• We undertook a prospective study to ascertain the accuracy of the noninvasive venous vascular laboratory in diagnosing clinically suspected deep-vein thrombosis. Of 88 patients studied with contrast venography and bilateral plethysmography, 41% had normal venograms; 15%, isolated venous thrombosis in the calf; and 44%, proximal deep-vein thrombosis. Clinical and Doppler examinations could not differentiate these three groups. The composite venous score was reliable at extreme values. By considering scores of 3.9 or less as negative, 8.0 or more as positive, and 4.0 to 7.9 as indeterminate, a suitable combination of sensitivity (75%), specificity (100%), and overall accuracy (90%) was obtained. Low-risk patients with scores of 3.9 or less should be observed. Patients with scores of 8.0 or greater can undergo anticoagulation without contrast venography, which is mandatory for patients with indeterminate venous scores and high-risk patients with scores of 3.9 or less.
(Arch Surg 1983;118:1024-1028)