Objective To assess the effect on microbiotia development of a standard control formula, a formula with a composition close to human milk, Nan, and investigate the equivalence between Nan and Nan + a probiotic, Bifidobacterium longum (BL999; Morinaga, Japan).

Methods Randomized, double blind, controlled trial of 3 groups: control formula; Nan : 30/70 Casein/Whey ratio (vs.70/30), lower protein (1.8 vs.2.5 g/100 Kcal); lower phosphate (31 vs.70 mg/100 Kcal) and higher lactose levels (11.1 vs. 4.5 g/100 Kcal); Nan + BL999 (2×107 cfu/g) and a reference BF group. Healthy singleton term infants received the study formulae for 4 months. Primary outcome was stool bifidobacteria counts (log CFU/g: Median (IQR)) measured by FISH method at 2 months of age.

Results Bifidobacteria counts were significantly higher for BF (10(0.9), p<0.0001, n = 35; 49.7% of total bacteria (TB)), or Nan (10.0(0.8), p<0.0001 n = 33; 52.6% TB) or Nan +BL999 (9.8(1.4), p<0.01 n = 33; 35.3% TB) groups than controls (9.2(3.5), n = 33; 31.9%TB). The same profile between study formulae was observed with Enterobacteria (8.5, 8.4, 8.1% vs. 13.1% TB respectively) or Clostridium (5.6, 9.4, 7.5 vs. 18.9% TB respectively).

Conclusions These data show that infant formula with a composition closer to human milk (with or without BL999) is more bifidogenic, and leading to a closer intestinal microbiotia profile to BF, than a standard formula.