Diagnosis of deep venous thrombosis by clinical signs and symptoms is unreliable, but contrast venography is relatively expensive and invasive. We therefore evaluated the use of impedance plethysmography as a noninvasive alternative in 426 consecutive outpatients with clinically suspected acute deep venous thrombosis. Four sequential impedance plethysmograms were obtained on days 1, 2, 5, and 10 of the study. In 289 patients (68 percent), the results of all four studies were normal, and these patients were not treated with anticoagulants. One of these patients may have had a minor pulmonary embolus during the 10-day study period. During a six-month follow-up of all patients, none of the 289 patients whose plethysmograms were normal died of venous thromboembolism or presented with suspected pulmonary embolism. In 137 patients (32 percent), the impedance plethysmograms were abnormal; 117 (85 percent) had the abnormal results on their first test, and 20 (15 percent) had them on subsequent tests. All patients with abnormal plethysmograms also underwent contrast venography, which confirmed the diagnosis of deep venous thrombosis in 92 percent. We conclude that the diagnostic accuracy of repeated impedance plethysmography compares favorably with that of venography and that the technique is a safe and effective noninvasive approach to the diagnosis and care of outpatients with clinically suspected acute deep venous thrombosis.