How effective is stretching for injury prevention? A large-scale systematic review with meta-analysis

Background: Sports injuries can impact athletes’ availability, threaten their career and have substantial financial impact. Stretching is commonly used as a prevention strategy. However, controversial evidence, methodological shortcomings, and substantial time elapsed since previous key data syntheses justify performing an updated systematic review. Our goal was to assess the effects of stretching interventions on injury risk.Methods: We followed the PRISMA 2020 guidelines. Eligible studies included humans performing stretching and at least one comparator group; sports-related injury data was required. Sources: six electronic databases (CINAHL Complete, Embase, PubMed, Scopus, SPORTDiscus, Web of Science), reference lists, citation tracking, and consultation of external experts. Risk of bias was assessed using Cochrane tools. Raw injury data was used to assess Odds Ratios (OR), with OR <1 meaning a preventive effect for stretching. Results were presented with 95% confidence intervals (CIs) and prediction intervals (PIs). An inverse-variance random-effects model was employed using the DerSimonian and Laird estimator. Certainty of evidence was assessed using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) framework.Results: Over 300,000 records were screened, with 19 studies fulfilling the eligibility criteria (>9,000 participants, >80% male, age range 14 to 59 years), of which 15 were used for the meta-analyses. Interventions lasted 4 to 24 weeks and mostly implemented static unassisted stretching with varying volumes. Half the studies used passive controls, and half used diverse active comparators. Risk of bias was high for ~90% of randomized trials and serious for all non-randomized trials. No preventive effect was detected for stretching versus all other comparators (k = 15, n = 9,002; OR = 0.945, p = 0.396, 95% CI 0.828 to 1.078, 95% PI 0.773 to 1.155), passive controls (k = 10, n = 6,273; OR = 0.923, p = 0.209, 95% CI 0.814 to 1.046, 95% PI 0.796 to 1.069), or active controls (k = 6, n = 2,744; OR = 1.190, p = 0.312, 95% CI 0.849 to 1.667, 95% PI 0.738 to 1.919). Likewise, no preventive effect was found for muscle, tendon, and ligament/joint capsule injury risk, or across 19 sensitivity analyses, 3 subgroup analyses, and 3 secondary analyses. GRADE denoted very low certainty of evidence for all outcomes except bone injury risk (low certainty).Conclusion: Despite the very low to low certainty of evidence as per GRADE judgments, the data consistently showed no chronic effect of stretching on injury risk across multiple analyses. No study assessed the acute effects of stretching on injury risk. Given the studied samples, these conclusions apply mostly to male aged 14 to 60 years performing static unassisted stretching, with little to no information about other populations or interventions. Several caveats are discussed, as well as implications for practice, policy-making and future research.Registration: PROSPERO: CRD42024616532. OSF project (https://osf.io/umj8h/) and registration (https://osf.io/ybdvw).