Background: Injured-list (IL) time reflects the clinical and competitive impact of pitching injuries, but determinants of cumulative injury burden have rarely been evaluated league-wide using integrated workload and pitch-tracking data.
Purpose: To leverage large-scale public data to identify pitcher characteristics associated with cumulative time-loss IL burden due to shoulder and elbow injuries, accounting for exposure and role.
Study design: Case-control study.
Methods: This study analyzed aggregated data from the study period between 2015 and 2025 for 1445 Major League Baseball pitchers (≥1000 pitches). Injury burden was defined as cumulative IL days per pitcher over the study period and stratified into shoulder (S), elbow (E), and a total combined (S+E) category using public injury designations. Candidate predictors included pitcher demographic characteristics, advanced performance metrics, pitch-tracking characteristics, and within-game pitch-to-pitch delta measures reflecting adjacent-pitch variation in velocity and release characteristics. We used negative binomial models to estimate incidence rate ratios (IRR) for IL days, accounting for differences in pitching exposure by including the natural logarithm of total pitches thrown as an exposure offset and adjusting for pitching role using the percentage of appearances as a starting pitcher. Candidate predictors were screened in univariate models and then entered into multivariable models.
Results: The cohort accrued 147,765 total IL days (47,022 shoulder; 100,743 elbow). In multivariable models, higher strikeout rate (IRR, 1.24), 4-seam velocity (IRR, 1.18), and 4-seam spin (IRR, 1.13) were independently associated with greater combined S+E burden (all P < .001). Joint-specific models revealed divergent signals: greater release extension was associated with lower shoulder injury burden (IRR, 0.87; P < .001), whereas higher pitch-to-pitch velocity delta (IRR, 1.08; P = .009) was specifically associated with greater elbow burden. Univariate screening identified several additional factors significantly associated with cumulative and joint-specific outcomes, including arm angle, pitch usage (eg, changeups and sliders), and "Plus" metrics.
Conclusion: Cumulative injury burden is driven by a high-performance phenotype, but risk signals diverge meaningfully by joint. Shorter release extension and greater pitch-to-pitch velocity delta emerged as distinct, potentially modifiable correlates of shoulder and elbow injury burden, respectively. These findings identify mechanical and consistency-based signals that warrant prospective investigation to refine individualized workload management and injury prevention.
Keywords: injury surveillance; pitch-tracking; pitching; ulnar collateral ligament; workload.
© The Author(s) 2026.