In a prospective study, we compared real-time B-mode ultrasonography, using the simple criteria of common femoral and popliteal vein compressibility (Compression US), and Doppler ultrasound, using a standardized technique (Doppler US), with contrast venography in 158 consecutive outpatients symptomatic for deep-vein thrombosis of the lower limbs (DVT). For proximal vein thrombosis, the sensitivities documented for Compression US and Doppler US were 100% (95% CI: 90% to 100%) and 89% (95% CI: 76% to 96%), respectively. This difference is not statistically significant (p = 0.056). For all thrombi (including isolated calf-vein thrombosis), however, the sensitivity of Compression US was significantly higher than that of Doppler US (95% and 76%, respectively; p < 0.04). Compression US was normal in all patients with normal venogram (specificity, 100%; 95% CI: 95% to 100%), while Doppler US was abnormal in two patients with normal venogram (specificity, 98%; 95% CI: 92% to 100%). The specificities of the two tests did not differ significantly. The results of our comparison suggest that Compression US is superior to Doppler US in the detection of DVT in symptomatic outpatients.