The efficacy of ultrasound compared with ascending venography for the detection of deep venous thrombosis immediately after total knee arthroplasty was assessed after a 2-year interval. One hundred thirty-seven patients were eligible for the study; however, 31 patients received only one of the screening methods and a color Doppler examination was inconclusive in six patients. Therefore, 100 patients had a Doppler examination and a venogram. Overall, the sensitivity of ultrasound was 85%, the specificity 97%, the positive predictive value 85%, the negative predictive value 97%, and the accuracy 95%. The sensitivity in the calf was 83%, in the popliteal vein 86%, and in the femoral vein 100%. Two years ago, the initial assessment of ultrasound for the detection of deep venous thrombosis after surgery in patients who had total joint arthroplasty revealed a 75% sensitivity, 99% specificity, 91% positive predictive value, 97% negative predictive value, and 97% accuracy. The sensitivity in the calf was 83%; the sensitivity in the popliteal vein was 40%; and the sensitivity in the femoral vein was 50%. After 2 years of using this screening test with one technician and one radiologist, an improvement with this noninvasive technique was shown. However, it was found that Doppler imaging is not as sensitive as venography for detecting calf thrombi. Any imaging technique should be validated by each institution to determine the validity of the instrument and the learning curve of the technician administering the examination.