Reliability of knee extensor and flexor torques during continuous concentric-eccentric cycles.

Peak and average torques of the knee extensors and flexors were determined on three occasions within ten days. On each occasion, 20 healthy women and 15 healthy men performed three continuous concentric-eccentric cycles with no pause between muscle actions or between cycles at 45 degrees and 90 degrees/sec angular velocities. Peak and average torques did not vary significantly or systematically over the three test occasions. Intraclass correlation coefficients (ICCs), based on one occasion, ranged from 0.79 to 0.91 for peak torques and from 0.75 to 0.88 for average torques. At 45 degrees/sec angular velocity, the ICCs for peak torque tended to be greater than those for average torque, whereas at 90 degrees/sec angular velocity the ICCs were similar whether calculated using peak or average torque. The ICCs during knee extension were greater than those during knee flexion. Eccentric muscle actions produced significantly higher torques than did concentric actions (p less than 0.01). A continuous concentric-eccentric cycle protocol which emphasizes teaching the subject to perform the test maneuvers and using at least three submaximal practice contractions and one maximal practice contraction can produce reliable isokinetic torques.