We retrospectively reviewed the records of 2050 arthroscopic knee surgeries performed at The Orthopedic Specialty Hospital from January 1993 to December 1994. The number of clinically detected deep venous thromboses, with confirmation by duplex ultrasonography, was determined. Prospectively, preoperative and postoperative duplex ultrasonographic images were completed on 239 patients divided into 2 groups: those undergoing nonligament, intraarticular arthroscopic surgery (N = 131) and those undergoing arthroscopically assisted ligament surgery and extraarticular or osteotomy surgery (N = 108). For the retrospective study, the incidence of deep venous thrombosis was 0.24%. Prospectively, seven total deep venous thromboses were identified (rate, 2.9%), with five being identified within 8 days of surgery in asymptomatic patients (rate, 2.1%). There were no statistically significant associations or correlations between the development of deep venous thrombosis and patient personal data or surgical variables, respectively. The difference in the rate of deep venous thrombosis between the two prospective groups was not statistically significant; however, patients who had more invasive surgery tended to be at higher risk for developing deep venous thrombosis. A cost-benefit analysis did not support the routine use of duplex ultrasonography to detect deep venous thrombosis in patients undergoing arthroscopic knee surgery.