Acute caffeine and carbohydrate strategies for fatigue-related sport-specific performance in team-sport athletes: a systematic review and multilevel meta-analysis.

BACKGROUND

Although caffeine and carbohydrate are established ergogenic strategies for physical performance, it remains unclear whether these benefits translate to the decision-making and sport-specific skills that determine team-sport performance under fatigue.

METHODS

PubMed, Web of Science Core Collection, and SPORTDiscus were searched on 28 July 2026. Thirty-one reports were retained; their reported analyzed sample sizes summed to 441 athletes. Twenty-three independent quantitative study clusters, with reported analyzed sample sizes summing to 341 athletes, contributed 67 effects. Multilevel random-effects models nested effect sizes within studies; primary assumptions were r = 0.5 and rho = 0.80. Inference used CR2 robust standard errors and Satterthwaite degrees of freedom.

RESULTS

Soccer supplied 11 of the 23 quantitative clusters; every other sport category supplied four or fewer. Broad caffeine was centered near zero (Hedges' g = 0.040, 95% CI -0.211 to 0.291; p = 0.691; df = 4.41). Tier 1a perceptual-motor and Tier 1b technical/tactical estimates were imprecise. Broad carbohydrate was positive but uncertain (g = 0.256, 95% CI -0.094 to 0.605; p = 0.118; df = 4.96), and one study cluster materially influenced this estimate. All six GRADE comparisons were rated very low certainty. The two-study mental-fatigue calculation was too fragile for conventional inference.

CONCLUSIONS

Current evidence does not support assuming that the established physical ergogenic effects of caffeine translate to fatigue-related decision-making or sport-specific skill. Carbohydrate may provide a small benefit, but the evidence remains sparse and highly uncertain. Sensitivity to study composition was driven primarily by one influential cluster, whose omission increased the pooled estimate. The carbohydrate estimates apply only to heterogeneous ingested-beverage protocols and should not be extrapolated to mouth-rinse or gel strategies. Parallel syntheses do not establish intervention superiority or combined-strategy synergy.