Gravity correction and knee isokinetic torque parameters and hamstring-to-quadriceps (H:Q) ratios: reference equations from a large ACL cohort.

BACKGROUND: Inconsistency in applying gravity-correction during knee isokinetic dynamometry limits direct comparisons across sports medicine literature. This study aimed to develop and internally validate reference equations relating knee peak torque parameters measured with and without gravity correction in an anterior cruciate ligament (ACL) cohort. METHODS: 439 individuals with a confirmed ACL rupture underwent concentric isokinetic knee extension and flexion testing at 60, 180, and 300°/s using a HUMAC Norm dynamometer. Peak torque values and hamstring-to-quadriceps (H:Q) ratios were calculated using both gravity-corrected and uncorrected protocols. Relationships were modelled using linear mixed-effects models with a participant-level random intercept, fitted separately by sex, velocity and outcome. RESULTS: Gravity-corrected extensor torque was 4.2-11.1% greater, flexor torque 13.1-25.5% smaller, and H:Q ratios 17.2-33.1% smaller than uncorrected values, with differences increasing with angular velocity. Marginal R2 ranged from 0.95 to 0.99 for knee extensors, 0.90 to 0.97 for knee flexors, and 0.68 to 0.87 for H:Q ratios. Cross-validated prediction error was small for torque (RMSE 3.3-5.1 Nm; LoA ± 6.5 to ± 10.0 Nm) but larger and velocity-dependent for the H:Q ratio (RMSE 3.6-7.9 percentage points; LoA ± 7.1 to ± 15.5 points). CONCLUSIONS: These equations provide an internally validated reference for approximating peak torque between gravity-corrected and non-gravity-corrected protocols within a comparable cohort and testing set-up; the H:Q equations should be applied with caution, particularly at higher velocities. Neither protocol is inherently superior, but explicit reporting of whether gravity correction was applied is essential to avoid confounding when datasets are compared or pooled.