Prospective study of color duplex ultrasonography compared with contrast venography in patients suspected of having deep venous thrombosis of the upper extremities

Background: The optimal strategy for diagnosis of deep venous thrombosis (DVT) is less well established for the upper extremities than for the lower extremities. Duplex color ultrasonography can be difficult to perform in the upper extremities because of their anatomy, and contrast venography is often indicated. Moreover, limited data exist on the use of duplex color ultrasonography in this setting. Objective: To determine the accuracy of duplex ultrasonography for diagnosis of DVT of the upper extremities. Design: Prospective study of duplex ultrasonography compared with venography. Setting: A teaching hospital in Amsterdam, the Netherlands. Patients: 126 consecutive inpatients and outpatients with suspected DVT of the upper extremities. Measurements: Contrast venography was obtained after duplex ultrasonography and was judged independently. A three-step protocol, involving compression ultmonography, color ultrasonography, and color Doppler ultrasonography, was used. Sensitivity, specificity, and likelihood ratios for ultrasonography as a whole were calculated. The independent value of each step was assessed. Results: Venography and ultrasonography were not feasible in 23 of 126 patients (18%) and 1 of 126 patients (0.8%), respectively. Results of ultrasonography were inconclusive in 3 patients. Venography demonstrated thrombosis in 44 of 99 patients (44%); in 36 patients (36%), thrombosis was related to intravenous catheters or malignant disease, Sensitivity and specificity of duplex ultrasonography were 82% (95% Cl, 70% to 93%) and 82% (Cl, 72% to 92%), respectively. Venous incompressibility correlated well with thrombosis, whereas only 50% of isolated flow abnormalities proved to be thrombosis-related. Conclusions: Duplex ultrasonography may be the method of choice for initial diagnosis of patients with suspected thrombosis of the upper extremities. However, in patients with isolated flow abnormalities, contrast venography should be performed