Compression sonography in patients with indeterminate or low-probability lung scans: lack of usefulness in the absence of both symptoms of deep-vein thrombosis and thromboembolic risk factors.

OBJECTIVE: We sought to determine whether compression sonography could be eliminated in the evaluation of patients who lacked both thromboembolic risk factors and symptoms of deep-vein thrombosis and who had an indeterminate or low-probability lung scan. MATERIALS AND METHODS: The medical records of 155 consecutive patients who underwent bilateral lower-extremity sonography after an indeterminate or low-probability lung scan were reviewed. The presence of thromboembolic risk factors and deep-vein thrombosis symptoms and the result of sonography were recorded. Patients were divided into two groups: group 1 consisted of patients with either thromboembolic risk factors or deep-vein thrombosis symptoms, and group 2 consisted of patients without thromboembolic risk factors and without deep-vein thrombosis symptoms. The incidences of deep-vein thrombosis in groups 1 and 2 were compared by use of a two-tailed Fisher's exact test. RESULTS: Thromboembolic risk factors or deep-vein thrombosis symptoms were found in 109 of 155 patients (70%) (group 1). Deep-vein thrombosis was found in nine of 108 patients (8%) in group 1. Both thromboembolic risk factors and deep-vein thrombosis symptoms were absent in 47 of 155 patients (30%) (group 2). Deep-vein thrombosis was found in none of 47 patients (95% confidence interval, 0-8%) in group 2. The difference in the incidences of deep-vein thrombosis in groups 1 and 2 approached statistical significance (p = .0579). The negative predictive value of the absence of both thromboembolic risk factors and deep-vein thrombosis symptoms in excluding deep-vein thrombosis was 100% (95% confidence interval, 93-100%). CONCLUSION: If both symptoms of deep-vein thrombosis and thromboembolic risk factors are absent, the usefulness of lower-extremity sonography in detecting deep-vein thrombosis in patients with an indeterminate or low-probability lung scan is low. The manner in which these findings may be used to modify individual practice patterns will undoubtedly depend on the rate of detection of deep-vein thrombosis at a given institution.