Quantifying Knee Strength Deficits and Muscular Contribution to Torque Generation in ACL-Injured Adolescents Using A Simple Musculoskeletal Model

BACKGROUND: Surface electromyography (sEMG) is commonly used to elucidate the effect of anterior cruciate ligament (ACL) injury on neuromuscular function. To facilitate comparisons, sEMG is normalized to a known value, such as peak activation during a maximum voluntary isometric contraction (MVIC). However, voluntary muscle inhibition after an injury compromises one’s ability to achieve a true MVIC value. Musculoskeletal modeling can be used to quantify this strength deficit by comparing a theoretical MVIC torque to the actual MVIC torque. PURPOSE: This study sought to evaluate voluntary knee extensor and flexor strength deficits and individual muscle contribution to peak torque generation in ACL injured (ACLi) and uninjured (CON) adolescent populations using a simplified subject-specific modeling framework. METHODS: Thirty-nine ACLi (25 females) and 39 CON (25 females) adolescents (12-17years) completed knee extension and flexion MVICs on an isokinetic dynamometer. Peak experimental torque (TE) was identified. A subject-specific modeling framework used normalized sEMG of the quadriceps, hamstrings, and gastrocnemius muscles to determine theoretically ideal torque (TI) for each exercise, assuming agonist muscles were fully activated. Strength deficit ratios (TE/TI) and individual muscle contribution to TE were computed. Group mean differences were compared using independent t-tests. RESULTS: ACLi demonstrated significantly lower extension (2.53±0.94 vs 3.07±0.57Nm/kg, p<0.001) and flexion (1.14±0.50 vs 1.37±0.31Nm/kg, p<0.001) peak TE compared to CON (FIG 1.A). Significant between group differences in sEMG of antagonist muscles were observed. Both groups demonstrated similar TI values (FIG 1.B-upper). Significantly lower trends in strength ratios (TE/TI) were maintained between groups (FIG 1.B-lower). However, percent between group differences were minimized from 17.8% to 6.3% for knee extension, and 16.7% to 10.0% for knee flexion, when TE is expressed relative to TI. Significantly lower medical gastrocnemius contribution to flexion TE was also observed in the ACLi (24.7±0.08%) compared to CON (30.9±10.1%, p=0.037). CONCLUSION: Significant between group differences in peak extensor or flexor torques were observed. Our model confirmed a strength deficit in ACLi compared to CON, however, between group differences were less prominent when described relative to TI. This may be attributed to between group differences in antagonist contributions (i.e. co-contraction) during MVIC exercise. Simplified modeling frameworks that can incorporate muscle activations as well as torque generation may be more appropriate for evaluating functional outcomes of ACLi populations in a clinical setting.