CT angiography in the evaluation of acute pulmonary embolus

OBJECTIVE. The purpose of this study was to assess the appropriate use of CT angiography (CTA) in the diagnostic evaluation of acute pulmonary embolism (PE). MATERIALS AND METHODS. We reviewed a total of 575 CT angiograms obtained to evaluate for PE at a large level 1 trauma teaching hospital from January 2004 through March 2005. Various clinical settings were used for 267 inpatient (46%), 258 emergency department (45%), and 50 outpatient (9%) studies. We excluded CTA performed for other reasons, repeated CTA, and patient records with incomplete clinical data. On the basis of chart review in which the investigators were blinded to final diagnoses, pretest probability of PE according to the Wells criteria was retrospectively assigned to each patient. D-dimer values, when obtained, also were reviewed. The diagnosis of PE was based on final CTA reports. RESULTS. PE was diagnosed in 9.57% of 575 patients. Positivity rates by location were 32 (12%) of the 267 inpatients, 22 (8.5%) of the 258 emergency department patients, and one (2.0%) of the 50 outpatients. Three (< 1%) of the 575 patients had high probability of PE, even though 351 patients had gone directly to CTA. Of the other 572 patients, 158 (28%) had intermediate and 414 (72%) low probability of PE. In the high, intermediate, and low probability groups, two (67%), 24 (15%), and 29 (7%), respectively, of the patients had PE. A D-dimer assay was performed for 224 (39%) of the 575 patients. Thirty-nine (17%) of the 224 patients had normal results (< 0.5 mu g/mL); 107 (48%), intermediate results (0.6-2.0 mu g/mL); and 78 (35%), abnormal results (> 2.0 mu g/mL). In the emergency department cohort, 151 (59%) of 258 patients underwent a D-dimer assay. Thirty-two (21%) of the 151 patients had normal results; 81 (54%), intermediate results; and 38 (25%), abnormal results. Only one patient with a normal D-dimer level and three patients with intermediate D-dimer levels had PE, the equivalent of 3% of each group. The number of CTA examinations ordered for patients with normal and intermediate D-dimer results was 146 (25% of the 575 total studies). Twenty-two (8%) of the 258 emergency department patients had PE, and clinical suspicion of PE was high for 11 (50%), intermediate for 10 (45%), and low for one (5%) of those patients. CONCLUSION. Our data showed suboptimal use of the Wells criteria and subjective overestimation of the probability of PE before ordering of CTA. Although a definitive acceptable PE positivity rate for CTA has not been established, the 10% yield represents overuse of CTA as a screening rather than a diagnostic examination.