Dietary and Lifestyle Correlates of Gut Microbiota Diversity and Composition in Healthy Adults: A Cross-Sectional Pilot Study Informing Longitudinal Research

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Background/Objectives: Modifiable dietary and lifestyle factors may contribute to gut microbiota variation but are often evaluated separately. This pilot study examined four exposures across complementary microbiome domains and used association magnitude and uncertainty to inform longitudinal study design. Methods: Twenty healthy adults (10 females and 10 males) provided a fresh-stool specimen and completed the exposure questionnaire on the day of specimen production. Exposures comprised Mediterranean diet adherence, ultra-processed food consumption frequency, physical activity, and sleep characteristics. Microbiome outcomes comprised independently recalculated OTU-level Shannon diversity (primary outcome) and Pielou’s evenness, genus-level Bray–Curtis and Aitchison distances, centered log-ratio coordinates of eight prespecified genera, and the natural-log-transformed Prevotella/Bacteroides ratio. Associations were evaluated using age- and sex-adjusted linear regression and permutational multivariate analysis of variance, with false-discovery-rate control, influence diagnostics, and robustness analyses. Scenario-based sample-size calculations were restricted to age- and sex-adjusted cross-sectional linear regression models. Results: Complete exposure and microbiome data were available for all participants. No association was supported after false-discovery-rate correction. However, the observed effect estimates were not uniformly negligible: maximum partial R2 values were 0.140 across the alpha-diversity models, 0.210 for the UPF–Akkermansia association, and 0.108 for the Prevotella/Bacteroides log-ratio, whereas the maximum marginal R2 for overall community composition was 0.0649. Estimates were imprecise and participant-sensitive, and robustness analyses did not alter the overall interpretation. Conservative scenarios for age- and sex-adjusted cross-sectional regression models, based on assumed partial R2 values of 0.03–0.05, yielded planning benchmarks of 275–465 analyzable participants, or 306–582 enrolled participants assuming 10–20% unusable observations. Conclusions: This pilot does not establish the absence of diet–microbiota or lifestyle–microbiota relationships. Instead, it provides a reproducible framework for prioritizing hypotheses and planning adequately powered longitudinal studies using repeated, temporally aligned exposure and microbiome measurements.

This publication has 41 references indexed in Scilit: