Author(s): LEKATSAS, N., KEKELEKIS, A., KELLIS, E., Institution: AUTH, Country: GREECE, Abstract-ID: 862

INTRODUCTION:
Jump-based screening is widely used; however, simple performance outcomes such as jump height show limited ability to identify athletes at increased risk of lower-limb injury. Force–time characteristics may provide more injury-relevant information, although prospective evidence in elite male professional football is limited. This study examined the association between preseason jump-derived neuromuscular metrics and non-contact lower-limb injuries in elite male football players.
METHODS:
Professional players competing in the top national division in Greece completed a standardized preseason neuromuscular assessment including bilateral countermovement jump (CMJ), single-leg CMJ (SL-CMJ), and drop jump (DJ) testing using force-plate technology (Vald). Selected variables reflected performance capacity, force–time asymmetries, reactive strength, and landing-related characteristics. Non-contact lower-limb injuries resulting in ≥24 h time-loss were prospectively recorded throughout the season using established injury surveillance definitions. Associations between preseason metrics and injury occurrence were examined using logistic regression and multivariable modeling approaches.
RESULTS:
Bivariate point-biserial correlations showed small to moderate associations between preseason metrics and injury. The strongest positive correlations were observed for CMJ concentric force asymmetry (r ≈ 0.43) and CMJ RSI mod (r ≈ 0.39), followed by CMJ peak power (r ≈ 0.28). Shorter DJ contact time was moderately associated with higher injury incidence (r ≈ −0.35). Other variables demonstrated weak or inconsistent associations. Boxplots revealed substantial overlap between injured and non-injured players, with no clear discriminative thresholds. A multivariable logistic regression model including all standardized predictors demonstrated modest predictive performance under leave-one-out cross-validation (AUC ≈ 0.65; accuracy ≈ 0.68). In the final model, injury odds increased with higher CMJ RSI mod (OR = 3.15 per SD), CMJ concentric force asymmetry (OR = 2.27 per SD), and CMJ peak power (OR = 1.71 per SD). Lower injury odds were observed for DJ jump height (OR = 0.46), DJ RSI (OR = 0.68), and DJ contact time (OR = 0.73). Other asymmetry measures showed smaller effects (ORs ≈ 1.33–1.46 or <1.0), likely influenced by sampling variability. Predicted injury probabilities were higher on average in injured players, but substantial misclassification occurred.
CONCLUSION:
Consistent with previous literature, preseason jump performance metrics alone were insufficient for injury risk screening in elite male football players. Force–time asymmetries and reactive strength characteristics provided additional, albeit modest, injury-relevant information when interpreted within a multifactorial framework. These findings support the use of standardized jump-based assessments as part of an integrated monitoring strategy rather than as standalone injury prediction tools.