Real-world creatine supplementation: a large-scale cross-sectional study of use, knowledge, and experiences

Abstract

Background: Creatine is among the most extensively studied dietary supplements in sports nutrition, yet real-world supplementation patterns, knowledge, and outcomes remain poorly characterized. Prior survey research has focused on competitive athletes and younger male samples, even as the evidence base has expanded to support applications in healthy aging, cognition, and female physiology. Whether use patterns, perceived benefits, and adverse effects among the broader and more diverse population now using creatine align with trial evidence, and whether they differ by gender identity and age, has not been systematically assessed. Methods: An online cross-sectional survey was distributed to approximately 110,000 users of a supplement tracking application (SuppCo, Inc.) with recorded creatine use, yielding 2,178 analyzable responses from individuals reporting current or past-12-month creatine use (1,077 women and 1,101 men; median age, 55 years). The survey assessed supplementation patterns, goals, creatine knowledge (six-item battery scored via generalized partial credit item response theory [IRT] modeling), side effects, satisfaction, and information needs. Multivariable analyses included logistic regression for perceived benefit, modified Poisson regression for side-effect prevalence, and OLS regression for satisfaction, with Benjamini-Hochberg correction for multiple comparisons. As an exploratory analysis, a mixed graphical model (MGM) characterized conditional associations across measurement domains, and IRT discrimination parameters were combined with network centrality to identify candidate keystone misconceptions. Results: Despite 98% intending to continue supplementation and mean satisfaction of 7.8/10, only 55% perceived benefit. Cognitive performance was the most commonly endorsed goal (67.8%) yet showed the largest within-person unmet-expectation rate (64.4% of cognitive-goal endorsers did not perceive a matched cognitive benefit; McNemar p