The urinary excretion of soya isoflavones and gut microflora metabolites was investigated in infants and children who had been fed soya-based infant formulas in early infancy. These infants and children were compared with cows'-milk formula-fed controls, to determine at what age gut microflora metabolism of daidzein to equol and/or O-desmethylangolensin (O-DMA) was established, and whether exposure to isoflavones in early infancy influences their metabolism at a later stage of development. Sixty infants and children (aged 4 months–7 years) participated in the study; thirty in each of the soya and control groups. There were four age groups. These were: 4–6 months (seven in the soya group and seven in the control group); 7–12 months (seven in the soya group and nine in the control group); 1–3 years (six in the soya group and eight in the control group); 3–7 years (ten in the soya group and six in the control group). Urine samples were collected to measure isoflavonoids by MS, and faecal samples were collected to measure gut-health-related bacterial composition, by fluorescent in situ hybridisation with oligonucleotide probes, and metabolic activity. A soya challenge (typically a soya yoghurt alternative product containing 4·8g soya protein and on average 22mg total isoflavones) was given to control-group infants (>6 months) and children, and also to soya-group children that were no longer consuming soya, to determine their ability to produce equol and/or O-DMA. Urinary genistein, daidzein and glycitein were detected in all infants (4–6 months) fed soya-based infant formula; O-DMA was detected in 75% of infants but equol was detected in only 25%. In the controls (4–6 months), urinary isoflavonoids were very low or not detected. In the older age groups (7 months–7 years), O-DMA was found in the urine samples of 75% of the soya group and 50% of the controls, after the soya challenge. Equol excretion was detected in 19% of the soya-group infants and children, and in only 5% of the controls. However, in the oldest (3–7 years) children, the proportion excreting O-DMA and equol was similar in both groups. Faecal bacterial numbers for bifidobacteria (P<0·001), bacteroides and clostridia (P<0·05) were significantly lower for the soya group compared with the control group. There appears to be no lasting effect of early-life isoflavone exposure on isoflavone metabolism.