Evaluating Creatine for Female Athletes: A Meta-Analysis of Performance and Body Composition

While creatine is one of the most rigorously tested sports supplements available, much of the foundational literature historically centered on male cohorts. In recent years, female-specific trials have attempted to map out how creatine impacts women across different sports, training states, and physiological phases. However, synthesizing this data has proven challenging due to wide variations in study design. The preceding systematic review in this area gathered 27 studies and could only describe their performance results narratively, which left the average magnitude of creatine's effect across those correlated outcomes unknown.

A newly published systematic review and multilevel meta-analysis¹ set out to close that gap by estimating the average magnitude of creatine's effects on exercise performance, physiological outcomes, and body composition specifically in females engaged in exercise, sport, or structured training.

Key Takeaways

• Small performance benefits: The study identified a small positive effect of creatine supplementation on selected strength, power, and high-intensity performance measures in exercising females.
• Modest body composition changes: Creatine supplementation showed small effects on lean mass-related measures, though the included outcomes lacked a uniform favorable direction.
• Low certainty of evidence: Due to significant variation in study protocols, risk of bias, and methodological limitations, certainty was graded low for performance and body composition, and very low for the overall and physiological estimates.

The Study Design

The researchers searched five databases — PubMed, Scopus, Web of Science, SPORTDiscus, and CINAHL — for experimental studies published between 1 January 1996 and 5 March 2026. Two reviewers independently screened the records, extracted the data, and appraised each study's risk of bias. Of 13,244 records identified and 160 reports assessed in full text, 34 studies comprising 692 female or female-eligible participants entered the systematic review.

Of these, 21 studies were compatible with the researchers' multilevel random-effects models, contributing 170 distinct effects from 424 analyzable participants. The research team applied CR2 small-sample robust inference to handle the complex, correlated data, ran sensitivity analyses covering graph-digitized and reconstructed percentage-change data as well as plausible within-study sampling correlations, and utilized the GRADE² framework to assess the certainty of the overall estimate and each quantitative domain.

The Findings

The overall model yielded a small positive effect for creatine supplementation in females (Hedges g³ = 0.23, CR2 robust 95% CI 0.09 to 0.37). When broken down into specific domains, the results remained modest:

• Performance: A small positive effect (g = 0.30, 95% CI 0.06 to 0.54) on parameters like strength, power, and high-intensity performance.
• Body Composition: A small positive effect (g = 0.24, 95% CI 0.06 to 0.43) on lean mass and related metrics, though the researchers noted that these outcomes were combined without a uniform favorable direction.
• Physiological Outcomes: The estimate for physiological effects was uncertain (g = 0.10, 95% CI -0.17 to 0.38), with the interval crossing zero.

Critically, the certainty of the evidence was graded as "low" for performance and body composition, and "very low" for the overall and physiological estimates.

Analyzing the Study: Strengths & Limitations

Strengths and Reputability

This study earned a top score for design rigor in our weighted rubric, and the analytical machinery behind it is genuinely careful. Two reviewers independently screened, extracted and risk-of-bias appraised every study rather than one person working alone. The multilevel random-effects models let the authors account for multiple, correlated outcomes reported within the same primary study—a common issue in sports science literature that simpler meta-analyses mishandle by treating every effect as independent. CR2 small-sample robust inference and the sensitivity analyses further guard the estimates against the modest number of contributing studies.

Two caveats belong alongside that score. The review was registered on the Open Science Framework, but the authors state plainly that the registration was retrospective—filed after the work rather than before it. Prospective registration is what stops analytical goalposts from moving once the data are in view; a retrospective entry documents the protocol publicly but cannot make that guarantee. Second, the pool is of experimental creatine-versus-control studies, not exclusively randomized controlled trials, so the synthesis inherits whatever control and allocation weaknesses the primary literature carries.

Limitations

Despite the rigorous statistical design, the meta-analysis is bottlenecked by the quality and size of its source material. Pooling data from three decades of research yielded fewer than 1,000 total participants (n=692), with only 424 making it into the primary analysis models. Some of the pooled effects were not taken from the papers directly but digitized from published graphs or reconstructed from reported percentage changes—the authors ran sensitivity analyses precisely because those inputs are approximations rather than reported numbers. Substantial heterogeneity exists among the included studies, spanning different sports, training states, and protocols. Additionally, risk-of-bias concerns within the primary literature heavily downgraded the certainty of the findings, making it impossible to draw firm, sport-specific conclusions.

Conflicts of Interest

The authors declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The abstract does not state who funded the work.

References

1. Lin Y, Zhu W, Hong B. Effects of creatine supplementation on exercise performance, physiological outcomes, and body composition in females engaged in exercise or training: a systematic review and multilevel meta-analysis. Frontiers in nutrition. 10.3389/fnut.2026.1921827

Glossary

1. Meta-analysis: A statistical analysis that combines the results of multiple scientific studies to improve estimates of the size of the effect and resolve uncertainty. ↩

2. GRADE: Grading of Recommendations, Assessment, Development, and Evaluations. A transparent framework for developing and presenting summaries of evidence and providing systematic approaches for making clinical practice recommendations. ↩

3. Hedges g: A statistical measure of effect size used in meta-analyses, indicating the standardized difference between two means. An effect size of 0.2 is generally considered small. ↩