 mmtig] on inspiration) and muffled heart sounds. There may be a history of cancer, uraemia, autoimmune disease (e.g,, lupus) or trauma. © Oesophageal rupture: Chest pain following violent vomiting or oesophageal instrumentation. This is immediately available and should be performed stat. It is always helpful to have old ECGs for comparison. Look for © SI-elevation: ST-elevation with reciprocal depressions defines a STEMI (Figures 3.2a-b). Localise the infarct: ST-elevation in V1-3 in anterior STEMI (left anterior descending artery), V4-6 in lateral STEMI (left circumflex artery), Il, IIT and aVE in inferior STEMI (right coronary artery). In inferior STEMI, do right-sided Jeads for right ventricular infarction (important to identify, as it carries implications for management). © ST-depression: It implies coronary ischaemia, which may be angina, unstable angina or NSTEMI. Look out for widespread ST depressions with ST-elevation in aVR—this isa high-risk feature suggesting left main coronary artery or proximal left anterior descending artery disease, and is considered a STEMI-equivalent (see the ECG in Section ‘Clinical Case’). © Twave changes: T-wave inversions (Figure 3.2d) may be a sign of cardiac ischaemia, but are less specific than ST-elevations or depressions; they can also be seen in normal individuals, left ventricular hypertrophy, pulmonary embolism and other conditions. Repeat the ECG and compare against old ones—T-wave changes that are32 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS new (not present on old ECGs) or dynamic (changes with repeat ECGs) are more worrisome. Certain T-wave features are particularly high-risk: deeply inverted or biphasic I-waves in V2-3 (Wellens’ syndrome), T-wave inversion inaVL (may precede development of ST-elevations in inferior leads) and hyperacute T-waves (very straight ST upslope rather than a curve), © Left bundle branch block: Widened QRS (> 120ms) with dominant s in V1 and ‘M’ shaped QRS complex in V6" (Figure 3.2e). Chest pain plus left bundle branch block (LBBB; especially if new) is suspicious of myocardial infarction (MI). Look for old ECGs—in an old LBBB, interpreting ST changes can be tricky.” © Pulmonary embolism: An $,Q,7; pattern (Figure 3.2/) is pathognomonic but rare; more commonly, the ECG shows sinus tachycardia, T inversion or evidence of right heart strain (eg., right bundle branch block, P pulmonale). © Cardiac tamponade: Low voltage QRS complexes. © A normal ECG does not rule out myocardial infarction. Repeat ECGs for dynamic changes. Figure 3.2. The ECG in chest pain. dot pepe ti SE ae ATI ay 7 pf (@) Anterior STEMI: A 60-year-old gentleman presents with crushing chest pain and diaphoresis. ECG shows ST-elevations in V2-4. These ST-elevations are concave downwards ‘sad face’), as compared to the concave upwards (smiley face’) ST-elevations in normal early re-polarisation. Coronary angiography revealed 99% stenosis in the left anterior descending artery. (Conta) * § helpful mnemonic: WiLLiaM (in left bundle branch block, there is ‘W’ in V1 and ‘M’ in V6) vs. ‘MaRRoW (in right bundle branch block, there is ‘M’ in V1 and “W’ in V6). 2 Consider the Sgarbossa criteria (beyond the scope of this text).An Approach to Chest Pain 33 Ph tf : : ce | Go oo 1 ee (b) Inferior STEMI: A 48-yeer-old gentleman presents with chest pein and dyspnoea. ECG shows ST-elevations in lead ll Ill and aVF, with reciprocal ST depressions in land aVL. Deep ST depressions InV1-3 suggest concomitant posterior infarction. This patient should receive a right-sided ECG (for right ventricular infarction) and close monitoring for arrhythmias (an inferior STEMI may involve the sinoatrial and ztrioventricular nodes). Coronary angiography revealed 100% stenosis in the right coronary artery. (9 Pericarditis: 21-year-old infantry trooper presents with chest pain. His ECG shows widespread ST-elevations (lead II >Ill) and PR depression (best seen in inferior leads). Contrast this with the STEMI ECGs in (a] and (b). He recovered well with anti-inflammatory medication. ee pop See eee tat Ee (¢) Tinversions: A 70-year-old lady with a past medical history of diabetes presents with a 12-hr history of vague chest ‘tightness: ECG reveals T-inversions in V2-4 (the T-wave in V1 can normally be inverted) and q-waves in lead Il. These changes were not present on old ECGs. Her troponin returned grossly elevated and coronary angiogram revealed 95% occlusion of the left anterior descending artery. (Contd)ba oon ade fo Decal heh / cy oe : Pend ninedinnlftentrenctnendnendy / i (Vo MUA Bee ea poafa Jn JV RA AAA! Nai brian ai yy : 4 () Left bundle branch block (LBB3): A 60-year-old gentleman presents with sudden-onset diaphoresis, but no chest pain. He is known to have diabetes, and had defaulted all medication: ECG shows a new LBBB (\ ned QRS, ‘M-shape’ notched QRS in V5-6, and ‘W-shape' slurring in V1-2). Soon after this ECG was taken, he developed ventricular fibrili and could not be resuscitated. Troponin subsequently came back grossly elevated. Because of diabetic neuropathy, he was pain-free despite having a myocardial infa Itt de yyy? L he A nonce? ‘ Pablo aad aalieraragerartel bata nba | (6) Pulmonary embolism: A young lady with a history of multiple miscarriages presents with chest pain and dyspnoea after a long flight. ECG reveals 5:07 and a right bundle branch block. CT pulmonary angiogram confirmed pulmonary embolism. She was later found to have anti- Phospholipid syndrome and received lifelong anticoagulation. ohh Rear ma ppp ror (@) Underlying ischaemic heart disease: This 50-year-old lady gave a history of stable angina. Serial ECGs showed no dynamic changes, and troponins were normal. However, her ECG shows ‘q-waves in lead Ill, indicating an old inferior myocardial infarction, as well as small lateral R-waves in lateral leads (V4-6) with poor R-wave progression (these R-waves should be positive, rather than negative), suggesting loss of myocardium in the lateral wall of the left ventricle. She was eventually found to have triple vessel disease.An Approach to Chest Pain 35 threatening causes: Subsequent inve: Subsequent workup in the acute setting involves Cardiac enzymes: Do two sets of troponin, 3-6 hr apart. A negative first set does not exclude myocardial infarction, as values may take 3 hr to rise. Today's high-sen- sitivity troponins are good for picking up myocardial infarction (and ruling out if negative), at a cost of reduced specificity; they may be positive in renal failure, myo- carditis, arrhythmias and pulmonary embolism. © Chest X-ray (CXR): For pulmonary oedema, pneumothorax (Figure 4.2f), widened mediastinum (dissection), pneumomediastinum (oesophageal rupture). Other tests based on clinical suspicion: © For PE: Bedside ultrasound is helpful to visualise right heart strain. CT pulmonary angiogram confirms pulmonary embolism. D-dimers are only useful to rule out PE if pre-test probability is low (Le. do not orderif PE is clinically suspected) (Chapter 4). © Foraortic dissection: Bedside ultrasound for a widened aortic root > 3 cm, and CT angiogram if stable. © CT thorax: For oesophageal rupture. © Echocardiography: For cardiac tamponade. 2. Other acute causes requiring treatment ‘The above evaluation will also identify these causes of acute chest pain which, while usually not as immediately life threatening as the prior ones, do require treatment. © Pneumonia with pleurisy: Pleuritic chest pain with fever, cough sputum produc- tion and dullness and/or crackles on respiratory examination. CXR reveals consolidation. Continue with a septic workup including inflammatory markers (full blood count [FBC] and C-reactive protein [CRP]), sputum and blood cultures. © Pericarditis: Chest pain is pleuritic in nature, and there may be a pericardial fric- tion rub. The diagnosis is usually made when the ECG finds widespread ST-elevation (lead I] > lead II1), with reciprocal ST depression in aVR and V1, and PR depression. (Figure 3.2c). Look for an underlying aetiology—infective, neoplastic, uremic, post-infarction and autoimmune. © Myocarditis: Presentations vary and can be vague; a typical scenario is a young, previously healthy patient, who comes in with a viral prodrome (eg., fever, malaise, arthralgiz) and subsequently develops chest pain. Myocardial inflammation results in elevated troponins, arrhythmias (bundle branch block, high-degree atrioven- tricular blocks, etc.) and acute heart failure. Myocarditis can be mild or rapidly fatal. © Rib fractures: Unfortunately, these are usually undiagnosed until a CXR is done and looked at carefully! © Referred pain from cholecystitis: Consider if history is suggestive (Table 3.1) (cee Chapter 9).36 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS © Herpes zoster: Pain may precede the dermatomal vesicular eruption, such that patients receive work up for cardiac disease, without positive findings. The diagnosis, is revealed days later when vesicles appear. In the absence of red-flag history and examination findings, and normal initial investi- gations (ECG, CXR and troponins), move on to consider other less serious causes. LEE i | Workup Having excluded the serious acute causes of chest pain, consider (3) whether chest pain could be due to an underlying cardice disease, and (4) other diagnoses. Me Le TUE ee Te sy @ Ischaemic heart disease: Some patients with anginal symptoms have underlying ischaemic heart disease, and may require further testing. First, decide on the pre- test probability of ischaemic heart disease, which increases with age, a history of ‘classic angina’ (i.e, typical quality and duration, provoked by exertion or stress and relieved by rest or nitrates), cardiovascular risk factors (eg., diabetes, hyperlipi- daemia, hypertension) and suspicious ECG changes (Figure 3.2g). Next, if pre-test probability is at least moderate, consider non-invasive stress testing to confirm the diagnosis and obtain prognostic information. Choose an appropriate modality of stress test—for example, (a) treadmill ECG in patients who can exercise and have a normal rest ECG, (b) exercise stress echocardiography in patients who can exercise but have a non-interpretable ECG (e.g., bundle branch block or pacing) and (© pharmacologic stress imaging for patients who cannot exercise. The results of the stress test determine the therapeutic approach taken; patients with intermediate to high-risk criteria on non-invasive testing generally receive coronary catheterisation, which is both diagnostic and therapeutic (stenting). A final consideration—in patients with ongoing symptoms, unstable angina or high-risk ECG (eg., suspected left main occlusion), stress testing is risky and it may be preferable to proceed directly to coronary catheterisation. © Aortic stenosis (AS) or hypertrophic cardiomyopathy (HOCM): Identification of left ventricular hypertrophy on ECG (eg., Sokolov criteria: S-wave V1 + tallest R-wave in V5-V6 > 35 mm), an ejection systolic murmur or concomitant syncope, prompts concern for AS or HOCM. Chest pain can develop in these conditions, and is usually a bad sign (implying myocardial hypoperfusion caused by increased wall stress and myocardial oxygen demand). Echocardiography confirms the diagnosis. Meee Cre Non-cardiac diagnoses often present with chest pain. Gastrointestinal causes commonly mimic cardiac ischaemia. © Gastroesophageal reflux disease: Heartburn and reflux ave classic, but may also present as ‘chest pain’. Symptoms are exacerbated by lying down, food, alcohol and aspirin, and relieved by antacids.An Approach to Chest Pain 37 © Oesophageal spasm: Mimics cardiac ischaemia or aortic dissection—pain is retro- sternal + radiating to the back, and relieved by nitrates. © Referred pain from sub-diaphragmatic aetiologies: For example, biliary colic and peptic ulcer disease. ‘Musculoskeletal causes: © Costochondritis: Pain localised to a single point on the chest well, exacerbated by palpation and chest wall movement. © Cervical spondylosis causing nerve root compression: Usually a shooting pain radiating from neck to arms, worse on neck movements (Chapter 51). © Shoulder joint arthritis: Pain exacerbated on shoulder movements, may be local- ised on shoulder joint palpation (Chapter 49). Psychiatric: Anxiety and related disorders (panic disorder, depression with anxious features, somatisation disorder) are relatively common causes of chest pain, but are diagnoses of exclusion. Caveat: Due diligence must always be done to exclude serious causes of chest pain. The presence of known anxiety or gastroesophageal reflux does not exclude a serious cause of chest pain developing. Symptoms of myocardial infarction can be non-specific; vomiting can occur in acute myocardial infarction (AMI) and does not always point to a gastrointestinal aetiology. (HH Using what you have learnt, pen down your approach to the Clinical Case at the start of the chapter BEFORE reading the discussion below. Case Discussion This is a middle-aged gentleman with multiple cardiac risk factors (hypertension, obesity, smoking, ethnicity). His pain is most likely due to cardiac ischaemia—it has a typical character, is relieved by rest and associated with diaphoresis. The history of intermittent chest tightness, which has been getting worse, may well be unstable angina instead of gastroesophageal reflux disease (GERD). ECG is significant for diffuse ST-depression with ST-elevation in aVR: this is @ high-risk ECG suggestive of left main coronary artery occlusion, proximal left anterior descending artery ‘occlusion or severe triple vessel disease. He either has unstable angina or NSTEM|; troponin levels distinguish the two (normal in unstable angina, elevated in NSTEMI). While awaiting blood tests, he should receive emergent treatment (including antiplatelets). He will likely require an inpatient coronary angiogram and intervention (coronary stenting or coronary artery bypass grafting). Key 1. Initial evaluation of chest pain focuses on the acute life threatening causes: Lessons. acute coronary syndrome, aortic dissection, (tension) pneumothorax, pul- monary embolism, cardiac tamponade and oesophageal rupture.38 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS Initial investigations in chest pain include an ECG, cardiac enzymes and investigations (eg., pulmonary angiogram and CT angiogram) lered if indicated by clinical suspicion. These may reveal the life threatening causes above, or other causes requiring treatment. Consider cardiac stress testing in patients with symptoms of angina, and tisk factors. Common Pitfalls Patients with acute myocardial infarction can have ‘normal’ ECG and troponin, when done very early. If there is a high clinical suspicion but a normal ECG, repeat serial ECGs (e.g, 15 min later), looking for dynamic changes. Not every patient with an elevated troponin has myocardial infarction. There are other causes, especially renal failure, heart failure and arrhyth- mies. Questions REFLECT! Have you ever seen a STEMI patient wheeled through the Emer- gency Room door? How did the patient look? Would you have suspected the diagnosis even before an ECG? DISCUSS! Two to three sets of cardiac enzymes (troponin and CK-MB) are routinely ordered. Do all patients with chest pain require cardiac enzymes, and is it necessary to order both troponin and CK-MB? EXPLORE! A patient complains of episodic musculoskeletal-sounding chest pain. However, he has multiple cardiovascular risk factors. In this scenario, how do you differentiate cardiac chest pain requiring treatment, from musculoskeletal pain, for which he can be safely reassured? GO FURTHER! Why do we distinguish between primary and secondary pneumothorax? What is the significance of this distinction?39 Chapter An Approach to Dyspnoea 4 and Hypoxaemia Clinical Case A 56-year-old gentleman with a history of diabetes, ischaemic heart disease and chronic obstructive pulmonary disease (COPD), is admitted for breathlessness and cough. The cough had started 2 days ago; it was productive of thick sputum. Breathlessness started suddenly 5 hr ago. There is no chest pain or fever. On examination, his vitals are BP 154/92, HR 137, RR 27, $p02 85% on room air. He appears uncomfortable; each breath is effortful and he is barely able to speak. Examination is remarkable for bilateral basal crepitations and scattered rhonchi. How would you approach his breathlessness? Dyspnoea is a sensation of shortness of breath or difficulty breathing. ‘Desaturation’ refers to low peripheral oxygen saturation, as picked up by a pulse oximeter and reflects hypoxzemia, that is, low blood oxygen concentrations. Patients may have signs of res- piratory distress (e.g,, tachypnoea and use of accessory muscles of respiration). The causes of dyspnoea may be divided by organ system and time course (Table 4.1). Table 4. Causes of Dyspnoea by Organ System and Time Course Upper airway Pneumonia copp obstruction Asthma Interstitial lung disease Asthma Exacerbation of COPD Tumour Pneumothorax Pleural effusion Pulmonary embolism Acute pulmonary Decompensatedheart Heart failure oedema, including AMI | failure Renal failure Anaemia Decompensated Neuromuscular cirrhosis weakness Metabolic acidosis Neuromuscular weakness, Anxiety AMI, acute myocardial infarction; COPD, chronic obstructive pulmonery disease.40 ALGORITHMS IN DIFFERENTIAL DIAGNOSIS. I | Approach Patients with dyspnoea can be unstable. Assess airway, breathing and circulation; look for features of respiratory distress and whether the patientis toxic. Decide if the patient requires immediate airway support (.e, intubation in most cases), if there is time for an azterial blood gas (ABG) and quick evaluation before taking further decisions, or if one can afford a full history and exemination. Providing supplemental oxygen is important, but remember that it will not address the underlying cause. Begin with a directed history, focused chest examination (Table 4.2), bedside investiga- tions (ECG and blood gas) and a chest X-ray (CXR). Identify the likely clinical picture (Table 4.3). Bear in mind that patients may have multiple causes of dyspnoea, for example, myocardial infarction on top of chronic obstructive pulmonary disease (COPD). Arterial Blood Gas Consider performing an ABG if the patient looks unwell, or if peripheral oxygen satu- rations are low. It is rapid and provides a lot of information. ¢ pO; and PaOz/FiO> ratio (‘P/F ratio’): Low PaO2 (< 60 mmHg) confirms hypox- aemia. This is only meaningful when read together with the inspired oxygen concentration (FiO2)—for instance, a PaO, of 75 mmHg would be normal if the patient were on room air, but poor if he/she were on 50% oxygen. Use the ratio of PaQ; to FiO, (i.e., PaO,/FiO,) to correct for varying FiO, concentrations (FiO, can be calculated: nasal prongs, 21% + 4% per 11 02; venturi mask, as set; face mask, ~50%; non-rebreather mask, 70—100%). Any P/F < 300 is worrying. © PCOx Ifhypoxaemia is established, PCO, distinguishes type 1 (PCO, < 45 mmHg) vs. type 2 (PCOz > 45 mmHg) respiratory failure. Type 2 respiratory failure reflects hypoventilation and cannot simply be overcome by increasing FiO, (ie., by providing supplemental oxygen). © pH: For acidosis (pH < 7.35). Distinguish respiratory (*PCO,) from metabolic ({ HCO,, compensatory | PCO,) acidosis, both of which can present with dyspnoea (cee Chapter 16). ¢ Alveolar-arterial (A-a) gradient: This can be calculated and helps to distinguish causes of hypoxaemia. — Reduced A-a gradient: Lung disease (ventilation/perfusion mismatch, diffusion abnormality) or shunts. — Normal A~a gradient: Hypoventilation, low inspired FiO, The implication of anormal A-a gradient is that the cause of hypoxaemia is not pulmonary.