Effects of Blueberry Consumption on Fecal Microbiome Composition and Circulating Metabolites, Lipids, and Lipoproteins in a Randomized Controlled Trial of Older Adults with Overweight or Obesity: The BEACTIVE Trial

Abstract

Background/Objectives: Generous consumption of phytonutrient-rich foods, including blueberries, provides benefits to multiple physiologic and metabolic systems. This study explored the potential that regular, generous blueberry intake could favorably modulate fecal microbiome composition in sedentary older (>60 years) men and women with overweight or obesity (BMI ≥ 25 to 32 kg/m²). Methods: Participants (n = 55) were randomized to daily consumption of either lyophilized blueberry powder (equivalent to 1.5 cups of blueberries) or an indistinguishable placebo powder; both groups participated in weekly supervised exercise classes. Fecal samples were collected at 0 and 12 weeks and frozen. Following this, 16S rRNA gene sequencing was used to profile each participant's fecal microbiome. Blood biomarkers of cardiometabolic health were measured via nuclear magnetic resonance spectroscopy (NMR) pre- and post-treatment. Results: Comparing the baseline and endpoint results for the blueberry (n = 15) and placebo (n = 19) groups, there were no significant overall compositional differences or differences in the level of diversity in the fecal microbiome. However, in subjects whose diet included blueberry powder, there was a significant enrichment (p = 0.049) in the relative abundance of Coriobacteriales incertae sedis, a taxonomic group of bacteria that facilitates the metabolism of dietary polyphenols. The placebo group exhibited significant reductions in total cholesterol, LDL-C, non-HDL-C, total LDL-P, large LDL-P, and ApoB, while the blueberry group exhibited significant reductions in total HDL-P and ApoA-I after 12 weeks compared to baseline. Conclusions: Generous blueberry consumption may upregulate the ability of the older human gut to utilize dietary polyphenols by altering the fecal microbiome. Longer, larger-scale studies with blueberries or blueberry powder are needed to observe improvements in cardiometabolic risk factors in older adults with overweight or obesity.

Keywords: blueberry intake; cardiometabolic risk; fecal microbiome; gut microbiome; lipoproteins; nuclear magnetic resonance; polyphenols.

Figure 1

Consort diagram for the BEACTIVE study.

Figure 2

Maximum likelihood phylogeny of all amplicon sequence variants (ASVs), with branches colored by class. The number of ASVs and the proportion of ASVs in each class is noted.

Figure 3

Stacked bar plot of relative abundances at (a) phylum, (b) class, (c) order, (d) family, and (e) genus taxonomic levels in blueberry (left of each plot) and placebo (right of each plot) samples. Samples were ordered by decreasing mean relative abundance of the most abundant taxa across baseline/endpoint, with participants separated by vertical lines. Each participant has baseline and endpoint samples grouped together from left to right. Taxa were sorted by mean relative abundance, and only those with at least 2.5% prevalence in at least one sample were kept; otherwise, they were aggregated in the “other” category. At most, the top ten taxa were shown in each plot, with all others aggregated in the “other” category.

Figure 4

Composition and diversity analysis. (a) Samples clustered using pairwise Bray–Curtis dissimilarities of species-resolved microbiome profiles by participants. Microbes without species-level resolution were aggregated at the lowest taxonomic rank available. Tips were colored and labelled using the participant ID and then appended with either ‘e’ for endpoint or ‘b’ for baseline. Arrows represent samples from the same participant that were not clustered together. (b) Principal coordinate analysis of Bray–Curtis dissimilarity on genus and species rank colored by the blueberry (left) and placebo (right) treatment groups and time points. Filled ellipse represents 95% confidence ellipses. PERMANOVA with p-value and R-squared are labeled on each of the subfigures. (c) The Shannon diversity was measured at the genus (238 unique genera taxa) and species (380 unique species taxa) level for both blueberry and placebo groups at the baseline and endpoint. The boxplot shows the median, first, and third quartiles, and the whiskers indicate ±1.5 × interquartile range. The p-value is displayed on top of the paired comparison boxes.

Figure 5

Differential abundance analysis of microbes. Volcano plots with log2 fold change (L2FC) (endpoint vs. baseline) on the x-axis and −log10-adjusted p-value (Benjamini–Hochberg correction) on the y-axis for two treatment groups; (a) blueberry group (left panel) and (b) placebo group (right panel). Data points represent taxa, which were aggregated using the lowest taxonomic rank classified for each amplicon sequence variants and are labeled with prefixes indicating the rank (‘s’ = species, ‘g’ = genus, ‘f’ = family). Colors indicate whether taxa are enriched (red, L2FC ≥ 1) or depleted (blue, L2FC ≤ −1) in endpoint samples relative to baseline samples. Light green taxa are those that did not have a large effect size (−1 < L2FC < 1). (c) Relative abundance of Coriobacteriales incertae sedis between baseline and endpoint samples for the blueberry group (left panel) and placebo group (right panel). Points represent individual samples. Lines connect samples from the same subject.