Mechanisms, injury patterns, and kinematic features of hamstring strain injuries in football (soccer): a systematic review and meta-analysis of video-analysis studies.

Background and objective: Match video analysis enables retrospective examination of the movement tasks, match situations, and body postures immediately before and after hamstring strain injury (HSI) in football, thereby providing information on the injury process that is difficult to obtain directly from conventional clinical or imaging data. This study aimed to systematically evaluate video-based studies of football matches and to compare the relative reporting frequencies of several prespecified key epidemiological contrasts. In addition, single-arm proportion pooling was used to provide descriptive background information, and exploratory analyses were conducted on kinematic postures and sex differences. Methods: Seven databases were systematically searched from inception to February 12, 2026. Studies of 11-a-side football were included if they used video analysis to assess the inciting moment of real-match HSI and reported injury mechanisms, situational patterns, or kinematic characteristics. The primary analyses focused on five prespecified pairwise comparisons. Study-specific 2 × 2 contingency tables were constructed, and odds ratios (ORs) with 95% confidence intervals (CIs) were pooled using random-effects models. The primary OR analyses were restricted to male samples. Single-arm proportions were pooled using a generalized linear mixed model. Analyses related to kinematics and sex were all treated as exploratory. Results: A total of 7 studies involving 342 HSI cases were included. The main comparisons showed that non-contact was reported more frequently than indirect contact (OR = 2.51, 95% CI 1.31, 4.82), and sprint-type was reported more frequently than mixed-type (OR = 2.01, 95% CI 1.04, 3.87). In contrast, sprint vs. stretch, offensive vs. defensive, and first half vs. second half showed no clear differences. The descriptive pooled results suggested relatively high pooled proportions for sprint-type, running-related action, and non-contact. Exploratory analyses showed that samples associated with different mechanisms were generally similar in knee-joint posture, whereas differences may exist in hip-joint and trunk alignment; cross-study exploratory comparisons by sex did not identify any results that remained consistently significant after multiple-comparison correction. Conclusion: Existing video-analysis evidence from football matches shows that HSI occurs more commonly in situations without external contact, and that sprint-type HSI is more frequent than mixed-type HSI; however, the relative differences between offensive and defensive phases and between the first and second halves remain unclear. These findings may provide a basis for hypotheses in football HSI prevention research, suggesting that future studies may consider examining non-contact injury scenarios in the context of high-speed running, eccentric load tolerance, and neuromuscular control during the single-leg support phase; however, the existing video-analysis evidence itself cannot directly demonstrate the preventive effects of specific training strategies. The above conclusions are limited by factors such as the small number of included studies, inconsistent definitions of injury mechanisms, and insufficient standardization of video analysis, and future research should further accumulate cross-sex samples and harmonize reporting standards. Systematic review registration: This systematic review and meta-analysis has been registered in PROSPERO (www.crd.york.ac.uk/prospero), identifier CRD420261335634.