Biomechanical Determinants of Performance and Injury Risk During Cutting: A Performance-Injury Conflict?

This study used a cross-sectional design to determine the relationship between COD biomechanics and COD performance (completion time, GCT, exit velocity) and injury risk (peak KAMs and peak KIRMs), following an associative strategy [103]. Participants performed six 90° cuts (Fig. 1) from their right limb and 3D motion and GRF analysis was used to explore the joint kinetic, kinematic, and GRF determinants of performance and injury risk during cutting, similar to the methodological procedures of previous research [70, 104, 105].

A minimum sample size of 48 participants was determined from an a priori power analysis using G*Power (Version 3.1, University of Dusseldorf, Germany) [106]. This was based upon a previously reported correlation value of 0.472 (lateral foot plant distance to peak KAM) [70], a power of 0.95, and α level of 0.05. Lateral foot plant distance theoretically should be a key variable linked to the performance–injury conflict, because of the requirement to generate medio-lateral impulse [53, 62, 70, 93] for faster performance, and the increased moment arm distance between the GRF vector and knee joint centre increasing peak KAMs, thus ACL injury risk [31‐33, 62, 70]. As such, 61 male athletes (mean ± SD; age: 20.7 ± 3.8 years, height: 1.77 ± 0.06 m, mass: 74.7 ± 10.0 kg) from multiple sports (soccer n = 43, rugby n = 10, cricket n = 7, field hockey n = 1) participated in this study. For inclusion in the study, all athletes had played their respective sport for a minimum of 5 years and regularly performed one game and two structured skill-based sessions per week. All athletes were free from injury during the study and none of the athletes had suffered a prior traumatic knee injury such as an ACL injury. At the time of testing, players were currently in-season (competition phase). The investigation was approved by the Institutional Ethics Review Board (HSR1617-02), and all participants were informed of the benefits and risks of the investigation prior to signing institutionally approved consent or parental assent documents to participate in the study.

The warm-up [104], marker placement [30, 104, 105, 107], 3D motion analysis [30, 104, 105, 107], and procedures were based on previously published methodologies [81, 104, 105, 107]; thus, a brief overview is provided here. Each participant performed six acceptable trials of a 90° pre-planned side-step cut (Fig. 1) as fast as possible and were provided with standardised footwear to control for shoe–surface interface (Balance W490, New Balance, Boston, MA, USA). Completion time was assessed using two sets of Brower timing lights placed at hip height at the start and finish (Draper, UT, USA). Marker and force data were collected over the PFC and FFC using ten Qualisys Oqus 7 (Gothenburg, Sweden) infrared cameras (240 Hz) operating through Qualisys Track Manager software (Qualisys, version 2.16 (Build 3520), Gothenburg, Sweden) and GRFs were collected from two 600 mm × 900 mm AMTI (Advanced Mechanical Technology, Inc, Watertown, MA, USA) force platforms (Model number: 600900) embedded into the running track sampling at 1200 Hz. Using the pipeline function in visual 3D, joint coordinate (marker) and force data were smoothed using a Butterworth low-pass digital filter with cut-off frequencies of 15 and 25 Hz, based on a priori residual analysis [108], visual inspection of motion data, recommendations by Roewer et al. [109], and to preserve the GRF signal to explore kinetic determinants. Additionally, we have previously reported good agreements (ρ = 0.768–0.859) for peak KAM participant ranking between 15 and 25 Hz and matched cut-off frequencies (12–12 Hz, 15–15 Hz, 18–18 Hz) [110]; thus, participants were likely to display similar rankings between conditions. Lower limb joint moments were calculated using an inverse dynamics approach [111] through Visual 3D software (C-motion, version 6.01.12, Germantown, USA) and were defined as external moments and normalised to body mass. Joint kinematics and GRF were also calculated using visual 3D, while GRF braking and propulsive characteristics were normalised relative to body weight, with vertical, anterior–posterior, and medial–lateral corresponding to Fz, Fx, and Fy, respectively.

A full description of variables along with definitions, abbreviations, and calculations are provided in Supplementary Material 1. Briefly, lower-limb joint moments were calculated over the FFC and lower-limb joint and trunk angles were also calculated and assessed at initial contact, peak, and range of motion of the FFC. Peak and mean GRF braking and propulsive characteristics were also calculated. Weight acceptance (braking) was defined as the point of initial contact to maximum knee flexion and push-off (propulsion) was defined the point of maximum knee flexion to toe-off. PFC braking forces were also assessed for an indication of braking strategies, and horizontal COM velocity profiles at PFC touch-down to determine approach velocity, FFC touch-down, and FFC toe-off to determine exit velocity (Supplementary Material 1) were also examined. COD performance dependent variables were completion time, FFC GCT, and exit velocity, while injury risk dependent variables were peak KAMs and peak KIRMs and were used as surrogates of ACL injury risk [31, 32].

All statistical analyses were performed in SPSS v 25 (SPSS Inc., Chicago, IL, USA) and Microsoft Excel (version 2016, Microsoft Corp., Redmond, WA, USA). Normality was inspected for all variables using a Shapiro–Wilks test. To explore the biomechanical determinants of performance and injury risk-dependent variables, Pearson’s (for parametric data) and Spearman’s (for non-parametric data) correlations were used, similar to previous research [60, 70]. Correlations were evaluated as follows: trivial (0.00–0.09), small (0.10–0.29), moderate (0.30–0.49), large (0.50–0.69), very large (0.70–0.89), nearly perfect (0.90–0.99), and perfect (1.00) [112]. A correlation cut-off value of  ≥ 0.40 was considered relevant according to Welch et al. [51] who also investigated the biomechanical determinants of cutting performance. Thus, correlations greater than this value are only reported. Stepwise multiple regression analysis was also performed to explore the relationship between the abovementioned variables and key primary performance and injury risk variables. Only significantly correlated variables that were parametric were considered for the Stepwise multiple regression analysis, and no more than 6 variables were inputted into the model to ensure a minimum 10:1 participant to independent variable ratio was present [113]. Statistical significance was defined as p ≤ 0.05 for all tests. A minimum of four trials was used for each participant [61], and an average of individual trial peaks for each variable was used [104, 114].

Shorter completion times were significantly (p ≤ 0.001) and very largely associated with greater FFC touch-down (ρ = − 0.752) (Fig. 2) and exit velocities (r = − 0.733); largely associated with faster approach velocities (ρ = − 0.660), greater peak (r = − 0.641) (Fig. 2) and mean resultant propulsive forces (r = − 0.530) and medial–lateral propulsive forces (r = − 0.588 to − 0.627), shorter approach times (ρ = 0.620), greater mean horizontal propulsive forces (r = 0.608), greater peak KIRMs (ρ = − 0.539), shorter PFC and FFC GCTs (ρ = 0.551–0.581), and greater PFC (r = 0.551) and FFC mean horizontal braking forces (r = 0.535); and moderately associated with greater mean FFC resultant braking forces (r = − 0.484), greater peak vertical propulsive forces (r = − 0.449) and horizontal propulsive forces (r = − 0.460), greater peak KAMs (r = − 0.412), greater initial foot progression angles (r = − 0.411), and lower hip flexion range of motion (r = 0.406). Stepwise multiple regression analysis revealed that greater exit velocities, greater peak resultant propulsive forces, greater PFC mean horizontal braking forces, and greater initial foot progression angles together could explain 64.2% (r = 0.801, adjusted 61.6%, p = 0.048) of the variation in completion time. The regression equation is presented in Table 2.

Shorter GCTs were significantly (p ≤ 0.001) and largely associated with greater lateral foot plant distances (ρ = 0.626), lower peak knee and hip flexion angles and range of motion (ρ = 0.603–0.670) (Fig. 2), and lower peak lateral trunk flexion angles and range of motion (ρ = 0.595–0.623).

Faster exit velocities were significantly (p ≤ 0.001) and very largely associated with shorter completion times (r = − 0.733) and greater FFC touch-down velocities (ρ = 0.725); largely associated with greater mean and peak medial–lateral (r = 0.638–0.651) and resultant propulsive forces (r = 0.549–0.568), shorter FFC GCTs (ρ = − 0.569), greater peak vertical (r = 0.540) and horizontal propulsive forces (r = 0.500), and greater approach velocities (ρ = 0.533); and moderately associated with greater mean vertical propulsive forces (r = 0.499), shorter PFC GCTs (ρ = − 0.484), greater peak KFMs (r = 0.482), lower hip flexion range of motion (ρ = 0.470), greater FFC mean vertical braking forces (r = 0.456), and greater PFC mean horizontal braking forces (r = − 0.430). Stepwise multiple regression analysis revealed that greater FFC peak medial–lateral propulsive forces and peak greater KFMs together could explain 47.8% (r = 0.690, adjusted 45.7%, p = 0.019) of the variation in exit velocity. The regression equation is presented in Table 2.

Greater peak KAMs were significantly (p ≤ 0.001) and largely associated with greater peak KIRMs (ρ = − 0.557) (Fig. 2); and moderately associated with greater FFC touch-down velocities (ρ = − 0.491) (Fig. 2), greater peak KAAs (r = − 0.468) (Fig. 2), greater FFC mean vertical, horizontal, and resultant braking forces (r = 0.434–0.497), and shorter completion times (r = − 0.412). Stepwise multiple regression analysis revealed that greater FFC mean vertical braking forces and greater peak KAA together could explain 42.6% (r = 0.652, adjusted 40.6%, p < 0.001) of the variation peak KAM. The regression equation is presented in Table 2.

Greater peak KIRMs were significantly (p ≤ 0.001) and largely associated with greater peak KAMs (ρ = − 0.557), greater FFC touch-down velocities (ρ = − 0.551), shorter completion times (ρ = 0.539), greater approach velocities (ρ = − 0.534), and greater FFC peak resultant braking forces (ρ = − 0.505); and moderately associated with greater FFC mean (ρ = − 0.468) and peak vertical braking forces (ρ = − 0.475), and greater FFC peak resultant braking forces (ρ = − 0.458).