Associations between Diet, the Gut Microbiome, and Short-Chain Fatty Acid Production among Older Caribbean Latino Adults

Background: Caribbean Latino adults have disproportionately high prevalence of chronic disease; however, underlying mechanisms are unknown. Unique gut microbiome profiles and relation to dietary quality may underlie health disparities.

Objectives: To examine the dietary quality of an underrepresented group of Caribbean Latino older adults with high prevalence of chronic disease; characterize gut microbiome profiles in this cohort; determine associations between dietary quality, gut microbiome composition, and short-chain fatty acid (SCFA) production; examine associations of clinical factors (body mass index, type 2 diabetes [T2D] status, and laxative use) with gut microbiome composition.

Design: The study design was cross-sectional.

Participants/setting: Recruitment and interviews occurred at the Senior Center in Lawrence, MA, from September 2016-September 2017. A total of 20 adults aged ≥50 years, self-identified of Caribbean Latino origin, without use of antibiotics in 6 months or intestinal surgery were included in the study.

Exposure and outcome measures: Diet was assessed by two, 24-hour recalls and dietary quality was calculated using the Healthy Eating Index 2015 and the Mediterranean Diet Score. The gut microbiome was assessed by 16S rRNA sequencing and fecal SCFA content. Anthropometrics (ie, weight and height) were measured by a trained interviewer, and self-reported laxative use, and other self-report health outcomes (ie, T2D status) were assessed by questionnaire.

Statistical analyses: Faith Phylogenetic Diversity (alpha diversity) and unique fraction metric, or UniFrac (beta diversity) and nonphylogenetic metrics, including Shannon diversity index (alpha diversity) were calculated. Spearman correlations and group comparisons using Kruskal-Wallis test between alpha diversity indexes and nutrient intakes were calculated. Patterns in the microbiome were estimated using a partitioning around medoids with estimation of number of clusters, with optimum average silhouette width. Log odds were calculated to compare predefined nutrients and diet score components between microbiome clusters using multivariable logistic regression, controlling for age and sex. Pearson correlation was used to relate SCFA fecal content to individual nutrients and diet indexes. Final models were additionally adjusted for laxative use. Differences in lifestyle factors by gut microbiome cluster were tested by Fisher's exact test.

Results: Generally, there was poor alignment of participant's diets to either the Mediterranean Diet score or Healthy Eating Index 2015. Range in the Healthy Eating Index 2015 was 36 to 90, where only 5% (n=1) of the sample showed high adherence to the Dietary Guidelines for Americans. Mediterranean Diet scores suggested low conformance with a Mediterranean eating pattern (score range=2 to 8, where 45% scored ≤3 [poor adherence]). The gut microbiome separated into two clusters by difference in a single bacterial taxon: Prevotella copri (P copri) (permutational multivariate analysis of variance [PERMANOVA] R²=0.576, ADONIS function P=0.001). Significantly lower P copri abundance was observed in cluster 1 compared with cluster 2 (Mann-Whitney P<0.0001). Samples in the P copri dominated cluster 2 showed significantly lower alpha diversity compared with P copri depleted cluster 1 (Shannon diversity index P=0.01). Individuals in the P copri dominated cluster showed a trend toward higher 18:3 α-linolenic fatty acid intakes (P=0.09). Percentage of energy from total fat intake was significantly, positively correlated with fecal acetate (r=0.46; P=0.04), butyrate (r=0.50; P=0.03) and propionate (r=0.52; P=0.02). Associations between dietary intake and composition of the gut microbiome were attenuated by self-report recent laxative use. Individuals with T2D exhibited a significantly greater abundance of the Enterobacteriales (P=0.01) and a trend toward lower fecal content of butyric acid compared to subjects without T2D (P=0.08). Significant beta diversity differences were observed by weight (Mantel P<0.003) and body mass index (Mantel P<0.07).

Conclusions: Two unique microbiome profiles, identified by abundance of P copri, were identified among Caribbean Latino adults. Microbiome profiles and SCFA content were associated with diet, T2D, and lifestyle. Further research is needed to determine the role of P copri and SCFA production in the risk for chronic disease and associated lifestyle predictors.