Mobile-bearing (MB) total knee arthroplasty (TKA) was developed as an alternative to the established fixed-bearing (FB) design because of theoretical advantages. Short-term studies comparing these designs have not shown any differences in clinical and radiographic results. We compared the results at 7 years of a randomised study of patients undergoing TKA using either a FB or a MB variant of the same prosthesis. Fifty-two patients (52 knees) with an average age of 70 years received a FB posterior-stabilized prosthesis, and 50 patients (52 knees) with an average age of 72 years, a MB prosthesis. All implants were cemented and the patella was routinely resurfaced. Preoperatively, there were no differences between the two groups, and surgical procedure and postoperative protocol were the same for both. At an average follow-up of 7.1 years, no significant differences of FB over MB design could be demonstrated with respect to the American Knee Society score (AKSS), pain score, a questionnaire of general health (SF-12 score), range of motion (ROM), or complication rates. Radiographs showed no significant difference in prosthetic alignment or evidence of loosening. Two knees with a MB design required reoperation, one for persistent joint stiffness and another to treat septic loosening. One patient with a MB prosthesis with signs of tibial component loosening was asymptomatic. We conclude that at mid-term follow-up there is no evidence to prove the superiority of MB over FB TKA with regard to the clinical and radiographic results.