Ruling out pulmonary embolism in patients with high pretest probability

Murtaza Akhter*, Jeffrey Kline, Bikash Bhattarai, Mark Courtney, Christopher Kabrhel

*Corresponding author for this work

Research output: Contribution to journalArticle

Abstract

Introduction: The American College of Emergency Physicians guidelines recommend more aggressive workup beyond imaging alone in patients with a high pretest probability (PTP) of pulmonary embolism (PE). However, the ability of multiple tests to safely rule out PE in high PTP patients is not known. We sought to measure the ability of negative computed tomography pulmonary angiography (CTPA) along with negative D-dimer to rule out PE in these high-risk patients. Methods: We analyzed data from a previous prospective observational study conducted in 12 emergency departments (ED). Wells score criteria were entered by providers before final PE testing. PE was diagnosed by imaging on the index ED visit, or within 45 days, demonstrating either PE or deep vein thrombosis (DVT), or if the patient died of PE during the 45-day, follow-up period. Testing threshold was set at 1.8%. Results: A total of 7,940 patients were enrolled and tested for PE, and 257 had high PTP (Wells >6). Sixteen of these high-risk patients had negative CTPA and negative D-dimer, of whom two were positive for PE (12.5% [95% confidence interval (2.2%-40.0%)]). One of these patients had a DVT on CT venogram and the other was diagnosed at follow-up. Conclusion: Our analysis suggests that in patients with high PTP of PE, neither negative CTPA by itself nor a negative CTPA plus a negative D-dimer are sufficient to rule out PE. More aggressive workup strategies may be required for these patients

Original languageEnglish (US)
Pages (from-to)487-493
Number of pages7
JournalWestern Journal of Emergency Medicine
Volume19
Issue number3
DOIs
StatePublished - May 2018

ASJC Scopus subject areas

  • Emergency Medicine

Fingerprint Dive into the research topics of 'Ruling out pulmonary embolism in patients with high pretest probability'. Together they form a unique fingerprint.

Cite this