A single center, prospective, epidemiologic study was conducted to estimate the incidence of deep venous thrombosis detected by venography in patients in Japan undergoing total hip arthroplasty or total knee arthroplasty without prophylactic anticoagulant therapy. Venograms of 164 patients who had total hip arthroplasty and 138 patients who had total knee arthroplasty were evaluated. The incidences of deep venous thrombosis were 22.6% in patients who had total hip arthroplasty and 48.6% in those who had total knee arthroplasty. The incidences of proximal deep venous thrombosis were 9.8% in patients who had total hip arthroplasty and 14.5% in those who had total knee arthroplasty. Statistical analysis revealed that the type of operation influenced the development of deep venous thrombosis. Patients who had total knee arthroplasty were 3.2 times more likely to have deep venous thrombosis develop than were patients who had total hip arthroplasty. Body mass index and age were identified as statistically significant risk factors.