Low-dose oral caffeine supplementation may improve aerobic time-trial performance in healthy adults, according to a systematic review and meta-analysis published in Nutrients.
The review included 48 randomized, placebo-controlled crossover trials involving 689 healthy adults aged 18 to 59 years. Participants received acute pre-exercise oral anhydrous caffeine, administered as capsules or aqueous solutions, before aerobic time-trial tests lasting at least 3 minutes. Eligible studies had to report performance as time-to-completion outcomes.
The review used narrow inclusion criteria. Studies were excluded if they used coffee, energy drinks, chewing gum, mouth rinses, sprays, combination ergogenic supplements, fractionated dosing, time-to-exhaustion protocols, high-intensity interval training protocols, or non-time-based outcomes such as power output, total work, or distance covered.
Researchers searched PubMed, Embase, and the Virtual Health Library for studies published through July 1, 2022. Caffeine doses were categorized as low, approximately 1.3 to 3 mg/kg; moderate, 4 to 6 mg/kg; or high, greater than 6 mg/kg. The moderate-dose category reflected the available literature: no eligible studies used doses between 3.1 and 3.9 mg/kg.
Most included studies evaluated cycling performance. Of the 48 studies, 33 involved cycling, 10 involved running, 2 involved rowing, and 1 each involved skiing, swimming, and triathlon.
Low-dose caffeine was evaluated in 16 trials and was associated with reduced aerobic time-trial completion time compared with placebo, with a standardized mean difference (SMD) of −0.27 (95% CI, −0.44 to −0.11; p = .001; I² = 0%). Moderate-dose caffeine was evaluated in 36 trials and was also associated with reduced completion time, with an SMD of −0.52 (95% CI, −0.77 to −0.28; p < .0001; I² = 73%).
The larger SMD for moderate-dose caffeine should not be interpreted as a proportionally larger real-world performance gain, the investigators cautioned. When raw time-trial outcomes were assessed across eligible studies, low-dose and moderate-dose caffeine corresponded to mean completion-time reductions of approximately 2.14% and 2.18%, respectively.
In funnel plot analyses, two studies fell outside the 95% confidence interval limits, both in the moderate-dose category. The authors noted that Hodgson et al. used a single-blind design, while Guest et al. included a substantially larger sample than most eligible trials and had a low proportion of participants with the CYP1A2 CC genotype, which is commonly associated with slower caffeine metabolism and a potentially lower ergogenic response. In sensitivity analyses excluding those two studies, the authors reported that the apparent difference between low- and moderate-dose effects was partly attributable to these data, and cautioned that the moderate-dose estimate warrants careful interpretation as a result.
In an exploratory subgroup analysis limited to studies reporting VO₂max or VO₂peak, moderate-dose caffeine was associated with improved performance in both trained and highly trained participants. The authors noted this analysis was not confirmatory: it excluded studies that did not report VO₂max/VO₂peak data, as well as studies whose participants showed substantial variability in aerobic fitness that precluded clear classification by training status. Among trained participants, the effect estimate was SMD −0.81 (95% CI, −1.59 to −0.02; p = .04; I² = 91%). Among highly trained participants, the estimate was SMD −0.94 (95% CI, −1.35 to −0.53; p < .00001; I² = 0%). The between-subgroup difference was not statistically significant.
No eligible studies evaluated caffeine doses greater than 6 mg/kg using the review's required time-to-completion outcomes. The authors noted that high-dose caffeine has been studied in other exercise contexts, but many of those trials used time-to-exhaustion protocols, non-time-based outcomes, fractionated dosing, or lacked the mean and standard deviation data needed for this meta-analysis. As a result, the review does not establish whether higher-dose caffeine strategies improve aerobic time-trial performance or whether potential benefits would outweigh adverse effects.
The authors also highlighted interindividual variability in caffeine responsiveness. Citing other research, they noted that CYP1A2-related variants may contribute to differences in ergogenic response, with AA and AC genotypes more often associated with benefit after caffeine ingestion and the CC genotype, commonly classified as slower caffeine metabolism, associated with attenuated or less consistent responses. Evidence for other receptor-related polymorphisms remains limited.
Several limitations affect clinical interpretation. The sample was predominantly male, with 642 male participants and 47 female participants. Only 11 trials were classified as having low risk of bias, while 37 had unclear risk of bias, largely because of insufficient reporting of randomization or allocation procedures. Eight trials had high risk of bias related to blinding. The available data also did not support robust subgroup analyses by sex or caffeine timing.
The findings are most directly applicable to healthy adults aged 18 to 59 years engaging in aerobic time-trial exercise, the population studied in the included trials. The review did not quantify adverse events; the authors called for future studies evaluating the risk–benefit profile of high-dose caffeine in relation to both performance and adverse effects such as anxiety, heart palpitations, headaches, insomnia, and gastrointestinal symptoms.
"Pre-exercise use of low caffeine doses (1.3–3 mg/kg) can enhance generalized aerobic time-trial performance," wrote lead study author Gabriel L. Martins, of Plenitude Education Institute and the University of São Paulo, and colleagues. "In addition, the use of moderate caffeine doses (4–6 mg/kg) appears to promote a more consistent ergogenic effect."
The authors concluded that low and moderate caffeine doses may represent effective supplementation strategies for aerobic time-trial performance, while calling for future studies addressing individual sensitivity, genetic contributors to caffeine responsiveness, and high-dose risk–benefit profiles. The review also found insufficient evidence to support robust subgroup analyses by sex or caffeine timing.
Disclosures: Funding support was reported from the National Council for Scientific and Technological Development, São Paulo Research Foundation, and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior. The authors declared no conflicts of interest.
Source: Nutrients
