The rotating platform flexion (RPF) Sigma total knee prosthesis (DePuy; Warsaw, Indiana) was designed for maintaining the contact of the condyles with their corresponding tibial plateau throughout the high-flexion range. However, this requires an additional 3-mm bone cut of the posterior condyles. Compared to the conventional design, this modification is intended to improve the flexion range. This hypothesis was tested by studying the increase in flexion (flexion gain, range of motion [ROM], active flexion) of 59 consecutive patients who had received the hyperflex design implant (RPF), whose preoperative mobility values were retrospectively compared to these same values in another 59 consecutive matched patients who had received an implant with the conventional design of the same implant (rotating platform [RP]) between June 2005 and June 2006. Postoperative mobility was measured visually with a goniometer. Only osteoarthritic knees were eligible to be included. Knees with more than 20 degrees flexion contracture or less than 90 degrees flexion, and patients with a body mass index (BMI) greater than 30 were excluded. Both groups were comparable with regard to age, preoperative mobility values, and BMI. The sex ratio differed significantly, but preoperative mobility did not differ significantly in male and female patients in the RP and in the RPF groups. The difference in sex ratio did not appear to be a bias influencing preoperative mobility. Overall, the flexion gain was correlated to preoperative flexion (r=-0.75, p<0.001). The flexion gain in the RPF group was significantly greater than in the RP group (13+-20 versus 6+-13; p=0.02) as was the ROM gain (10+/-17 degrees versus 4+/-12 degrees; p=0.02). However, the one-year active mean flexions were not significantly different (118+/-14 degrees versus 116+/-6 degrees; p=0.47). In patients whose preoperative flexion was less than 120 degrees (18 and 27 RPF prostheses), the flexion and ROM gains were significantly greater in the RPF group (23+/-16 degrees versus 14+/-16 degrees; p=0.03 and 26+/-18 degrees versus 17+/-9 degrees; p=0.05), and the mean one-year active flexion was also greater in the RPF group (124+/-13 degrees versus 116+/-8 degrees, p=0.02). In patients with more than 120 degrees of preoperative flexion, the flexion and ROM gains and the final mean flexions in both groups were comparable. In particular, there were nine patients in the RP group and ten patients in the RPF group whose flexion decreased. Thus, the Sigma RPF prosthesis provided a significant additional flexion gain in patients with 90-120 degrees preoperative flexion, and less than 20 degrees flexion contracture. Patients with a preoperative flexion greater than 120 degrees were exposed to a decrease in flexion range whichever implant was used, RP or RPF. Level 3, therapeutic study.