Table 2.

Summary of the 6 included probiotic studies

Reference, study design, study location, overall biasa	Study enrollment and intervention duration	Interventions used in study groupsb	Fecal sample collection, tests, compositionc	Fecal microbial endpoints	Health outcomes
Garcia-Rodenas et al. 2016 Double-blind, multicenter Greece MODERATEa	Infants enrolled by day 3 of life, on study formula for 6 mos (complementary foods allowed after 4 mos)	1) Ct (CTRL): starter formula, n=44 (recruited) and n=31 (AN) 2) Lr (PRO): CTRL + L. reuteri 1.2×109 CFU/l, n=44 (recruited) and n=31 (AN) Total: N=88 (recruited) and N=62 (AN)	Randomized groups further stratified based on vaginal (V) or C-section (C) delivery n=40 samples collected at 2 weeks and 4 mos: CLr, n=9; VLr, n= 11; CCt, n=10; VCt, n=10 Composition: 454 16S rRNA pyrosequencing	Enterobacteriaceae (relative abundance, %) reduced in CLr v. CCt at 2 weeks (P=0.004) but not at 4 mos Actinobacteria (relative abundance, %) increased in CLr v. CCt at 2 weeks (P=0.015) but not at 4 mos, specifically due to Bifidobacterium Lactobacillus (relative abundance, %) higher in CLr v. CCt at 2 weeks (P=0.027) and 4 mos (P=0.051) Lactobacillus (relative abundance, %) higher in VLr v. VCt infants at 2 weeks (P=0.045) and 4 mos (P=0.012) Microbial richness and diversity as measured by Chao1 and Shannon indices did not differ among formula groups.	Not assessed
Gil-Campos et al. 2012 Double-blind, multicenter Spain MODERATEa	Infants enrolled at 1 mo of age, on study formula for 5 mos (complementary foods introduced according to ESPGHAN guidelines)d	1) CTRL: standard formula + GOS (0.3g/100mL), n=71 (ITT) and n=63 (PP) 2) PRO: CTRL formula + L. fermentum CECT-5716 1×107 CFU/g, n=66 (ITT) and n=63 (PP) Total: N=137 infants (ITT)	n = unknown number of samples collected at 1 (baseline), 4, and 6 mos of age Fecal SCFA: gas chromatography Fecal IgA concentration: ELISA Composition: colony-plating and nested qPCR for Lactobacillus fermentum	Lactobacillus, Bifidobacterium, Clostridium, Bacteroidaceae (CFU/g) did not differ between groups at each time point, but both groups showed significant increases in these bacterial groups over time (statistic not provided) L. fermentum CECT5716 detected alive in 53% of PRO group and only 2 CTRL infants (statistic not provided)	Diarrhea incidence rate lower in PRO v. CTRL (P=0.018)
Hascoët et al. 2011 Double blind, single center France MODERATEa	Infants enrolled at or before day 7 of life, on study formula for 4 mos	1) CTRL: standard formula, n=38 (ITT) and n=33 (PP) 2) STUDY: formula with low protein and phosphate, high lactose, predominantly whey protein, n=39 (ITT) and n=32 (PP) 3) PRO: STUDY formula + B. longum BL999 2×107 CFU/g, n=40 (ITT) and n=32 (PP) 4) BF reference, n=73 (ITT) and n=44 (PP) Total: N = 190 (ITT) and N = 140 (PP)	n=140 samples collected at 1 and 2 mos Quantification of Bifidobacterium longum BL999 via plating followed by qPCR Fecal IgA concentration: ELISA Composition: FISH using 16S-rRNA targeted oligonucleotide probes	B. longum BL999 detected in PRO infants at 1 mo, but not at 2 mos	Soft stool frequency higher in PRO v. STUDY (P<0.05)
Maldonado et al. 2019 Double blind, multicenter Spain MODERATEa	Infants enrolled at 1 month of age, on study formula until 12 mos of age (complementary foods introduced according to ESPGHAN guidelines)d	1) CTRL: standard formula, n=77 (ITT) and n=61 (PP) 2) Lf (PRO): CTRL formula + L. fermentum CECT5716 Lc40, 107 CFU/g, n=83 (ITT) and n=65 (PP) 3) Bb (PRO): CTRL formula + B. breve CECT7263, 107 CFU/g, n=76 (ITT) and n=63 (PP) Total: N=236 (ITT) and N=189 (PP)	n=236 samples collected at baseline, 4, 6, 9, and 12 mos Composition: qPCR specifically targeting, Lactobacillus spp., Bifidobacterium spp., Clostridium spp., Bacteroides spp., and Escherichia coli	Lactobacillus higher in Bb v. CTRL at 4 mos (P<0.001) Bifidobacterium lower in Lf v. CTRL at 4 mos (P=0.038)	Lower risk of long episodes of crying in Bb v. CTRL (P=0.001)
Papagaroufalis et al. 2014 Double blind, multicenter Greece HIGHa	Infants enrolled at 0–72 hours of life, on study formula until 28 days of age	1) CTRL: starter infant formula, n=44 (ITT), n=35 (PP day 28), and n=31 (PP days 112, 168) 2) PRO: CTRL + L. reuteri DSM-17938 1.2×106 CFU, n=44 (ITT), n=36 (PP day 28), and n=31 (PP days 112, 168) Total: N=88 (ITT), N=71 (PP day 28), and N=62 (PP days 112, 168)	n=71 (day 28), 62 (day 112) samples collected at 14 and 112 day visits Composition: PCR for quantification of L. reuteri and FISH, specifically bifidobacteria, lactobacilli, Enterobacteriaceae, Clostridium difficile	Bifidobacterium, Lactobacillus, and L. reuteri detectability (presence/absence) higher in PRO v. CTRL at day 14 (P=0.005–0.032) and day 112 (P=0.006–0.024)	Lower number of spitting events (median) in PRO v. CTRL at day 28 (P=0.048) and 4 mos (P=0.047). Lower frequency of hard stools (P=0.001) and higher percentage of soft stools (P=0.018) in PRO v. CTRL at day 28
Wu et al. 2016 Double blind, single center China HIGHa	Infants enrolled at or before day 7 of life, on study formula until 6 mos of age (no information given on complementary foods)	1) CTRL: commercially available standard formula, n=148 (EN) and n=129 (AN) 2) PRO: CTRL formula + B. longum BB536 1×107 CFU/g, n=152 (EN) and n=135 (AN) Total: N=300 (EN) and N=264 (AN)	n=264 samples collected at 2, 4, and 11 mos of age Composition: Plating techniques for total bacteria count, lactobacilli count, and Enterobacteriaceae count, PCR for bifidobacteria	Bifidobacteria (log CFU/g) higher in PRO v. CTRL at 2 mos (P<0.0001) and 4 mos (P=0.0096) Bifidobacteria/Enterobacteriaceae ratio higher in PRO v. CTRL at 2 mos (P<0.0001) and 4 mos (P=0.03)	Not assessed

^aOVERALL BIAS: The overall bias rating based on ratings of five individual-domain-level bias ratings (see Figure 3 and Supplemental Table S1).

^bCTRL, control group; PRO, probiotic group; BF, breast-fed group. Other groups are specified by study. Interventions: GOS, galactooligosaccharide; CFU, colony-forming units. We also specify intent-to-treat (ITT), enrolled (EN), analyzed (AN), and per-protocol (PP) analyses if given in the article.

^cSCFA, short-chain fatty acid; FISH, fluorescence in situ hybridization; qPCR, quantitative polymerase chain reaction; ELISA, enzyme-linked immunosorbent assay.

^dESPGHAN guidelines: an authoritative guidance for the incorporation of complementary foods into the infant diet.