In a prospective randomized double-blind study we compared the postoperative recovery and early results of two groups of 30 patients having total knee arthroplasty with minimally invasive techniques using either a mini-subvastus or a modified "quadriceps-sparing" approach. All knees were implanted with the same posterior-stabilized prosthesis (LPS-Flex, Zimmer, Warsaw, IN) by the same surgeon with the same dedicated set of downsized instruments. Epidural anesthesia with the same postoperative analgesia and rehabilitation protocol was used in all patients. Evaluation was performed preoperatively, postoperatively in the first week, and at 1 and 3 months. In five cases in the "quadriceps- sparing" group, the incision was extended a few cm to facilitate exposure. Tourniquet time, estimated blood loss, and postoperative pain were similar in the two groups. Active straight leg raising was achieved half a day earlier, on average, in the mini-subvastus group (1.9 vs 1.4 days). Average maximum active flexion was similar in the two groups at each interval, and reached 117 degrees and 119 degrees at 3 months for the mini-subvastus and "quadriceps-sparing" group, respectively. We believe there was no difference between the mini-subvastus and "quadriceps-sparing" approach in relation to short term recovery or early results.