Safety and efficacy of a probiotic-containing infant formula supplemented with 2'-fucosyllactose: a double-blind randomized controlled trial

Abstract

Background: Human milk oligosaccharides (HMOs) have important and diverse biological functions in early life. This study tested the safety and efficacy of a starter infant formula containing Limosilactobacillus (L.) reuteri DSM 17938 and supplemented with 2'-fucosyllactose (2'FL).

Methods: Healthy infants < 14 days old (n = 289) were randomly assigned to a bovine milk-based formula containing L. reuteri DSM 17938 at 1 × 107 CFU/g (control group; CG) or the same formula with added 1.0 g/L 2'FL (experimental group; EG) until 6 months of age. A non-randomized breastfed group served as reference (BF; n = 60). The primary endpoint was weight gain through 4 months of age in the formula-fed infants. Secondary endpoints included additional anthropometric measures, gastrointestinal tolerance, stooling characteristics, adverse events (AEs), fecal microbiota and metabolism, and gut immunity and health biomarkers in all feeding groups.

Results: Weight gain in EG was non-inferior to CG as shown by a mean difference [95% CI] of 0.26 [-1.26, 1.79] g/day with the lower bound of the 95% CI above the non-inferiority margin (-3 g/day). Anthropometric Z-scores, parent-reported stooling characteristics, gastrointestinal symptoms and associated behaviors, and AEs were comparable between formula groups. Redundancy analysis indicated that the microbiota composition in EG was different from CG at age 2 (p = 0.050) and 3 months (p = 0.052), approaching BF. Similarly, between sample phylogenetic distance (weighted UniFrac) for BF vs EG was smaller than for BF vs CG at 3-month age (p = 0.045). At age 1 month, Clostridioides difficile counts were significantly lower in EG than CG. Bifidobacterium relative abundance in EG tracked towards that in BF. Fecal biomarkers and metabolic profile were comparable between CG and EG.

Conclusion: L. reuteri-containing infant formula with 2'FL supports age-appropriate growth, is well-tolerated and may play a role in shifting the gut microbial pattern towards that of breastfed infants.

Trial registration: The trial was registered on ClinicalTrials.gov ( NCT03090360 ) on 24/03/2017.

Keywords: 2’fucosyllactose; Growth; Gut microbiome; Human milk oligosaccharides; Infant formula.

Conflict of interest statement

Mailis Amico, Bernard Berger, Norbert Sprenger, and Colin I. Cercamondi are current employees of Société des Produits Nestlé S.A. Guus A. M. Kortman is an employee of NIZO Food Research BV. The remaining authors declare that they have no competing interests that are relevant to the content of this article.

© 2022. The Author(s).

Figures

Fig. 1

Flow chart of study subject disposition. BF, breastfed group; CG, control group; EG, experimental group; ITT, intention-to-treat. *Non-completion includes infants that dropped-out of the study before the 6 months of age visit. Infants dropping out between the 4 and 6 months of age visits might still be part of the full analysis or per protocol set. The numbers of the per protocol sets are for the primary endpoint which was not assessed in BF

Fig. 2

Anthropometric z-scores in the intention-to-treat population for a weight-for-age, b length-for-age, c head-circumference-for-age, and d weight-for-length from baseline (≤ 14 days of age) to 6 months of age. Values are means with SD as whiskers. No statistical differences between the feeding groups were observed at any time point using a robust ANCOVA comparing the changes from baseline in the anthropometric z-scores between the feeding groups and adjusted for baseline value, mode of delivery, sex, study center, and visit

Fig. 3

Comparison of the gut microbiota composition between the three feeding groups at baseline, 1, 2 and 3 months of age. BF, breastfed group; CG, control group; EG, experimental group. Redundancy analysis on the OTU level, assessing the effect of feeding on gut microbiota composition. OTUs were used as response data and feeding was explanatory data, the bacterial families that contributed most were plotted supplementary. The covariance attributed to confounders delivery mode and “at least 1 episode of antibiotics treatment before the 3 months visit” (only at 3 months of age) was first fitted by regression and then “partialled out” (removed) from the ordination. The (unadjusted) variation explained is indicated on the axes. a Baseline (≤ 0.5 months of age); variation explained by feeding was 0.8%, p = 0.002. b 1 month of age; variation explained by feeding was 2.8%, p = 0.002. c 2 months of age; variation explained by feeding was 3.3%, p = 0.002. d 3 months of age; Variation explained by feeding was 2.9%, p = 0.002. Sample size at each timepoint was: 71–81 in CG, 86–93 in EG and 25–29 in BF

Fig. 4

Relative abundance of a Bifidobacterium in the three feeding groups, b Bifidobacterium in the formula groups by delivery mode, c Lactobacillus and d Peptostreptococcaceae at baseline (≤ 0.5 months of age), 1, 2 and 3 months of age in the three feeding groups. BF, breastfed group; CG, control group; EG, experimental group. Groups were compared per time point by Kruskal–Wallis test, followed by pairwise comparisons with Dunn’s posthoc test. * = p < 0.05, ** = p < 0.01, *** = p < 0.001. Boxplots are displayed as Tukey whiskers. Sample size at each timepoint was: 71–81 in CG, 86–93 in EG and 25–29 in BF

Fig. 5

Fecal microbiota qPCR results by feeding group for a Clostridioides difficile, b Clostridium perfringens, c Klebsiella pneumoniae at baseline (≤ 14 days of age), 1, 2, and 3 months of age. Values are means of 104 copies/mg fecal dry weight with SE as whiskers. Feeding groups were compared using log-transformed data in a Mixed Model Repeated Measures adjusted for baseline values, sex, mode of delivery, antibiotic use, study center, and visit. BF, breastfed group; CG, control group; EG, experimental group. # = p < 0.05 for EG compared to CG. * = p < 0.05 for CG compared to BF. ^ = p < 0.05 for EG compared to BF. Sample size at each timepoint was: 38–61 in CG, 47–70 in EG and 17–22 in BF for Clostridioides difficile; 64–74 in CG, 71–92 in EG and 23–26 in BF for Clostridium perfringens; 65–68 in CG, 72–84 in EG and 21–24 in BF for Klebsiella pneumoniae

Fig. 6

Comparison of the gut microbiota composition between the formula feeding groups at baseline, 1, 2 and 3 months of age stratified by delivery mode. CG, control group; EG, experimental group. Redundancy analysis on the OTU level at baseline, 1, 2 and 3 months of age stratified by delivery mode. OTUs were used as response data and feeding*delivery mode was explanatory data, the bacterial families that contributed most were plotted supplementary. The covariance attributed to “at least 1 episode of antibiotics treatment before the 3 months visit” (only at 3 months of age) was first fitted by regression and then “partialled out” (removed) from the ordination. The (unadjusted) variation explained is indicated on the axes. a Baseline (≤ 0.5 months of age). Variation explained by delivery mode was 2.7%, p = 0.002. b 1 month of age. Variation explained by feeding*delivery mode was 1.4%, p = 0.002. c 2 months of age. Variation explained by feeding*delivery mode was 1.2%, p = 0.002. d 3 months of age. Variation explained by feeding*delivery mode was 1.2%, p = 0.002. Sample size at each timepoint was: 71–81 in CG and 86–93 in EG