PART 1 MEDICAL MUSTS AND MUST NOTS: SIX ESSENTIALS OF WARD WORLD

Art consists of limitation. The most beautiful part of every picture is the frame.

Clinical decision making is distinct from that of virtually every other type of profession. At least two ingredients make it unique. One is that clinicians almost always base their assessments on physical contact with the patient. Second, experienced clinicians, at least subliminally, implement a two-tiered analytic process in framing a differential diagnosis of what is wrong with a given patient, which can be described as a bit like the classic London double-decker sightseeing bus. The first tier is based on the principle of worsts first—the urgent consideration of all the possible diagnoses that might literally kill someone and need to be excluded right away, regardless of their statistical rarity. The second tier is a more comprehensive list of considerations—the set of all or most of the possible causes of the patient’s problems, including both rarer and more common conditions. The latter can be viewed as the “all things considered” category. To illustrate this two-tiered approach, we selected the differential diagnosis of two of the most common and important presenting symptoms in clinical and emergency medicine: chest discomfort and right upper quadrant pain.

DIFFERENTIAL DIAGNOSIS OF CHEST DISCOMFORT

Chest Discomfort: Worsts First

There are multiple causes of chest discomfort (this problem is usually considered under the heading chest pain). However, patients with acute or chronic coronary syndromes—stable or unstable angina and even acute myocardial infarction—often emphatically deny having pain but will admit to having chest discomfort or chest “pressure.” Of these causes, only a handful of diagnoses are both life-threatening and treatable in the immediate term and, therefore, need to be considered urgently and excluded. These potentially lethal causes of chest discomfort (pain) can be grouped into three categories: cardiovascular, pleuropulmonary, and gastrointestinal (Exhibit 5.1).

1. Cardiovascular causes of chest discomfort include (a) acute coronary syndromes, (b) aortic aneurysms (dissecting or expanding), and (c) acute pericarditis. Of these, acute pericarditis is the least likely to be fatal but needs to be among those considered first because it (a) can lead to death from cardiac tamponade if an effusion is present, (b) is a contraindication to anticoagulation, and (c) may be associated with other life-threatening conditions, such as systemic lupus erythematosus and chronic renal failure—and in the case of an effusion due to malignancy.

2. Pleuropulmonary causes of chest discomfort, often but not always with dyspnea, include (a) pulmonary embolism, (b) pneumothorax, and (c) pneumonia.

3. Gastrointestinal causes of chest discomfort can be subgrouped into those originating above or below the diaphragm. In the former category is one major condition: (a) ruptured esophagus (Boerhaave’s syndrome). In the latter category are (b) gastroduodenal ulcers and (c) acute pancreatobiliary disease (primarily acute pancreatitis and/or cholecystitis). One (or, occasionally, a combination) of these potentially life-threatening conditions should always be considered in every patient presenting with chest pain/discomfort. Obviously, not every cause can be ruled out. There is a key difference between formulating a differential and testing for the components in the differential. Clinical acumen must always guide the most appropriate workup based on limited history and emergency department physical, under pressured circumstances.

The need to consider life-threatening causes (worsts first) is not meant to imply that every imaging and diagnostic technique available is to be used in every case. Consideration of worsts first is indicated as a cognitive reflex, not as a prescription to order tests as part of defensive medicine. A strong knowledge base coupled with careful history and physical exams remain the best antidotes to what Dr. Pat Croskerry has termed base rate neglect. Under this rubric, he includes the tendency to inflate the actual prevalence of a disease to avoid missing rare but life-threatening illnesses. In addition, the initial workup by the emergency department should be discussed in a critical but positive way to learn the basis on which certain conditions were ruled out or ruled in.

Note to Attendings.

When hearing a presentation of a patient admitted via the emergency department, a useful exercise is to freeze-frame the presentation at selected points to ask the ward team: What would you have done at this point? What options were available? This strategy encourages active listening, creates suspense, and most important, compels a critical assessment of the workup in real time.

Chest Discomfort: All Things Considered

Sometimes the initial assessment of chest discomfort does not uncover a definitive cause and, most important, does not reveal evidence of one of the major life-threatening conditions noted above. In such instances it is helpful to take a step back and consider the more general set of conditions that may be associated with chest discomfort. This lengthier (but still not complete) list is summarized in Exhibit 5.2 and obviously includes all of the conditions mentioned above.

EXHIBIT 5.2 More Comprehensive Differential Diagnosis of Chest Pain: “Onion Skin” or “Mental Body Scan” Approach

^aIncludes other nonatherosclerotic causes of infarction/ischemia, such as collagen vascular disease, coronary artery dissection, coronary artery spasm syndromes (Prinzmetal’s angina/variant angina), takotsubo (stress) cardiomyopathy, cocaine-related, and congenital heart disease (e.g., anomalous left circumflex artery).

A useful strategy for thinking about the global differential diagnosis of chest discomfort—more reliable than most mnemonic devices—is to work your way from the outside of the body inward in concentric rings (onion skinning or mental body scanning) and from head downward. The causes can include everything from left-sided dermatomal discomfort with herpes zoster, to cervical disk disease with radicular pain, to vertebral compression fractures with radiation of pain in an anterior direction.

An understandable and universal goal in crafting differential diagnoses is being comprehensive and crisp, in the mode of the New England Journal of Medicine’s (NEJM’s) Clinicopathologic Conferences (CPCs) or its Clinical Problem-Solving (CPS) cases. In the best of these cases, the discussant deftly delineates and distinguishes among a myriad of possibilities and seems, with magical finesse, to hone in on the correct diagnosis with the sharpness of Ockham’s razor (Chapter 8). It is an admirable model, much envied but much less often achieved in the hurly-burly of the real world, where cases play out prospectively and are not written up “after the fact” and where more than one diagnosis may be present.

Indeed, the success of the discussants of CPCs or CPS cases in the NEJM and other journals is remarkable and primarily a testament to the extraordinary diagnostic skills of these master clinicians. But keep in mind that these discussants are operating in a rather specialized arena where cases are handpicked based on their discussant’s expertise, the cases’ uniqueness, and also because these cases have definitive answers supported by “smoking gun” evidence, typically from biopsies or postmortem studies. The revelatory climax is often preceded by the phrase: “A procedure was done. …” Then we learn that the heart failure was due to Chagas disease (American trypanosomiasis) or the splenomegaly and fever to visceral leishmaniasis contracted on a recreational trip to Central America, or the mid-right lower abdominal pain was due to carcinoid and not to appendicitis. Indeed, heart failure due to Chagas disease is much more likely to show up in a case report than in your clinic practice in North America. Similarly, the Bayesian probabilities in the emergency department or ICU, where “all-comer” cases are self-presented, are very different from the pretest likelihoods in a highly selected CPC/CPS forum.

DIFFERENTIAL DIAGNOSIS OF RIGHT UPPER QUADRANT PAIN

As with chest pain, there are many causes of right upper quadrant (RUQ) pain, and this complaint comprises only a subset of the causes of abdominal pain. Similar to chest pain, only a few are both life-threatening and treatable in the immediate term. As with chest pain, begin with anatomy: hepatobiliary system, duodenum, pancreas, right kidney, mesenteric vessels, right hemidiaphragm, and right lower lung lobe. Next, consider the possible causes in terms of pathology (i.e., vascular, inflammatory, neoplastic, traumatic, etc.). An abdominal exam is essential in trying to elucidate a cause (e.g., rebound tenderness, bowel sounds, guarding). Exhibit 5.3 is a short list of treatable “worsts firsts.” A more exhaustive list is given in Exhibit 5.4

^aIn addition, consider rarer causes, such as thoracic spine pathology (multiple myeloma, metastases, osteoarthritis with vertebral compression, tuberculosis, rheumatoid spondylitis) and adrenal gland pathology [Waterhouse–Friderichsen syndrome (hemorrhagic adrenalitis), neuroblastoma, adrenal infarct].

“n-PLUS”: A FINAL NOTE ON CRAFTING DIFFERENTIAL DIAGNOSES

You should always consider every differential diagnosis to be “n plus,” where n is the sum of all the things you are considering and plus is the possibility of other alternative(s). The latter most often is something that you had not thought of at the time. Yet sometimes, in very special cases, it is something entirely new—the discovery of a novel syndrome. More often it is a somewhat off-beat or unexpected presentation of a well-known disease process, or a true forme fruste (highly atypical or incomplete presentation of a known disease, from the French for “incomplete form”). The “n plus” rule is an example of what some authors refer to more generically as “lateral thinking” or “thinking outside the box.” We prefer n plus since you can actually add a written or at least a mental line on the differential diagnosis to “Consider other possibilities,” making this an explicit rather than an implicit part of clinical evaluation.

EXAMPLE

An otherwise healthy 53-year-old man with a history of aortic insufficiency who was 20 years status post a mechanical prosthetic aortic valve replacement presented with high fevers and left shoulder pain. The obvious concerns included bacterial endocarditis with arthropathy and/or a primary septic joint or septic shoulder bursitis. Physical examination, initial laboratory studies, including a transesophageal echocardiogram with particular attention to the prosthetic aortic valve, and blood cultures were unrevealing. About 36 hours after hospital admission, with hydration, a community-acquired left lower lobe pneumonia with referred pain to the left shoulder became apparent.

Suggested Exercise

Try constructing two-tiered differential diagnoses of some other major symptoms and signs and compare notes with your colleagues and attendings. Some common acute differentials include fever, weight loss, headache, red eye, acute visual loss, epigastric pain, shortness of breath, diarrhea, weakness, and selected electrolyte disorders, among others.

As with lab tests, expecting the unexpected is a principle that applies to formulating a differential diagnosis. While worsts first and a subsequent onion peel-back approach is a good systematic way to narrow your differential, revisiting unexpected causes, as in the example above, is often enormously helpful.

Sir Arthur Conan Doyle is best known as the author and creator of the Sherlock Holmes series. Conan Doyle was also a physician who reportedly modeled his protagonist’s deductive reasoning skills on one of his own professors at the University of Edinburgh, Dr. Joseph Bell.

One of the Sherlock Holmes stories, “Silver Blaze,” features the disappearance of a famous horse before an important race, and the apparent murder of its trainer. The following interchange between Holmes and a Scotland Yard detective, Inspector Gregory, illustrates how what does not happen can be as important as what does.

Gregory: Is there any other point to which you would wish to draw my attention?

Holmes: To the curious incident of the dog in the night-time.

Gregory: The dog did nothing in the night-time.

Holmes: That was the curious incident.

Hepatobiliary	Hepatitis syndromes Hepatic abscess Liver tumors Hepatic congestion syndromes: Severe heart failure Budd–Chiari syndrome Cholecystitis/choledocholithiasis/cholangitis
Gastroduodenal	Penetrating ulcer
Pancreatic	Pancreatitis Pancreatic cancer
Right renal	Pyelonephritis Stones Renal vein thrombosis Tumor
Right colonic	Diverticulitis Tumor Volvulus or other obstructive lesion Gallstone ileus
Diaphragmatic	Tear Abscess
Lymphatic	Lymphoma
Right pleuropulmonary	Effusion/pleurisy Pulmonary embolism Right lower lobe infection or tumor