Pulmonary embolism diagnosis: Remember the history and physical exam

Abstract

Thromb Haemost 2009; 101: 7–8 The trickiest part of detecting acute pulmonary embolism (PE) is remembering to consider PE as a diagnostic possibility. But this does not mean that every patient in whom the diagnosis is considered should undergo chest computed tomography (CT) scanning with contrast or ventilation-perfusion lung scanning. Such an approach would cram our CT and Nuclear Medicine imaging laboratories with hundreds of thousands of unnecessary tests, expose patients to unnecessary irradiation and contrast agent, precipitate cases of renal failure and anaphylaxis from contrast, and slow our clogged Emergency Departments to a virtual standstill. Ordering a chest CT with contrast as a reflex, as soon as the diagnosis of PE is added to the differential diagnosis, is becoming common practice even though it is not the ideal way to practice medicine. The motivations for this wayward approach may be difficult to figure out. Is it the new way to practice “defensive medicine” and avert the charge of medical malpractice as our society becomes more litigious? Is it a mistaken belief that imaging almost anyone with possible PE will lead to a hastier or more efficient workup? Is it a genuine belief that we can best advocate for our patients by ordering PE imaging even when the likelihood of PE is rare, so that we “leave no stone unturned”? A more orderly and thoughtful approach to the PE work-up adheres to classical teachings and tradition. Namely, take a directed history, perform a relevant physical examination, establish a differential diagnosis, and then decide upon D-dimer testing versus PE imaging based upon the likelihood of finding PE. If the likelihood of PE is low, then order a D-dimer test and rule out PE if the D-dimer test is normal. If the likelihood of PE is not low, or if the D-dimer is elevated, then PE imaging usually makes sense. But in most Emergency Departments, the majority of patients in whom the diagnosis is considered have a low clinical probability of PE. How do we standardize our definition of “PE clinical probability”? The best way, proven time and time again in clinical trials, is to use the Wells criteria for clinical probability assessment of PE (1). These criteria are used successfully in clinical trials. But in the real world of daily clinical medicine, Wells criteria are almost never employed to make clinical decisions. It is unclear why the Wells criteria are not utilized. Each criterion requires a clinical skill set, as shown in Table 1. And there exists a broad consensus that clinical evaluation is important, as well as maintaining the fundamental approach of history, physical examination, and differential diagnosis. The Wells criteria are awkward to employ. There are three different point scores assigned to the seven criteria: 1 point for two criteria, 1.5 points for three criteria, and 3 points for two criteria. And when summing the point score, it is difficult to remember how many points constitute low probability, intermediate probability, and high probability for PE. So, we have a set of clinical probability criteria that work but that are rarely used outside of clinical trials. Simplification of the Wells criteria is in order (2). And validation of the simplification, even as done retrospectively in the current study by Douma et al. (3), provides reassurance and guidance.