The incidence of pulmonary embolism (PE) in osteoarthritic patients prophylaxed with low-dose coumadin after cemented total knee arthroplasty (TKA) was investigated prospectively. Each patient had a preoperative perfusion scan and a ventilation-perfusion scan on the seventh postoperative day. Pulmonary embolism was diagnosed by a high probability ventilation-perfusion scan or positive arteriogram. Patients with a moderate probability scan had an arteriogram to rule out PE. Pulmonary embolus was identified in 48 (5.6%) of 852 TKAs in 755 patients. Of these, six (0.7%) were symptomatic, and no fatal PE was identified. Age, gender, and weight did not show statistical differences comparing the PE and non-PE groups, nor did the incidences of previous PE, contralateral phlebitis, malignancy, and diabetes. A history of ipsilateral phlebitis increased the risk of PE from 5.2% to 13%, and a history of cardiac disease decreased the risk from 7.8% to 4.2%. Type of anesthesia, blood loss, tourniquet time, and prosthesis type were not significant factors. With the exception of previous contralateral phlebitis, traditional risk factors for PE were not found to increase risk of PE with low-dose coumadin prophylaxis. Spinal anesthesia that has been shown to be protective in total hip surgery was not a significant factor in this study.