Author(s): MAITÉ, C., MOHR, M., FEDEROLF, P. , Institution: UNIVERSITÄT INNSBRUCK , Country: AUSTRIA, Abstract-ID: 2102

INTRODUCTION:
Return-to-sport (RTS) assessments often involve evaluations of movement execution, e.g. in jump landings, to identify deficits in movement control and connected risks for re-injury [1]. For RTS after anterior cruciate ligament (ACL) injury, the current literature offers a plethora of kinematic variables in various jump tests that clinicians may use to guide the RTS decision [1]. To better integrate RTS movement assessments in fast-paced clinical decision-making, it may be useful to investigate the most sensitive jump test and variable combinations for detecting movement deficits. The purpose of the current study was to compare the sensitivity of three commonly used jump tests for detecting movement characteristics influenced by either an ACL injury history, a provoked fatigue status, or a combination of both.
METHODS:
A total of 43 volunteers were recruited into ACL group (n=21, 11 females) and control group (n=22, 12 females). 3D motion data (Vicon, 250 Hz) were recorded during a single-leg hop (SLH), unilateral counter movement jump (uCMJ) and a unilateral cross-over hop (COH) before and after a fatigue-inducing intervention (single-leg squats and step ups). Thirteen joint angles from lower body, trunk and pelvis (50ms after initial contact) representing the landing posture were calculated through inverse kinematics in OpenSim. One combined principal component (PC) analysis was computed for all three jumps to characterize kinematic synergies. Six distinct logistic regression models (three jumps, fatigued/non-fatigued, alpha = .05) were calculated to predict ACL injury history. Three additional models predicted fatigue status. The predictors consisted of twelve PC scores.
RESULTS:
In all three jump landings, the logistic regression models were able to detect an ACL injury history (Chi²= 4.974, p<.026) but not fatigue status (Chi²=2.165, p>.141). When predicting ACL injury history, the highest sensitivity (76%) and classification rates (77%) were achieved for SLH (p<.001) and uCMJ (p<.001) when participants were fatigued. The worst sensitivity (67%) and classification (63%) was achieved for non-fatigued COH (p=.026). The SLH and uCMJ models consistently relied on the same two PCs, which described the correlation between frontal and transverse knee, hip and trunk angles.
CONCLUSION:
In our data-driven analysis, the SLH and uCMJ appear more sensitive for detecting movement characteristics related to a previous ACL injury compared to the COH. Furthermore, our results support the recommendation to include fatigued conditions during RTS tests post-ACL injury [2]. The finding that our approach could not predict fatigue status could mean that (1) fatigue effects were non-systematic or (2) the between-subject variance in PC scores blurred smaller within-subject fatigue effects. Either way, the observation of an unsuccessful prediction makes it less likely that the classification by injury history was due to chance.

[1] Kaplan, Sports Health, 2019
[2] Dingenen, Sports Med, 2017