Gut microbiota, diet, and cognitive function in community-dwelling older adults: an association with Bifidobacterium in the Kyotango longevity cohort

Age-related changes in the gut microbiota have been implicated in cognitive aging, however population-based evidence linking specific microbial taxa, dietary factors, and cognitive function in older adults is still limited. We investigated the associations between gut microbial composition and cognitive function, and explored the dietary and nutritional correlates of cognition-associated microbial features in older adults living in the community. This cross-sectional study included 578 participants from the Kyotango longevity cohort after exclusion of nine participants with implausible reported energy intake. Cognitive function was assessed using the Mini-Mental State Examination (MMSE). The composition of the gut microbiota was characterized using 16S rRNA gene sequencing. Forty-seven genus- or family-level taxa were retained after taxonomic aggregation, abundance/prevalence filtering, centered log-ratio transformation, and standardization. Microbial associations with MMSE scores were evaluated using Spearman correlations with false discovery rate correction and prespecified multivariable linear regression. Robustness was assessed using heteroscedasticity-robust inference, bootstrap resampling, ordinal regression, penalized regression, and sensitivity analyses. Dietary and nutrient correlates of Bifidobacterium were examined in 535 participants with MMSE > 23. Statistical indirect-effect analyses were considered exploratory. Microbiota-wide analysis identified a weak positive association of Bifidobacterium with MMSE (Spearman ρ = 0.141, false discovery rate q = 0.030) and a weak inverse association of Enterobacteriaceae (ρ = − 0.128, q = 0.050; unrounded q = 0.049988). After prespecified multivariable adjustment, Bifidobacterium remained associated with higher MMSE (B = 0.290, 95% confidence interval 0.079–0.500), but its incremental contribution to explained variance was small (R2, 0.130 to 0.142; ΔR2 = 0.011) and the association was attenuated after excluding participants with MMSE ≤ 23. In the MMSE > 23 dietary subset, milk/yogurt and several dairy-related nutrient measures correlated with Bifidobacterium , whereas no individual food or nutrient showed an FDR-significant direct association with MMSE. Exploratory indirect-effect analyses did not establish a unique temporal ordering among age, Bifidobacterium , and MMSE. In community-dwelling older adults, higher Bifidobacterium abundance was modestly and independently associated with better cognitive function as assessed using the MMSE. Dietary exposures were associated with Bifidobacterium , but direct food–MMSE analyses yielded no FDR-significant associations and the milk/yogurt → Bifidobacterium → MMSE indirect association was not statistically detectable. The cross-sectional design precludes conclusions regarding temporal direction or causal mediation. Longitudinal and interventional studies are required to determine the biological and clinical relevance of these associations. The UMIN Clinical Trials Registry (UMIN-CTR), UMIN000019486 and UMIN000045216, registered on December 1, 2015, and August 22, 2021, respectively.