Cutting technique of ACL-Reconstructed professional handball players: on-field kinematic assessment using IMUs

Introduction Anterior cruciate ligament (ACL) injury is common among handball players, mostly occurring in non-contact cutting maneuvers [1]. While lab-based studies have identified several biomechanical risk factors, real-game duels in court settings remain unexplored [1]. In this regard, wearable technologies (IMUs) enable on-field assessments during dynamic tasks and ecologically detect potential injury risk factors. The aim of this study was to compare the biomechanics of ACL-reconstructed (ACLR) and healthy handball players using IMUs during on-field sport-specific injuryrelated tasks. Methods Twenty-five professional (first division) handball players (age: 23.7 ± 3.9 years; 14 males), including 7 ACLR players (cleared for RTS>24 months), were tested in their home-team handball court. Players were asked to perform two highdynamic cutting tasks, one mimicking a change of direction with passive opponent and one in response to a real-game opponent (Figure 1). Both tasks were performed six times (three per limb) per player, with an overall of 300 trials collected. Based on dominant hand and movement direction, each trial was classified as either a classic sidestep change of direction (SSC) or a crossover (XOVER). Lower limb and trunk kinematics were recorded using eight IMUs (100 Hz, MTw Awinda, Movella), validated for high-dynamic movement analysis [2]. Joint kinematics was assessed in the three anatomical planes at key events of the cut foot contact window: initial contact (IC), peak of knee flexion (pKF), peak values (maximum and minimum), and range of motion (ROM). Analysis was conducted separately for movement type (SSC and XOVER), comparing healthy and ACLR players in active and passive movements. Two-tailed Student’s t-tests were used to assess between-group differences for each of the 75 extracted features, with Cohen’s d effect size reported alongside p-values. Results During XOVER, 23/75 variables differed significantly between ACLR and healthy players, and 26/75 during SSC. A greater pelvis and trunk flexion (forward bending) was noted for the ACLR group, both during XOVER (p<0.001, d=0.57-0.70) and SSC (p<0.001, d=0.42-0.95). ACLR players exhibited a pronounced pelvis contralateral drop at pKF in XOVER (p=0.013, d=0.50), and at IC in SSC (p<0.001, d=0.90). ACLR players showed a greater tendency to knee valgus in both movements (p<0.017, d=0.47-1.27), and an increased ankle plantarflexion at pKF (p<0.031, d=0.40-0.60). Larger effect size in significant differences and more frequent ACL injury-related patterns [1] were observed in active compared to passive tasks. Discussion ACLR handball players adopted different cutting techniques compared to healthy controls, with an altered trunk-pelvis strategy, increased frontal/transverse rotation, and consistent knee valgus tendency during two typical cut maneuvers. These findings highlight the need for sport-specific assessments using wearable sensors to detect residual impairments, supporting data-driven return-to-play decisions.