Acute hemarthrosis is a frequent complication of hemophilia A. To test the hypothesis that recurrent hemarthrosis has an adverse effect on neuromuscular function, ten patients with hemophilia A and history of unilateral hemarthrosis of the knee were studied. The uninvolved (U) side was used as the control. The time since diagnosis, factor level, and the number of bleeding episodes in the involved (I) knee averaged 11.6 +/- 6.6 years, 7.7 +/- 5.1%, and 7.0 +/- 7.4, respectively. Neuromuscular function was evaluated on a Cybex 340 isokinetic dynamometer. Knee extensor strength measured at 60 degrees/sec was significantly (p less than .05) lower in the I side (mean = 58Nm) than in the U side (mean = 85Nm). Similarly, lower (p less than .05) total work (I = 479J, U = 656J) and average power output (I = 59W, U = 83W) values were obtained. Adjusting for thigh circumference did not eliminate any of these differences. Testing the extensors at faster angular velocities (180 degrees/sec and 240 degrees/sec) and the flexor muscle groups at the three speeds revealed no differences between I and U. X-rays showed minimal changes. These data show that patients with hemophilia A and history of unilateral hemarthrosis of the knee have neuromuscular dysfunction in the I extremity that precedes the appearance of radiologic evidence of joint pathology. It is suggested that strength training should be started early in the rehabilitation of hemophiliacs.