Effects of integrating blood flow restriction (BFR) into sport-specific on-field training on athletic performance and physiological adaptations in team-sport athletes: a systematic review and meta-analysis

To evaluate the additional effects of integrating blood flow restriction (BFR) into matched sport-specific on-field training (SOFT) on specific athletic performance and physiological outcomes in team-sport athletes. Eight databases (MEDLINE, PubMed, Embase, CENTRAL, Web of Science, Scopus, SPORTDiscus, and China National Knowledge Infrastructure) were searched up to July 30, 2026. Between-group differences in change scores were pooled as mean differences (MDs) or Hedges’ gΔ using restricted maximum likelihood random-effects models. Robustness was examined using alternative approaches to change-score standard deviations, pre–post correlation coefficients, and small-sample inference. RoB 2 and GRADE were used to assess outcome-level risk of bias and certainty of evidence. Six journal reports represented 5 independent trials, with a total reported sample size of 102 participants across the selected comparison arms. For 30-second all-out cycling, the relative peak power estimate was MD = − 0.52 W/kg (95% CI − 1.40 to 0.36), and the relative mean power estimate was MD = − 0.50 W/kg (− 0.85 to − 0.15). The estimates were gΔ=−1.55 (− 2.88 to − 0.21) for lower-limb 1RM, MD = − 1.00 mL/kg/min (− 2.63 to 0.62) for oxygen uptake, gΔ=−0.44 (− 1.02 to 0.14) for RSA mean time, and MD = − 1.75 cm (− 4.05 to 0.56) for CMJ height; negative values favoured BFR. Heterogeneity was high for strength (I²=86.4%), and its interval crossed zero under the alternative variance scheme; the interval for mean power crossed zero after exclusion of the Amani trial. Under modified Knapp–Hartung inference, the intervals for all six primary outcomes crossed zero. Certainty was low for five primary outcomes and very low for strength. Adding BFR to matched SOFT may improve relative mean power during 30-second cycling, but the certainty of evidence is low. Its effect on lower-limb maximal dynamic strength is very uncertain, and no clear additional benefits have been demonstrated for the other primary outcomes. The results are affected by variance assumptions and the small number of studies and remain insufficient to support routine use or establish safety. PROSPERO CRD420261296954.