Effects of repeated-sprint training in hypoxia on repeated-sprint and sprint performance in athletes: a systematic review and three-level meta-analysis

Repeated-sprint training in hypoxia (RSH) has been proposed as a strategy to enhance repeated-sprint and sprint performance in athletes. However, evidence regarding its effectiveness remains inconsistent, and no consensus has been reached regarding its true effects. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines, this systematic review and meta-analysis was registered in PROSPERO (CRD420261415952). English-language controlled intervention studies examining repeated-sprint training in hypoxia (RSH) in athletes or trained participants were retrieved from PubMed, Web of Science, Embase, the Cochrane Library, and Ovid from database inception to June 5, 2026. Hedges’ g was calculated from post-intervention means, standard deviations, and sample sizes, with positive values favoring RSH. A three-level random-effects model accounted for multiple effect sizes nested within studies. Ninety-five percent prediction intervals (PIs), outcome-domain subgroup analyses with Benjamini–Hochberg adjustment, leave-one-study-out analysis, a sensitivity analysis excluding voluntary hypoventilation at low lung volume (VHL), exploratory publication-bias assessment, a crossover-variance scenario analysis, and Grading of Recommendations, Assessment, Development and Evaluation (GRADE) assessment were conducted. Ten studies involving 216 unique athletes (224 condition-level observations because one eight-participant trial used a crossover design) and 25 effect sizes were included. The main three-level model showed a significant effect of RSH (Hedges’ g = 0.31, 95% CI [0.12, 0.50], 95% PI [0.01, 0.60], P = 0.001; I 2 Level 2 = 0.00%, Level 3 = 6.95%). Sprint time/RSA time (g = 0.34, 95% CI [0.06, 0.61], BH-adjusted P = 0.025) and sprint decrement (g = 0.53, 95% CI [0.12, 0.95], BH-adjusted P = 0.025) favored RSH, whereas power/work output remained uncertain (g = 0.24, 95% CI [ − 0.08, 0.57], P = 0.137). Leave-one-study-out results were robust. Crossover-variance scenarios preserved the pooled effect. Exploratory Egger regression was not statistically significant ( P = 0.072). RSH may improve repeated-sprint performance, including sprint-time/RSA-time and sprint-decrement outcomes, but the evidence is of very low certainty and does not establish a consistent benefit for power/work output. RSH should therefore be regarded only as a promising supplementary strategy pending larger, well-reported trials and more precise estimates of effects in new settings.