[Diagnosis of deep venous thrombosis of the legs: accuracy of ultrasonography using vein compression].

We prospectively compared real-time US findings with venographic results in the legs of 171 consecutive outpatients with clinically suspected deep venous thrombosis (DVT). In each leg the common femoral and the popliteal veins were evaluated with venography and US. The two examinations were independently performed and reviewed. Vein compressibility (VC), intraluminal echogenicity and response to Valsalva maneuver were evaluated with US. Venography detected DVT in 54/171 legs. 10/54 legs had isolated distal DVT. With VC, US was 100% specific, 87% sensitive and 96% accurate. 6/7 false negative US studies were due to isolated distal DVT; therefore US sensitivity for proximal DVT was 98%. The other two US diagnostic criteria were much less accurate than VC. VC-US is an accurate, cheap and reproducible test for the detection of DVT in symptomatic outpatients.