The purpose of this study was to evaluate the long-term isokinetic muscle performance after ACL surgery and to analyze by a multiple stepwise regression which factors (if any) predict the overall outcome. The study subjects were 119 patients who had had a complete rupture of the ACL and had been treated surgically at the Tarnpere University Hospital between 1981 and 1990. They were divided into the acute group (N= 62) and into the chronic group (N= 57) according to the time delay between the injury and the ACL surgery. The isokinetic strength measure ments of the quadriceps and hamstring muscles were performed on average 4 years after the operation using Cybex II and Cybex 6000 dynamometers (Lumex Inc.. Ronkokoma, NY, USA). The peak torques were determined at speeds of 60 and 180'/s and the peak work at speed of 180'Is. The measurements revealed that after the ACL surgery there was still a considerable thigh muscle strength deficit especially in extension of the injured extremity, the average extension strength deficit ranging from 9 to 20%. The extension strength deficit was significantly more prominent in the chronic (18-20%) than in the acute group (9-15%). The extension strength deficit was also significantly greater at the slower (15-20%) than higher (9-18%) speed of the isokinetic movement. In both the acute and chronic groups, a multiple stepwise regression analysis indicated that patellofemoral pain and flexion deficit of the knee were the factors that most frequently and significantly associated with the strength deficits. At the higher speed of theisokineticmovement, these two factors accounted 20% and 21% for the variation seen in the quadriceps strength deficit of the acute group and the chronic group. No correlation could be found between age, sex, height, weight, body mass index. length of the follow-up time. injury type, athletic activity level, immobilization method, knee stability, and the isokinetic muscular performance.