Contrast venography: reassessment of its role.

To compare contrast venography with noninvasive methods, 353 patients clinically suspected of having deep venous thrombosis were examined with venography and independently with combined Doppler flow sounds and plethysmography. Noninvasive examinations had a sensitivity of 96% and a specificity of 90%. Positive noninvasive tests had a 94% predictive value, and negative noninvasive tests had a 93% predictive value. The overall accuracy of the noninvasive tests was 94% (331 of 353) compared with venography. Since venography itself may be subject to misinterpretation, noninvasive examinations should be the preferred initial method for diagnosing deep venous thrombosis. Venography should be reserved for situations that require additional diagnostic confirmation.