To compare the sensitivity, specificity, and predictive values of compression ultrasonography (US) in postoperative orthopedic patients with those of (a) impedance plethysmography in postoperative patients and (b) compression US in symptomatic outpatients, the authors performed an investigator-blinded cohort study. One hundred thirty-four consecutive inpatients who had undergone elective knee-replacement surgery or surgery for a fractured hip and 65 consecutive outpatients with clinically suspected venous thrombosis who had undergone venography were evaluated. Compression US allowed detection of 11 of 21 (52.4%) proximal-vein thrombi but was insensitive to calf-vein thrombi in the orthopedic patients. Compression US had a significantly greater specificity and positive predictive value than impedance plethysmography for all thrombi in orthopedic patients; compression US also had greater sensitivity. The sensitivity of compression US for proximal-vein thrombi was significantly higher (92.1%) in symptomatic outpatients than in orthopedic patients. The authors conclude that compression US has significant advantages over impedance plethysmography in the detection of proximal-vein thrombi in patients who have undergone hip- or knee-replacement surgery.