Role of insulinemic and inflammatory dietary patterns on gut microbial composition and circulating biomarkers of metabolic health among older American men

Abstract

Chronic low-grade inflammation and hyperinsulinemia are linked with metabolic dysfunction and dysbiosis. This study investigated the role of dietary inflammatory and insulinemic potential on gut microbiome and circulating health biomarkers in older men. Data from the Osteoporotic Fractures in Men (MrOS) study were analyzed. Reversed Empirical Dietary Inflammatory Pattern (rEDIP), Empirical Dietary Index for Hyperinsulinemia (rEDIH), and Healthy Eating Index (HEI)-2020 scores were computed from food frequency questionnaire data. Stool samples were profiled using 16S rRNA sequencing. Elastic net regression identified diet-associated microbial profiles and multivariable-adjusted linear regression assessed diet-biomarker associations. Higher rEDIP, rEDIH, and HEI-2020 scores were positively associated with gut microbiota alpha diversity. Specific genera, including Intestinibacter and Lachnospira, associated positively, while Dielma, Peptococcus, Feacalitalea, and Negativibaccilus associated inversely with healthier dietary patterns. When evaluating changes in dietary patterns between baseline and visit 4 ( ~ 14 years), these genera tended to define rEDIP, rEDIH more than HEI-2020. In addition, higher dietary quality was linked to better biomarker profiles, including lower creatinine, sodium, triglycerides, and insulin resistance. Beneficial effects of higher dietary quality on health may be mediated by the ability of diet to regulate gut microbial composition and metabolic biomarker profiles.

Keywords: Dietary patterns; biomarkers of chronic diseases; gut microbiome.

Figure 1.

Combined flow chart of study design and participant selection: MrOS study cohort.

Figure 2.

Cross-sectional, longitudinal, and change-in-dietary-score analyses of dietary patterns and gut microbiome alpha diversity measures. The absolute values, predicted min, max, for Shannon, inverse Simpson, and Pielou’s evenness in each study type are presented in supplementary table 2. PD, percentage difference; SD, standard deviation; CI, confidence interval.

Figure 3.

Venn diagrams showing the number of microbial genera identified from elastic net regression analyses, across the three dietary patterns in a) cross-sectional, b) longitudinal, and c) change in dietary score analyses. reversed empirical dietary index for hyperinsulinemia (rEDIH), reversed empirical dietary index for inflammation (rEDIP), and the healthy eating index-2020 (HEI_2020).

Figure 4.

Microbial genera selected by elastic net regression in a) cross-sectional, b) longitudinal, and c) change-in-dietary-score analyses, aggregated at the genus level for each of the three dietary patterns (rEDIH, rEDIP, and HEI-2020). Each number within the heatmap represents the beta coefficient (expressed as a percentage), indicating the strength of the association between a specific microbial genus and the dietary pattern. Green shades indicate genera associated with healthier diets, while red/brown shades represent genera linked to more hyperinsulinemic, pro-inflammatory, and lower-quality diets. In addition, the genera presented in the figures represent the top “hits” representing 90% of the variance in the total number of genera associated with the dietary pattern. The cross-sectional analysis represents the primary analysis.

Figure 5.

Associations of dietary patterns (per 1 standard deviation increment) with biomarkers of renal functions, lipids, and insulin response. Values are absolute back-transformed biomarker concentrations (beta coefficients). Beta coefficients were derived from MV (multivariable) adjusted linear regression models. MV models were adjusted for total energy intake, BMI-continuous, age, physical activity, smoking status, number of supplement used, sex, race/ethnic groups, education levels, use of NSAIDS, marital status, alcohol drinking status, family history of cancer, and diabetes status. HDL, high-density lipoprotein; LDL, low-density lipoprotein; TC, total cholesterol; HOMA-IR, homeostatic model assessment of insulin resistance.