We reviewed the clinical outcomes of fifty primary total knee arthroplasties that had been performed with cement in forty patients who were considered morbidly obese (a Quetelet index of more than forty). These results were compared with those of 1768 similar procedures, performed during the same time-period by the same surgeon, in 1539 patients who were not morbidly obese (controls). At a mean of approximately five years postoperatively, there was a significant difference between the morbidly obese patients and the control group with regard to the knee and functional scores (84 and 53 points compared with 92 and 67 points; p < 0.00005 for both scores). No significant difference was detected, with the numbers available, with regard to the range of motion or the radiographic score (p = 0.77). The rate of perioperative complications was significantly higher in the morbidly obese patients (p < 0.00005). Of the fifty knees in these patients, eleven (22 percent) had a wound complication, five (10 percent) had an infection, and four (8 percent) had an avulsion of the medial collateral ligament. The five infections developed within twenty weeks after the operation, and three were associated with a wound complication. In comparison, thirty-five (2 percent) of the 1768 knees in the control group had a wound complication, eleven (0.6 percent) had an infection, and none had an avulsion of the medial collateral ligament. We concluded that total knee arthroplasty in morbidly obese patients can be successful but is associated with an increased rate of perioperative complications, including problems with wound-healing, infection, and avulsion of the medial collateral ligament. Alterations in the operative technique for soft-tissue closure and protection of the medial collateral ligament have decreased the rates of complications related to wound-healing and the medial collateral ligament.