Table 3.

Summary of the 6 included synbiotic 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
Abrahamse-Berkeveld et al. 2016 Double blind, multicenter Germany MODERATEa	Infants enrolled by day 35 of life, on study formula for 13 weeks	1) CTRL: standardized extensively hydrolyzed whey protein-based powder without synbiotics, n=111 (ITT) and n=57 (PP) 2) SYN: CTRL formula + 0.8 g/100 mL GOS/FOS (9:1 ratio) + 1.3×109 CFU/100 mL B. breve M-16V), n=100 (ITT) and n=45 (PP) Total: N=211 infants (ITT) and N=102 infants (PP)	n=60 (n=36 CTRL, n=24 SYN) samples collected at baseline, week 1, week 13 (subset) Fecal SCFA: gas chromatography Fecal pH: pH meter Composition: FISH using 16S rRNA-targeted oligonucleotide probes, specifically targeting bifidobacteria, lactobacilli, Bacteroides/Prevotella, Clostridium histolyticum/C. lituseburense, Enterobacteriaceae, C. coccoides/Eubacterium rectale	Bifidobacteria (%) higher in SYN v. CTRL at week 13 (P=0.014 (ITT)) C. coccoides/E. rectale cluster (%) lower in SYN v. CTRL at week 13 (P=0.013 (PP)) Potential pathogens (%, C. histolyticum/C. lituseburense ratio) lower in SYN v. CTRL at week 1 (P=0.003) and week 13 (P=0.013) in both ITT (P=0.043) and PP (P=0.058) Fecal pH lower and D-lactate concentration higher in SYN v. CTRL at week 13 (P=0.04)	Stool consistency score lower in SYN v. CTRL in ITT subpopulation in first 4 weeks (P=0.035), but not at 13 weeks Diaper (nappy) rash severity lower in SYN v. CTRL in ITT subpopulation in first 4 weeks (P=0.026), but not at 13 weeks
Cooper et al. 2016 Double blind, multicenter South Africa MODERATEa	Infants enrolled by day 3 of life, on study formula until 6 mos of age (complementary food allowed after 4 mos)	1) CTRL: standard formula, n= 214 (RAND) and n= 129 (completed 4 mos) 2) SYN: CTRL formula + BMOS (whey permeate containing GOS and 3’- and 6’-sialyllactose) (8 g/L in reconstituted formula) and a probiotic (B. lactis CNCM-I-3446 at 1×107 CFU/g of powder formula), n=207 (RAND) and n=138 (completed 4 mos) Total: N=421 (RAND) and N=267 (completed 4 mos)	n=168 samples collected at day 10; unknown fecal sample n collected at day 3, 4 weeks, and 3 mos Fecal IgA: ELISA Fecal pH: pH meter Composition: Plating methods, PCR used to determine B. lactis CNCM I-3446, Staphylococcus, enterobacteria, Escherichia coli, and Klebsiella counts FISH used to determine total bacterial counts, bifidobacteria, lactobacilli, Bacteroides, and Clostridium	Bifidobacteria (CFU) higher in SYN v. CTRL at day 28 (P=0.001) and 3 mos (P<0.001) in C-section-born infants Bifidobacteria (CFU) greater increase in SYN v. CTRL in vaginally born infants at day 28 (P<0.001) and day 84 (P<0.001) Bifidobacteria species detected (presence/absence) in higher proportion in SYN v. CTRL at days 10 and 28 only in C-section born infants (P<0.025) B. lactis detected (presence/absence) in higher proportion of infants in SYN v. CTRL at day 10, week 4, and 3 mos (P<0.025) Lactobacillus detected (presence/absence) in higher proportion in SYN v. CTRL at 3 mos for vaginally born infants, and day 28 only for C-section-born infants (no p-value given) Clostridium eubacteria detected (presence/absence) in lower proportion in SYN v. CTRL at 3 mos (P<0.025) Enterobacteriaceae detected (presence/absence) in lower proportion in SYN v. CTRL at days 10 and 28 in vaginally born infants (P<0.025) E. coli detected (presence/absence) in lower proportion in SYN v. CTRL at days 3 and 28 in vaginally born infants (P<0.025) Klebsiella spp. detected (presence/absence) in lower proportion of infants in SYN v. CTRL at days 3 and 28 in vaginally born infants (P<0.025) Staphylococcus lower in SYN v. CTRL in vaginally born infants at days 10 and 28 only (P<0.025) Fecal pH lower in SYN v. CTRL among both vaginally and C-section born infants at day 10 and week 4, but significance only remained at month 3 for the C-section-born infants	Harder stool consistency lower in proportion of infants in SYN v. CTRL among vaginally born (P=0.002) and C-section-born infants (P=0.001) up until 6 mos of age Formed stools greater in proportion of CTRL v. SYN infants among C-section-born (P=0.045) and vaginally born (P=0.055) infants Liquid stools (frequency) higher in SYN among C-section-born infants (P<0.001)
Meli et al. 2014 Double-blind, single center Italy MODERATEa	Infants enrolled by 14 days of life, on study formula until 4 mos of age	1) CTRL: standard infant formula, n= 84 (RAND), n= 63 (primary analysis), and n=57 (PP) 2) PRE: CTRL formula + 10 g/L BMOS (whey permeate containing GOS and 3’- and 6’-sialyllactose) in the reconstituted formula, n=99 (RAND), n=62 (primary analysis), and n=60 (PP) 3) SYN: CTRL formula + 10 g/L BMOS (whey permeate containing GOS and 3’- and 6’-sialyllactose) + probiotics B. longum (B1999) and L. rhamnosus (LPR), each at 2×107 CFU/g, n=98 (RAND), n=64 (primary analysis), and n=56 (PP) 4) BF reference group, n=30 (RAND), n=12 (primary analysis), and n=12 (PP) Total: N=281 (RAND), N=201 (primary analysis), and N=185 infants (PP)	n=71 samples collected in a subset of infants at 2 mos of age Composition: FISH analyzing the following bacterial species: bifidobacteria, lactobacilli, enterobacteria, clostridia, and Bacteroides; Bl999 and LPR quantified via plating technique	Fecal Bifidobacterium and Lactobacillus counts (median) higher in SYN v. CTRL at 2 mos (P<0.05) Clostridium counts (median) lower in PRE and SYN v. CTRL at 2 mos (P<0.05)	Daily stool frequency higher in PRE and SYN v. CTRL (P=0.0001) Lower odds of harder stools in PRE and SYN v. CTRL (P=0.0001, P=0.0003, respectively) Investigator-diagnosed colic lower in CTRL v. PRE (P=0.01)
Radke et al. 2017 Double blind, multicenter France, Germany, Netherlands MODERATEa	Infants enrolled at or before day 14 of life, on study formula from enrollment to 6 mos of age (complementary foods allowed at 4 mos)	1) CTRL: standard formula, n=180 (ITT), n=157 (PP) 2) SYN: CTRL formula + BMOS (whey permeate containing GOS and 3’- and 6’-sialyllactose) (8 g/L) + B. lactis (CNCM I-3446, 1×107 CFU/g), n=179 (ITT) and n=150 (PP) 3) BF reference, n=59 (ITT) and n=49 (PP) Total: N=359 (ITT) and N=307 (PP)	n=unknown samples collected in a subset of infants at 3 and 6 mos of age Fecal sIgA and alpha-1 antitrypsin: ELISA Fecal pH: pH meter Composition: FISH using 16S-rRNA targeted oligonucleotide probes (details not provided)	Bifidobacterium and lactobacilli counts higher in SYN v. CTRL at 3 mos (P<0.01) Clostridia/eubacteria counts lower in SYN v. CTRL at 3 mos (P < 0.01) B. lactis detection (presence/absence) higher in SYN vs CTRL at 3 (P<0.001) Fecal pH (mean) lower in SYN v. CTRL at 3 mos (P<0.001, ITT) Fecal sIgA (mean) concentrations (mg/L) higher in SYN v. CTRL at 3 mos (P<0.0001, P<0.0001, respectively) (ITT) Stool alpha-1 antitrypsin higher in SYN v. CTRL at 3 mos (P=0.03) (ITT)	Daily stool frequency higher in SYN v. CTRL in first 3 mos (P<0.005) Infants who ever had flatulence (proportion) higher in SYN v. CTRL at 3 mos (P<0.01)
Rozé et al. 2012 Double blind, multicenter France MODERATEa	Infants enrolled by day 3 of life, on study formula until 6 mos of age (no information given on complementary foods)	1) CTRL: standard formula, n=49 (RAND) and n=38 (PP) 2) SYN: CTRL formula + L. rhamnosus LCS-742 (1.4×108 CFU) and B. infantis M63 (1.4×108 CFU) + 96% GOS (0.4 g/100mL) and 4% scFOS (0.02 g/100mL); also enriched with alpha-lactalbumin, n=48 (RAND) and n=35 (PP) Total: N=97 (RAND) and N=73 (PP)	n=43, 34 samples collected at 1 and 6 mos, respectively Fecal sIgA: ELISA Composition: plate spreading for quantification of main genera, PCR for specific genus and species as well as 16S rDNA sequencing	Bifidobacteria presence did not differ between SYN v. CTRL at 1 and 6 months (P=0.14 and P=0.99, respectively) and colonization (CFU/g) did not differ at 6 months (P=0.07) Lactobacilli colonization (CFU/g) higher in SYN v. CTRL at 1 month (P<0.0001), but not at 6 mos Clostridium presence did not differ between SYN v. CTRL at 1 and 6 months (P=0.97 and P=0.29, respectively) Incidence (presence/absence) rate and colonization of staphylococci (CFU/g) were higher in SYN v. CTRL at 1 and 6 mos (P=0.02, P=0.02) Fecal sIgA concentrations were similar at 1 and 6 mos in SYN, but decreased from 1 to 6 mos in CTRL (no statistics given)	During the 3 days preceding the 1-mo clinical visit, SYN exhibited less crying or agitation, more quiet behavior v. CTRL (P=0.03 for ITT, P<0.02 for PP) SYN associated with reduced risk of atopic dermatitis during the study (P<0.05) Number of stools per day greater in SYN v. CTRL at 1 month (P=0.05)
Simeoni et al. 2016 Double blind, multicenter France, Poland MODERATEa	Infants enrolled at or before 14 days of life, on study formula until 3 mos of age	1) CTRL: standard, starter infant formula, n=37 (EN) and n=18 (PP) 2) SYN: CTRL formula + BMOS (whey permeate containing GOS and 3’- and 6’-sialyllactose) (8 g/L + B. lactis CNCM I-3446 (1×107 CFU/g), n=39 (EN) and n=21 (PP) 3) BF reference, n=39 (EN) and n=23 (PP) Total: N=115 (EN) and N=62 infants (PP)	n=62 samples collected at baseline, 6, and 12 weeks of age 16S rRNA gene sequencing Composition: qPCR used for total bacterial cell counts Lactobacillus and Bifidobacterium, as well as counts for individual species of bifidobacteria	Bifidobacterium (P<0.001), total bacteria (P<0.05), and B. longum (P<0.05) counts higher in SYN v. CTRL at 6 and 12 weeks Escherichia (count) decreased over time in all groups but less so in CTRL v. SYN (P<0.01). B. animalis present in most fecal samples in SYN, but nearly absent in CTRL (P<0.001) Diversity index (P<0.01) and fecal pH (P=0.016) higher in CTRL v. SYN at 6 weeks but not at 12 weeks	Liquid stool frequency significantly higher in SYN v. CTRL (no p-value reported)

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

^bBF, breastfeeding reference group; CTRL, control group; PRE, prebiotic group; SYN, synbiotic group. Other groups are specified by study. Interventions: BMOS, bovine milk oligosaccharide; CFU, colony-forming units; FOS, fructooligosaccharide; GOS, galactooligosaccharide; scFOS, short-chain fructooligosaccharide;. We also specify enrolled (EN), intent-to-treat (ITT), per-protocol (PP), randomized (RAND) analyses if given in the article.

^cELISA, enzyme-linked immunosorbent assay; FISH, fluorescence in situ hybridization; PCR, polymerase chain reaction; SCFA, short chain fatty acid; sIgA, secretory immunoglobulin A.