Intestinal Microbial Composition of Children in a Randomized Controlled Trial of Probiotics to Treat Acute Gastroenteritis

Abstract

Compositional analysis of the intestinal microbiome in pre-schoolers is understudied. Effects of probiotics on the gut microbiota were evaluated in children under 4-years-old presenting to an emergency department with acute gastroenteritis. Included were 70 study participants (n=32 placebo, n=38 probiotics) with stool specimens at baseline (day 0), day 5, and after a washout period (day 28). Microbiota composition and deduced functions were profiled using 16S ribosomal RNA sequencing and predictive metagenomics, respectively. Probiotics were detected at day 5 of administration but otherwise had no discernable effects, whereas detection of bacterial infection (P<0.001) and participant age (P<0.001) had the largest effects on microbiota composition, microbial diversity, and deduced bacterial functions. Participants under 1 year had lower bacterial diversity than older aged pre-schoolers; compositional changes of individual bacterial taxa were associated with maturation of the gut microbiota. Advances in age were associated with differences in gut microbiota composition and deduced microbial functions, which have the potential to impact health later in life.

Clinical trial registration: www.ClinicalTrials.gov, identifier: NCT01853124.

Keywords: bacteria; children; gastroenteritis; intestine; lactobacillus; microbiome; probiotics.

Conflict of interest statement

SF is supported by the Alberta Children’s Hospital Professorship in Child Health and Wellness. MS is the recipient of a Canadian Research Chair in Interdisciplinary Microbiome Research. YF is the recipient of the Canada Research Chair in Pediatric Drug Safety and Efficacy. PS is the recipient of a Canadian Research Chair in Gastrointestinal Disease and research funded by the Canadian Institutes of Health Research (MOP-89894 and IOP-92890) and received honoraria from Abbott Nutrition, Mead Johnson Nutritionals and Nestlé Nutrition. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2022 Horne, Freedman, Johnson-Henry, Pang, Lee, Farion, Gouin, Schuh, Poonai, Hurley, Finkelstein, Xie, Williamson-Urquhart, Chui, Rossi, Surette and Sherman.

Figures

Figure 1

Genus level bacteria taxa during treatment with probiotics. (A) Center log ratio normalized abundance of ASV aggregated to Lactobacillus at the genus level; (B) center log ratio normalized abundance for Lactobacillus ASV2; and (C) centre log ratio normalized abundance for Lactobacillus ASV13. (D) Linear discriminant analysis effect size (LEfSe) analysis identified the most differentially abundant taxa between probiotic and placebo measured at study day 5; (E–G) using linear mix model regression, center log ratios with normalized abundance of individual amplicon sequence variants (ASV) were significantly associated with both probiotic intervention and changes across time. Levels of Akkermansia increased between study day 0 and day 5 in participants randomly assigned to the probiotic treatment study group. Significance was assessed using linear mixed modeling, pairwise comparisons were performed using estimated marginal means, with Tukey’s multiple comparison testing. Significance is denoted as *P < 0.05, **P < 0.01, ***P < 0.001 versus placebo; NS denotes not statistically significant.

Figure 2

Probiotic treatment does not significantly alter gut microbiota composition or diversity. Stool samples were collected at the time of acute enteritis and entry into the study (day 0), 5 days after either probiotic or placebo (day 5) and 28 days following entry into the study (day 28). (A) alpha diversity, as measured by Shannon index; (B) alpha diversity species richness metric Chao1; (C) principal coordinate analysis of beta diversity, measured by Weighted Unifrac distances; (D) principal coordinate analysis assessed by unweighted Unifrac distances. Statistical significance was assessed using repeated measure (PERMANOVA); (E) relative abundance of gut bacteria characterized at the phylum level; and (F) relative abundance of the top 20 genus level taxa. Significance was assessed using two-sided permutation t-test, with multiple comparisons corrected by false discovery rate (FDR).

Figure 3

Detection of bacterial enteric pathogens and participant age both impact on the diversity of gut microbiota composition. (A) differences in alpha diversity associated with pathogen carriage status; (B) alpha diversity of age categories, as measured by the Shannon diversity index; (C) alpha diversity, measured by Shannon index, between probiotic and placebo treat groups within age categories. Statistical significance was assessed by two-way ANOVA, with Tukey’s multiple comparison testing; (D) principal coordinate analysis of unweighted Unifrac distances. Significance was assessed by using PERMANOVA; (E) pairwise dissimilarity distance comparison between age categories on study day 0, day 5 and day 28 using unweighted Unifrac distances. One-way ANOVA followed by Tukey’s multiple comparison testing; and (F) relative abundance of the top 20 genus level taxa, with statistically significant differences between age groups for Bifidobacterium, Blautia, Streptococcus, Anaereostipes and Klebsiella. Significance denoted by *P < 0.05, **P < 0.01 and ***P < 0.001. NS indicates no statistical significance.

Figure 4

Changes in the gut microbiota occur with increasing chronological age. (A) heat map of bacterial taxa, with significant interactions between age of study participants and sampling day, as determined by linear mixed model regression; (B–E) mean center log normalized abundances of specific bacteria; and (F) Venn diagram comparison of core ASV between age categories (<1.0 year, 1.0 -2.0 years, and >2.0 - <4.0 years old), determined at study entry (day 0).

Figure 5

Predicted functional changes in gut bacteria across study periods and between age categories. (A) heat map of center log ratio normalized abundance of functional pathways were significantly associated with participant age category and sampling day, as determined by linear mixed model regression; (B) principal component analysis of predicted functional pathways at all sampling time points for study participants <1.0 year and those >2.0 - <4.0 years of age; (C) principal component analysis of all subjects at all time points; (D) extended error plot of abundance of functional pathways greater in children >2.0 - <4.0 years old compared to those <1.0 year; and (E) extended error plot of abundance of functional pathways, which were increase in infants <1.0 year. Statistical significance was assessed by one-sided Welch’s t-test, with multiple comparisons corrected by FDR.

Figure 6

Relationships between the age of study participants, fecal sIgA concentration and genus level bacterial taxa abundance. (A) log transformed stool sIgA levels (μg/mL) at entry into the study (day 0) and days 5 and 28 post randomization compared between age categories (under one year of age (<1.0 yr), between one and two years of age (1.0 – 2.0 yrs), and over 2.0 and under 4.0 years of age (>2 yrs). Values are expressed as mean ± SEM. Statistical significance was assessed by using mixed model regression; (B) correlation heat map representing Spearman correlation coefficients between changes in center log ratio normalized abundance of bacterial taxa (at the genus level) and log stool IgA levels stratified by age category. Significance is denoted as: *P < 0.05, **P < 0.01 and ***P < 0.001, with multiple comparisons corrected by FDR.