A prospective randomized study was performed to evaluate the role of cold compressive dressings in the postoperative treatment of total knee arthroplasty (TKA). Eighty consecutive unilateral and ten bilateral primary total knee replacements were evaluated in terms of blood loss, pain relief, and range of motion. Patients in the cold compression group demonstrated an average of 548 ml in suction drainage, whereas those in the control group averaged 807 ml. This resulted in an average 3.1 mg hemoglobin drop in the cold compression group and 4.7 mg in the control group. When body habitus and weight were taken into account in the cold compression group, an average total blood loss of 1298 cc was calculated, with 744 ml arising from soft tissue extravasation. The corresponding total blood loss calculated average was 1908 ml in the control group, with 1101 ml attributed to soft tissue extravasation. Total injectable morphine per kilogram per initial 48 hours averaged 0.53 mg in the cold compression patients and 0.69 mg in the control patients. In the cold compression knees, range of motion averaged 86° before operation, 53° on postoperative day (POD) 7, and 77° on POD 14. In the control knees, range of motion averaged 88° before operation, 44° on POD 7, and 65° on POD 14. The use of cold compression in the postoperative period of TKA results in a dramatic decrease in blood loss. In addition, mild improvements are seen in early return of motion and injectable narcotic pain needs in the postoperative period.