The authors report the results of a prospective study examining the benefits of daily continuous passive motion combined with physical therapy, compared with physical therapy alone, in 103 consecutive osteoarthritic patients undergoing primary total knee arthroplasty. The first 51 patients received continuous passive motion initiated in the recovery room and the next 52 patients did not receive continuous passive motion. Both groups underwent an identical physical therapy protocol starting on the first postoperative day. At discharge, there was a significant increase in active flexion in the continuous passive motion group. There were no significant differences regarding pain, wound healing, knee swelling, wound drainage, pulmonary embolism, or length of hospital stay between the 2 groups. At 2 years, there were no clinical differences in the motion or knee scores. Knee manipulation was done for < 50 degrees flexion after the tenth postoperative day. There were 5 manipulations in the noncontinuous passive motion group and none in the continuous passive motion group. The entire costs associated with the 5 manipulations was $48,274 or $937 per patient not receiving continuous passive motion. The average daily inpatient rental of the machine was $60 per day. Continuous passive motion is efficacious in increasing short-term flexion and decreasing the need for knee manipulation without increasing costs.