Impact of synbiotics on gut microbiota during early life: a randomized, double-blind study

Abstract

Human milk is considered the optimal nutrition for infants and found to contain significant numbers of viable bacteria. The aim of the study was to assess the effects of a specific synbiotic combination at doses closer to the bacterial cells present in human milk, on intestinal bifidobacteria proportions (relative abundance), reduction of potential pathogens and gut physiological conditions. A clinical study was conducted in 290 healthy infants aged from 6 to 19 weeks. Infants received either a control infant formula or one of the two investigational infant formulas (control formula with 0.8 g/100 ml scGOS/lcFOS and Bifidobacterium breve M-16V at either 1 × 104 cfu/ml or 1 × 106 cfu/ml). Exclusively breastfed infants were included as a reference. Analyses were performed on intention-to-treat groups and all-subjects-treated groups. After 6 weeks of intervention, the synbiotics at two different doses significantly increased the bifidobacteria proportions in healthy infants. The synbiotic supplementation also decreased the prevalence (infants with detectable levels) and the abundance of C. difficile. Closer to the levels in the breastfed reference group, fecal pH was significantly lower while L-lactate concentrations and acetate proportions were significantly higher in the synbiotic groups. All formulas were well tolerated and all groups showed a comparable safety profile based on the number and severity of adverse events and growth. In healthy infants, supplementation of infant-type bifidobacterial strain B. breve M-16V, at a dose close to bacterial numbers found in human milk, with scGOS/lcFOS (9:1) created a gut environment closer to the breastfed reference group. This specific synbiotic mixture may also support gut microbiota resilience during early life.Clinical Trial Registration This clinical study named Color Synbiotics Study, was registered in ClinicalTrials.gov on 18 March 2013. Registration number is NCT01813175. https://ichgcp.net/clinical-trials-registry/NCT01813175 .

Conflict of interest statement

This study was financially supported by Danone Nutricia Research. Shugui Wang, Alexia Kakourou, Guus Roeselers, and Jan Knol are employees of Danone Nutricia Research. Nopaorn Phavichitr, Sirinuch Chomto, Ruangvith Tantibhaedhyangkul, Sukkrawan Intarakhao, Sungkom Jongpiputvanich, Orapa Suteerojntrakool, Chonikarn Visuthranukul, Anundorn Wongteerasut and Punnapatch Piriyanon declare no competing interests.

Figures

Figure 1

Study design.

Figure 2

Study population composition. Screen Failure (SF) is defined as “potential subject did not meet one or more inclusion criteria. Intention to treat (ITT) includes every subject randomized to treatment assignment.

Figure 3

FISH analyses of fecal samples from the control treatment, Syn4 dose treatment and breastfed (BF) reference group. After 6 weeks the Syn4 dose treatment (trt) resulted in a significantly larger increase in the relative abundance (proportion) of bifidobacterial from baseline compared to control (A). (B) Shows adjusted LS mean (95% CL)/n change from baseline in relative abundance of bifidobacterial for the control and the Syn4 groups. Syn4 treatment lead to a significantly larger decrease in the relative abundance of Eubacterium rectale–Clostridium coccoides from baseline compared to control (C). A longitudinal linear mixed model was used with intervention, time, study site as fixed factors, intervention by time as interaction term and subject as a random effect. Data are expressed as mean ± SE. *Statistically significant difference in change from baseline between treatment groups (p-value < 0.05) as assessed by the linear mixed-effect model.

Figure 4

q-PCR analyses showed that the Syn4 dose resulted in a significantly larger increase in prevalence of (mean percentage of infants with detectable) B. breve (A) and B. breve M-16V (B) and a significantly larger decrease of C. difficile prevalence (C) as compared to control treatment. Detected C. difficile genomic copy numbers decreased significantly more in the Syn4 group compared to the control group (D). The increase in total amount of bifidobacterial copy numbers from baseline was only significantly larger (p = 0.0079) at Week 6 in the Syn4 group compared to control (E). Increase in the B. breve M-16V copy numbers from baseline was significantly larger in the Syn4 group compared to control at both Week 6 and Week 8 (respectively p < 0.0001 and p < 0.0001) (F). A generalized linear mixed model (GLMM) was used with intervention, time, study site as fixed factors, intervention by time as interaction term and subject as a random effect for the analysis of the binary transformed (detected/non-detected) data and the estimation of prevalence of detected values. A longitudinal linear mixed-effect model was used with intervention, time, study site as fixed factors, intervention by time as interaction term and subject as a random effect for the genomic copy number analysis. Data are expressed as mean ± SE. *Statistically significant differences in change from baseline between treatment groups (p-value < 0.05) as assessed by the linear or generalized linear mixed model.

Figure 5

The Syn4 dose resulted in a significantly larger decrease in pH (A) and a significantly larger increase in the l-lactate concentration (B) and the proportion of acetate (C) as compared to control. The proportion of propionate (D) and butyrate (E) in the Syn4 arm decreased significantly more from baseline as compared to the control arm. Data are expressed as mean ± SE. *Statistically significant differences in change from baseline between treatment groups (p-value < 0.05) as assessed by a longitudinal linear mixed-effect model with intervention, time, study site as fixed factors, intervention by time as interaction term and subject as a random effect.

Figure 6

RDA plot for gut microbiota analysis by treatment (control, syn4 and reference) at baseline (A), Week 6 (B) and Week 8 (C). A dot represents each sample and different colors represent different treatment (trt) groups. Triangles indicate centroids of study groups. Statistical significance of differences between groups based on the resulting model was evaluated using an ANOVA like permutation test.