Clinical Question: Can physicians rely on a negative computed tomography (CT) scan to rule out suspected pulmonary embolism (PE)?
Setting: Various (meta-analysis)
Study Design: Systematic review
Synopsis: Investigators thoroughly searched multiple databases—including MEDLINE, the Cochrane Registry of Controlled Trials, and Science Citation Index—and relevant journals for English language articles that met the following selection criteria: use of contrast-enhanced CT as the initial triage test to rule out the diagnosis of acute PE, an appropriate clinical follow-up of at least three months, and a prospective design. The standard to establish the validity of testing to rule out PE was the rate of subsequent venous thromboembolic events after anticoagulation therapy was withheld. Two reviewers independently abstracted data, and a third party arbitrated discrepancies. Of 22 studies found during the initial search, 15 studies that evaluated a total of 3,500 patients met the minimum inclusion criteria. Seven of these met the criteria for level 1 diagnostic studies.
Three different CT modalities were evaluated: single detector row helical CT, multiple detector row helical CT, and electron-beam CT. Patient follow-up ranged from three to 12 months. The overall negative likelihood ratio of a venous thromboembolic event after a negative CT scan for PE was 0.07 (95% confidence interval, 0.05 to 0.11). There was no significant difference in the risk of a subsequent venous thromboembolic event based on the type of CT modality used. Compared with studies that used chest CT imaging only, the risk of subsequent venous thromboembolic events in studies that used additional imaging tests before chest CT was not reduced significantly. The reported negative likelihood ratio in this analysis compares favorably with that reported for pulmonary angiography (Henry JW, et al. Continuing risk of thromboemboli among patients with normal pulmonary angiograms. Chest May 1995;107:1375–8). A formal analysis found no evidence for significant publication bias, but there was some minimal heterogeneity among the results of the various trials.
Bottom Line: A negative CT scan is as accurate as pulmonary angiography in ruling out suspected PE. Physicians should strongly consider using clinical decision rules to accurately assess the pretest probability of PE in an individual patient, and then interpret diagnostic test results in light of this probability. For example, a negative CT scan in a low-risk patient rules out PE, whereas a negative CT scan in a high-risk patient may require further confirmation. (Level of Evidence: 2a–)