A study of 86 patients undergoing total knee arthroplasty (TKA) was performed to evaluate the role of cold compression. The patients were treated with cold compression or epidural analgesia for 3 days after TKA. Pain was measured on a visual analog scale, and total consumption of analgesics was recorded. Range of movement (ROM) was recorded before TKA until 3 weeks postoperatively. Weight bearing, blood loss, and time in hospital were recorded. Visual analog scale scores and analgesic consumption were equal in both groups. Range of movement at discharge was 75 degrees in the cold compression group vs 63 degrees in the control group. By 3 weeks' follow-up, ROM was 99 degrees vs 88 degrees. Mean Hb values averaged 120 mmol/L in the cold compression group vs 109 mmol/L in the control group after surgery. Mean time in hospital of patients with cold compression averaged 4.8 days vs 6.2 days in the control group. The study shows that cold compression therapy improves the control of pain and might thus lead to improvement in ROM and shorter hospital stay.