Effects of Lactobacillus reuteri supplementation on the gut microbiota in extremely preterm infants in a randomized placebo-controlled trial

Abstract

Extremely low birth weight (ELBW) infants often develop an altered gut microbiota composition, which is related to clinical complications, such as necrotizing enterocolitis and sepsis. Probiotic supplementation may reduce these complications, and modulation of the gut microbiome is a potential mechanism underlying the probiotic effectiveness. In a randomized, double-blind, placebo-controlled trial, we assessed the effect of Lactobacillus reuteri supplementation, from birth to post-menstrual week (PMW)36, on infant gut microbiota. We performed 16S amplicon sequencing in 558 stool samples from 132 ELBW preterm infants at 1 week, 2 weeks, 3 weeks, 4 weeks, PMW36, and 2 years. Probiotic supplementation results in increased bacterial diversity and increased L. reuteri abundance during the 1^st month. At 1 week, probiotic supplementation also results in a lower abundance of Enterobacteriaceae and Staphylococcaceae. No effects were found at 2 years. In conclusion, probiotics may exert benefits by modulating the gut microbiota composition during the 1^st month in ELBW infants.

Trial registration: ClinicalTrials.gov NCT01603368.

Keywords: Lactobacillus; Staphylococcus; extremely low birth weight infant; microbial diversity; microbiota; necrotizing enterocolitis; preterm infant; probiotic; randomized controlled trial; supplementation.

Graphical abstract

Figure 1

Flow chart of the study PMW36: post-menstrual week 36. ^aStudy product was discontinued by mistake after transfer to other hospital (n = 1). ^bStudy product was not administrated again by mistake after temporarily being withheld during nil oral (n = 3). ^cStudy product ran out temporarily at the study site (n = 3). ^dInsufficient amounts of DNA were recovered from extraction (n = 6). ^eLibrary preparation failed (n = 6). ^fSequencing failed (n = 12).

Figure 2

Gut microbiota α-diversity of ELBW preterm infants supplemented with L. reuteri or placebo Boxplots (median with 25% and 75% percentiles and 1.5× the interquartile range; diamond shape depicts the mean) showing the α-diversity (Shannon index), richness (observed ASVs), and evenness (Pielou’s evenness index) from 1 week to 2 years of life in ELBW preterm infants supplemented with L. reuteri (Lr) or placebo (Pl). PMW36, post-menstrual week 36; w, week; y, year. ∗∗∗p < 0.001, ∗∗p < 0.01, and ∗p < 0.05 with Mann-Whitney U test and p value adjustment for multiple comparisons with the method from Benjamini and Hochberg.

Figure 3

Clustering of the gut microbiota composition (β-diversity) of ELBW preterm infants supplemented with L. reuteri or placebo Non-metric multidimensional scaling (NMDS) of bacterial community composition from 1 week to 2 years of life across ELBW preterm infants supplemented with L. reuteri or placebo. (A–F) The ASVs that significantly contributed to the variance explained (envfit(); p < 0.01 and R² > 0.3) were classified at genus level, and only one genus for all ASVs pointing toward the same direction was displayed (Table S6). At 1 week, (A) the ellipses (confidence level 0.95) show Linköping and Stockholm because inclusion site also affected the bacterial community composition and Lactobacillus had different effects depending on the site. (G–I) The abundance of L. reuteri DSM 17938 (qPCR data) significantly (envfit(); p < 0.01 and R² > 0.3) correlated to bacterial community composition in the placebo group. Weight, length, and head circumference were adjusted for gestational age using the standard deviation score (Z score). ∗∗∗p < 0.001 with ANOSIM and p value adjustment for multiple comparisons with the method from Benjamini and Hochberg.

Figure 4

Taxonomic composition of the gut bacteria in ELBW preterm infants supplemented with L. reuteri or placebo Relative abundance of the dominant taxa is displayed at phylum (A), family (B), and genus (C) level. At family and genus levels, the taxa with a relative abundance of <1% across all samples and time points are grouped in “others.” ∗∗ indicates taxa that significantly differed in relative abundance between the L. reuteri and placebo groups (LEfSE; p = 0.01; Table S3).

Figure 5

Prevalence and abundance of L. reuteri DSM 17938 Prevalence (A) and abundance (B) of supplemented L. reuteri DSM 17938 in L. reuteri and placebo groups at different time points. (A) Percentage of infants with a stool sample positive for the supplemented strain. (B) Boxplots (median with 25% and 75% percentiles and 1.5× the interquartile range) show the abundance as the number of L. reuteri DSM 17938 bacteria per 1 g wet feces. Colored dots indicate the L. reuteri DSM 17938 abundance in individual stool samples positive for the supplemented L. reuteri strain; the number of L. reuteri DSM 17938 bacteria per 1 g wet feces for infants with a L. reuteri negative stool sample was set to 1 for graphical display; (n) indicates the number of infants with a stool sample positive for supplemented L. reuteri DSM 17938; (N) indicates the total number of infants with a stool sample in the L. reuteri or placebo group at the indicated time point. Prevalence and abundance between groups were compared using Fisher’s exact tests and Mann-Whitney U tests, respectively, and adjusted for multiple comparisons with the method from Benjamini and Hochberg. Significant differences in L. reuteri DSM 17938 abundance in the L. reuteri group across neonatal time points (1 week to PMW36) were tested for with a Kruskal-Wallis test with Dunn post hoc test and p value adjustment for multiple comparisons with the method from Benjamini and Hochberg. ∗∗∗p < 0.001; ∗∗p < 0.01.

Figure 6

Correlation between growth parameters and ELBW preterm infant gut microbiota composition (A and B) Head growth until 4 weeks of life correlated to microbial diversity at 1 week (simple linear regression; p = 0.007; adjusted R² = 0.06; A) and microbial richness at 2 weeks (simple linear regression; p = 0.035; adjusted R² = 0.03; B). (C and D) Head (head) and weight (weight) growth rate significantly (envfit(); p < 0.05 and R² 0.1–0.2) correlated to the microbial community composition at 1 week (C) and 3 weeks (D) of life.