Diagnostic Strategies Based upon Outcome Data in Pulmonary Embolism

Abstract & Commentary

Source: Musset D, et al. Diagnostic strategy for patients with suspected pulmonary embolism: A prospective multicentre outcome study. Lancet 2002;360:1914-1920.

This prospective, multicenter, outcome study was designed to evaluate a diagnostic strategy based on clinical pretest probability, spiral computed tomography (CT), and venous compression ultrasonography (US) of the legs in patients with suspected pulmonary embolism (PE). The main objective was to assess the safety of withholding anticoagulant treatment in patients with low or intermediate clinical probability of PE and negative findings on spiral CT and venous US.

After reasonable exclusions, 1041 consecutive inpatients and outpatients with clinically suspected PE were entered during a 19-month period in 14 centers in France. All patients were assessed for a pretest likelihood of PE (high, intermediate, or low) and then underwent spiral CT and bilateral US of the legs within 24 hours. Patients with negative spiral CT and US who were clinically assessed as having a low or intermediate clinical probability were left untreated. Patients shown to have either deep-vein thrombosis (DVT) on US, or PE on spiral CT, received anticoagulant therapy. Patients with a high pretest probability of PE, but negative spiral CT and US, underwent ventilation perfusion (VQ) lung scanning, pulmonary angiography, or both.

If either the US or the spiral CT was nondiagnostic and the other study normal, or if only isolated subsegmental emboli were seen on the spiral CT, the patient was classified as inconclusive and underwent VQ scanning, pulmonary angiography or both. All patients were followed for three months, with telephone interviews at one and two months and a clinic visit at three months.

Follow-up was completed for 98.8% of the study population. PE was diagnosed in 360 (34.6%) patients. Of patients with PE, 55 (15 %) had positive US despite negative spiral CT. Of 601 patients with negative spiral CT and US, 525 were classified as having low or intermediate pretest probability and 507 of them were not treated. Of these patients, nine (1.8%, 95% CI 0.8-3.3%) experienced DVT or PE during the three-month follow up: three patients had DVT, one PE, and five more deaths were judged as possibly related to PE. The remaining 76 patients were classified as having high pretest probability of PE, and VQ scanning or angiography was positive in four (5.3%, 95%CI 1.5-13.1%). The diagnostic strategy proved inconclusive in 95 patients (9.1%). In this group, PE was confirmed in eight patients (9.6%) by VQ or angiography.

Commentary by Stephanie B. Abbuhl, MD, FACEP

This is the first large outcome study to look at the strategy of clinical probability of PE, spiral CT, and US to diagnose PE and is a major contribution to the evolving literature in this area. The authors conclude that withholding anticoagulant therapy is safe in patients with a low or intermediate pretest probability, and a negative spiral CT and US. In this group, the rate of PE/DVT at follow-up was 1.8%, a rate similar to other major trials that have looked at outcomes using various diagnostic strategies. It is important to note that the safety of this strategy was most convincing in outpatients, in whom the follow-up PE/DVT rate was only 0.8% (95% CI 0.2-2.3), as compared to inpatients, with a rate of 4.8% and a wide 95% CI (1.8-10.1). While the safety of excluding PE by negative spiral CT and US in hospitalized patients is not confirmed by this study and will require further investigation, in the ED, this strategy now has persuasive outcome data to support it.

In patients with the diagnosis of PE, 15% had a positive US despite a negative spiral CT, emphasizing the need to combine US with spiral CT to safely rule out PE. This is analogous to previously studied strategies where nondiagnostic VQ scans have been combined with US to safely withhold anticoagulation. Of note, a significant limitation was that only single-detector-row spiral CT was used in this study. Newer generation multi-detector CT scanners will need to be studied to evaluate their impact on similar outcome studies.

One of the major limitations to the study was the rate of inconclusive results (9.1%) necessitating VQ or angiography. Only 12 patients (3.3% of all patients with detectable thrombi) had isolated subsegmental filling defects and negative US, three of whom had PE confirmed, the others had a normal VQ or angiogram, were not anticoagulated, and had no events at follow-up. Thus, this study does not resolve the ongoing controversy surrounding isolated subsegmental filling defects.

Also of interest is the fact that the US was done only once and not serially, as has been done in other outcome studies. One wonders if any of the deaths in the patients who were not anticoagulated could have been prevented with a second US done within one week from the initial exam. Since d-dimers were not included in the strategy, this study provides no information about how to incorporate d-dimers into this approach. Finally, it is important to remember that CT involves contrast injection and radiation and many patients will need other strategies because of renal insufficiency, contrast allergy or pregnancy. 

Dr. Abbuhl, Medical Director, Department of Emergency Medicine, The Hospital of the University of Pennsylvania; Associate Professor of Emergency Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA, is on the Editorial Board of Emergency Medicine Alert.