Impedance plethysmography using the cuff technique has been compared with venography, in 346 consecutive patients with suspected venous thromboembolism. The limbs were classified according to the venographic results as no thrombosis, proximal (popliteal, femoral, or iliac) vein thrombosis, and calf thrombosis. A discriminant analysis was performed. The impedance plethysmographic result was normal in 386 of 397 limbs which were normal on venography, a specificity of 97%, and abnormal in 124 of 133 limbs which showed proximal vein thrombosis, a sensitivity of 93%. Seventy three of 88 limbs wih calf vein thrombi had a normal impedance plethysmographic result. The sensitivity in 29 limbs with asymptomatic proximal vein thrombosis was 83%. Impedance plethysmography is an accurate method for detecting proximal vein thrombosis but had limitations which include the possibility of false positive results due to arterial insufficiency and muscle tension.