A prospective study of hand-held Doppler ultrasonography by emergency physicians in the evaluation of suspected deep-vein thrombosis.

Abstract

A prospective study was conducted in the emergency departments of two community hospitals to assess the accuracy of Doppler ultrasonography performed by emergency physicians in the evaluation of patients with symptoms suggestive of deep-vein thrombosis. Seventy-eight extremities in 76 patients were evaluated by both Doppler ultrasonography and contrast venography. Thirty-six Doppler examinations were normal, 26 were abnormal, and 15 were equivocal. Of the unequivocal examinations, three were false-negative and nine were false-positive. These data represent a sensitivity of 85%, specificity of 79%, negative predictive value of 92%, and positive predictive value of 65%. Negative predictive value improved to 100% among a subgroup of low-risk patients, and positive predictive value improved to 80% among a subgroup of high-risk patients. Our study suggests that Doppler ultrasonography is highly predictive of a negative venogram, particularly in the low-risk patient. Further study with a larger population is needed. The availability and noninvasiveness of Doppler ultrasonography make it a useful adjunct in the evaluation of suspected deep-vein thrombosis in the ED.