Diagnosing Pulmonary Embolism in the ED: What’s Tops in the Toolbox?

Abstract & Commentary

Source: Perrier A, et al. Diagnosing pulmonary embolism in outpatients with clinical assessment, D-dimer measurement, venous ultrasound, and helical computerized tomography: A multicenter management study. Am J Med 2004;116:291-299.

The authors analyzed 965 consecutive patients presenting to any of three European emergency departments (EDs) between October 2000 and June 2002 with complaints suggestive of pulmonary embolism (PE)—sudden or worsening dyspnea, chest pain without another etiology, or syncope. A clinical point score was calculated according to a published prediction rule,1 including previous PE or deep vein thrombosis (DVT), recent surgery, age greater than 60, pulse greater than 100/min, pO2 less than 60 mmHg, pCO2 less than 36 mmHg, and abnormal chest x-ray. Low, intermediate, and high probability were defined. Average age was 61; females composed 58% of the cohort. Chest pain was present in 70%, dyspnea in 66%, and syncope in 7% of cases.

On initial assessment, plasma D-dimer (Dd) was measured by enzyme-linked immunosorbent assay (ELISA) technique. Patients with Dd levels less than 500 mcg/L were followed clinically. Those with Dd levels greater than 500 mcg/L underwent leg compression ultrasound (US). If DVT was detected, anticoagulation treatment was begun. If no DVT was apparent, patients underwent helical computed tomography (CT). If PE was visualized on CT, anticoagulation was initiated. Those with low or intermediate clinical probability and normal CT and US were not anticoagulated or tested further. High clinical probability patients underwent pulmonary angiography. Those with inconclusive CT scanning (mostly due to motion artifact) underwent ventilation-perfusion (V/Q) scanning.

The overall incidence of PE in this symptomatic cohort was 222/965 (23%). Of the 891 patients with low or intermediate clinical probability of PE, 275 (31%) had normal Dd levels, compared with only 5/74 (6.8%) of those with high probability point scores (p < 0.0001). US confirmed DVT in 92/685 (13.4%) cases with elevated Dd levels. Helical CT was negative for PE in 458/593 (76%) of patients with normal leg US studies, positive for PE 124/593 (21%), and indeterminate in only 11 cases, in which five had low-probability V/Q scans. Overall, PE was diagnosed in 34/522 low-probability patients (7%; 95%; CI:5-9%), 125/369 intermediate probability cases (34%; 95%; CI:29-39%), and 63/74 high-probability patients (85%; 95%; CI:75-92). Only eight cases had high-probability point scores with elevated Dd levels but negative US and CT results; at angiography, two had PE, while six were normal.

During the three-month follow-up interval, 7/685 (1.0%; 95%; CI:0.5-2.1%) without PE or anticoagulation had a subsequent DVT (n = 2) or PE (n = 5); all were low-to-intermediate probability cases with normal US and CT examinations. Nine (4.1%) of those with initial PE had recurrent PE during follow-up. Major bleeding occurred in 7/222 (3.2%) of anticoagulated patients, of whom two died. Overall three-month mortality was 7.7% in patients with PE vs. 2.7% of those without PE. Costs for diagnosis in a hypothetical cohort of 1000 patients were roughly $200,000 less for the study algorithm vs. other published analyses. The authors conclude that their sequential-testing diagnostic algorithm is precise, safe, and cost-effective for detection and exclusion of PE in outpatients in the ED.

Commentary by Michael Felz, MD

Several features of Perrier’s data are distinctive to me: Dd performed quite well as first-step testing to suggest, or exclude, DVT and PE. I wonder how soon ELISA technology (as utilized in Europe) will replace rapid, less expensive latex methodologies in widespread use in this country. Furthermore, the role of V/Q scanning was limited in this analysis. Lack of thromboembolic events during a three-month follow-up was 99% among those with PE excluded, even without anticoagulation. In my view, the instruments studied by Perrier need to be "tops in the toolbox" for all ED physicians considering PE in symptomatic patients.

Dr. Felz, Associate Professor, Department of Family Medicine, Medical College of Georgia, Augusta, is on the Editorial Board of Emergency Medicine Alert.


1. Wicki J et al. Assessing clinical probability of pulmonary embolism in the emergency ward: A simple score. Arch Intern Med 2001;161:92-97.