For this study on the incidence and prevention of deep-vein thrombosis, we examined the data on 517 patients with 638 total knee replacements. All of the patients had postoperative venograms and 475 had postoperative perfusion lung scans. We collected data on known risk factors, tourniquet time, knee deformity, postoperative rehabilitation, and methods of prophylaxis, and performed extensive statistical evaluation. Forty-nine patients inadvertently did not receive prophylaxis, and in forty-one (84 per cent) of them ipsilateral deep-vein thrombosis developed. The incidence of ipsilateral thrombosis was 57 per cent in the 468 patients who did receive some form of prophylaxis. Ipsilateral thrombosis in the popliteal veins or thigh was seen in 11 per cent of the patients with unilateral total knee replacement, and contralateral thrombosis was noted in 3 per cent. Bilateral total knee replacement was associated with a 58 per cent incidence of ipsilateral deep-vein thrombosis in the calf and a 14 per cent incidence in the thigh. Pulmonary embolism was diagnosed clinically in nine patients (1.7 per cent), but was suggested on perfusion lung scans in thirty-nine patients (7 per cent). Twelve patients (2.3 per cent) received formal anticoagulant therapy. In no patient was the pulmonary embolism fatal. No specific high-risk population was identified. While no one prophylactic regimen was proved to be more effective than another in our series, we think that prophylactic measures should be part of the management of patients undergoing total knee replacement.