OBJECTIVE To determine the diagnostic criteria (phase I) and to assess the accuracy (phase II) of an objective Doppler-Valsalva pressure method as compared with contrast venography for the diagnosis of acute deep-leg-vein thrombosis in symptomatic outpatients. METHODS A two-phase prospective study in consecutive patients. Doppler ultrasound strip-chart recordings and venograms were independently analyzed by experienced observers. METHODS Referral-based medical clinics at university medical centers. METHODS One hundred and ten (phase I) and one hundred and fifty-five (phase II) patients who had clinically suspected venous thrombosis and were referred by their general practitioners were included. METHODS A normal Doppler test result was defined as a cyclic spontaneous signal (S-signal), a continuous S-signal with a Valsalva pressure of less than 6.5 mm Hg, or an absent S-signal with flow after cessation of the Valsalva maneuver. A continuous S-signal with a Valsalva pressure of 6.5 mm Hg or more or an absent S-signal without flow after cessation of the Valsalva maneuver were defined as abnormal test results. The accuracy indices for proximal vein thrombosis in phase II (155 patients; prevalence, 31%) were sensitivity, 91% (95% CI, 79% to 98%), and specificity, 99% (CI, 97% to 100%). All 3 patients with isolated calf-vein thrombosis had normal Doppler test results. CONCLUSIONS The objective Doppler method is an accurate, reproducible, and simple method for detection of venous thrombosis in symptomatic outpatients.