Which jump test is most sensitive for classifying ACL injury history in fatigued or non-fatigued athletes?

Abstract

Introduction & Purpose In fast-paced clinical decision-making, jump-landing movement assessments are rarely used to guide return-to-sport (RTS) decisions after an anterior cruciate ligament (ACL) injury. While these tests are crucial for identifying deficits in movement execution (e.g., dynamic valgus, extended knee, knee internal rotation) that could predispose to ACL re-rupture (Kaplan & Witvrouw, 2019), identifying these deficits can be challenging. To integrate RTS movement assessments into clinical practice, the first step is to identify the most sensitive jump-landing test that can best expose remaining deficits in neuromuscular control and resulting landing kinematics associated with ACL injury. Further, RTS tests are mostly performed in a non-fatigued state although fatigued individuals often fail to reach the RTS criteria (Dingenen & Gokeler, 2017). This exploratory study aimed to rank common jump-landing tests in their ability to identify (sensitivity and classification rate) deficits in whole-body landing kinematics following ACL injury, under both fatigued and non-fatigued conditions. Methods A total of 43 volunteers were recruited into ACL group (n = 21, 11 females, 3.9 years post ACL injury) and control group (n = 22, 12 females). 3D motion data (Vicon, 250 Hz) were recorded during a single-leg forward hop (SH), single-leg countermovement jump (CMJ), single-leg triple crossover hop (COH), and single-leg 90° medial rotation hop (MRH) before and after a fatigue-inducing intervention (single-leg squats and step-ups). Thirteen joint angles (50 ms after initial contact) from the lower body, trunk, and pelvis representing the landing kinematics were calculated in OpenSim. A correlation analysis was performed to check for multicollinearity issues. Eight distinct logistic regression models (four jumps, fatigued/non-fatigued, α = .05) were computed to predict ACL injury history based on a combination of joint angles. For each model, a backward stepwise method was employed to determine which combination of joint angles resulted in the highest classification rates. Results ACL injury history was successfully predicted in seven (χ² > 4.4, p < .036) of the eight logistic regression models (Figure 1). The SH in the fatigued condition (χ² = 22.5, p = .008) demonstrated the highest sensitivity of 85.7% and a classification rate of 86%. Significant predictors for this model were knee rotation (Odds ratios [OR] = 1.537), hip abduction (OR = 1.385), hip rotation (OR = 1.525), and ankle flexion (OR = 0.771). The CMJ (non-fatigued) showed the second-highest sensitivity of 81% and a classification rate of 81.4% (p = .027). Significant predictors were knee rotation (OR = 1.440), hip abduction (OR = 1.428), hip rotation (OR = 1.297), and pelvis rotation (OR = 0.664). The MRH in the non-fatigued condition exhibited the lowest sensitivity, and no classification was possible in the non-fatigued COH model (χ² = 3.6, p = .056). Discussion & Conclusion The fatigued SH was the most sensitive jump-landing test for classifying groups and detecting kinematic differences due to previous ACL injury. Thus, we recommended the SH over other tests for RTS decision-making, as it is most likely to reveal kinematic abnormalities. Higher classification rates in the fatigued condition support RTS testing in fatigued states to better identify movement deficits. To integrate jump-landing assessments into RTS decisions in a feasible way, our future work will involve 3D motion capture using pose estimation algorithms on video data.