EM Nerd-An Addendum to The Case of the Incidental Bystander

The October 2016 NEJM publication of the PESIT trial by Prandoni et al set the world of social media ablaze (1). There were the standard fear mongering headlines noting that 1 in 6 patients admitted to the hospital following syncope had a pulmonary embolism. This was followed by a reactionary flurry of commentary by the FOAM community explaining why these results were likely a misrepresentation a far more benign reality. But until recently we have had no attempt to externally validate Prandoni et al’s findings.

Published as a research letter in JAMA Internal Medicine on May 8^th 2017, Verma et al conducted a retrospective analysis of the GEMINI database identifying patients admitted to the hospital after a syncopal event (2). The authors attempted as best they could to replicate the inclusion criteria used in the PESIT trial, including patients in their final analysis with an ICD-10 code of syncope in the Emergency Department or at hospital discharge. They excluded patients who were receiving anticoagulation prior to admission, had a history of prior pulmonary emboli, or were pregnant.

The authors examined how often an investigation for a venous thromboembolism (VTE) (defined by D-dimer assay, compression US, CTPA, or V/Q scan) occurred in this population. Of the 1305 patients admitted to the hospital for syncope included in the final analysis, 146 patients (11.2%) underwent at least one investigation for VTE during their inpatient course. PE was diagnosed in 11 of the 73 (15.1%) patients who received a CTPA or V/Q scan and a DVT was identified in 10 of the 67 (14.9%) patients who underwent compression US. The authors reported the overall prevalence was 1.4%.

These results are in stark contrast to results of the PESIT cohort who received diagnostic imaging for PE in 41.9% of the cohort and identified an acute clot in the pulmonary vasculature in 17.1%. And while Verma et al’s publication is limited by its retrospective methodology, its results are fairly similar to previous studies examining this question. And so how do we reconcile these drastically differing results? Likely these studies represent the same underlying truth, viewed from two drastically different perspectives.

The PESIT trial was a well done prospective, multicenter trial and yet its findings are not only incongruent with our shared clinical experience, but also with the remainder of the literature examining the epidemiology of syncope. Either we have a complete misconception of syncope and its causes, or the findings presented by the PESIT authors represent a cautionary tale of overdiagnosis.

The PESIT trial represents a specific diagnostic strategy taken to an extreme. Simply put, if one is to go looking for pulmonary emboli in any population they will most certainly identify a cohort of patients with pulmonary emboli. In fact, the more often one engages in this diagnostic endeavor, the more often they will be rewarded by their efforts (3,4). But a pulmonary embolism diagnosed on imaging does not directly translate into the morbidity that is traditionally thought to be associated with pulmonary embolism. A number of studies have demonstrated that neither the size of the clot nor the location of the clot are independent predictors of death or morbidity, but rather that patient level factors are far more important predictors of poor outcomes (5,6). And so it is important to know the clinically milieu that surrounds any radiologically confirmed PE.

In the case of the Verma et al study, this was not a prospective effort. Rather the authors performed a retrospective analysis of patients admitted to the hospital following a syncopal event. As such, only patients in whom the treating clinician chose to pursue the diagnosis of VTE underwent any diagnostic testing. This is in direct contrast the PESIT trial, where all patients admitted for syncope underwent a workup for pulmonary embolism independent of the presence or absence of any clinical symptoms. And so what the PESIT trial represents is the natural history of empirically applied screening protocol in all patients admitted for syncope independent of their clinical presentation. While the Verma et al cohort and earlier data sets are likely a more accurate representation of the true prevalence of clinically important thrombotic disease in a similar cohort.

Anecdote is often considered a dirty word in scientific circles. It is often used in vain to emotionally manipulate providers into making decisions that are not otherwise supported by the literature. As such, it is easy to discredit these stories as scientifically unhelpful. But in some ways these stories provide us with the context surrounding the radiological evidence of thromboembolic disease. In fact, without clinical context to guide our diagnostic inquiries, it is impossible to determine the clinical relevance of any radiologically diagnosed PE.

Sources Cited:

Prandoni, Paolo et al. Prevalence of Pulmonary Embolism among Patients Hospitalized for Syncope. New England Journal of Medicine,375,16;1524-1531 2016
Verma AA, Masoom H, Rawal S, Guo Y, Razak F, for the GEMINI Investigators. Pulmonary Embolism and Deep Venous Thrombosis in Patients Hospitalized With SyncopeA Multicenter Cross-sectional Study in Toronto, Ontario, Canada. JAMA Intern Med. Published online May 08, 2017
Ritchie G, McGurk S, McCreath C, Graham C, Murchison JT. Prospective evaluation of unsuspected pulmonary embolism on contrast enhanced multidetector CT (MDCT) scanning. Thorax2007;62:536-40.
Schultz DJ, Brasel KJ, Washington L, Goodman LR, Quickel RR, Lipchik RJ, et al. Incidence of asymptomatic pulmonary embolism in moderately to severely injured trauma patients. J Trauma2004;56:727-3
Den exter PL, Van es J, Klok FA, et al. Risk profile and clinical outcome of symptomatic subsegmental acute pulmonary embolism. Blood. 2013;122(7):1144-9.
Jain CC, Chang Y, Kabrhel C, et al. Impact of Pulmonary Arterial Clot Location on Pulmonary Embolism Treatment and Outcomes (90 Days). Am J Cardiol. 2017;119(5):802-807.