In 173 nursing-home residents averaging 86 years of age (77.5% female), diabetes mellitus — present in 51 participants with a mean HbA1c of 6% — selectively depleted four bacterial taxa: Clostridium_Clostridiaceae, Haemophilus, Actinomycetaceae, and Micrococcaceae. These differences persisted after adjusting for age, sex, and BMI. Critically, neither alpha diversity (Shannon index p=0.747) nor beta diversity shifted with DM status, though nutritional risk assessed by the Geriatric Nutritional Risk Index (GNRI) produced a marginal beta-diversity signal (R²=0.009, p=0.032). No interaction between DM and GNRI was detected.
The finding that diabetes reshapes microbial composition without altering diversity metrics is clinically meaningful and methodologically underappreciated. Standard diversity indices can mask targeted bacterial losses that may still carry functional consequences — particularly relevant for Clostridiales, whose members produce short-chain fatty acids critical to intestinal barrier integrity and immune regulation. This population represents a crucial but chronically understudied demographic; most microbiome-diabetes research draws from younger, community-dwelling cohorts where metabolic disease is more aggressive and HbA1c values higher. The near-normal glycemic control here (mean HbA1c 6%) suggests medication effects — notably metformin, a known microbiome modulator — may confound the signal. The study cannot establish causality, the GNRI effect size is negligible, and 16S rRNA sequencing limits functional resolution. Still, this is a confirmatory and pragmatically important contribution, reinforcing that diabetes leaves a compositional fingerprint even in well-controlled, frail elderly populations.
