Aims Considerable effort has been made to categorise the bacterial composition of the human gut and correlate findings with gastrointestinal disease. The infant gut has long been considered sterile at birth followed by rapid colonisation by pioneer microbiota. We examined first-pass meconium from healthy term infants to confirm/refute sterility.

Methods Healthy mothers were approached following vaginal delivery. First-pass meconium within 24 h of delivery were obtained from healthy, breastfed infants. Antibiotic use was an exclusion criterion- mother within 7 days or infant after birth. Stools were processed in triplicate for fluorescent in-situ hybridisation (FISH) with16S rRNA-targeted probes against all bacteria; Bifidobacterium; Bacteroides-Prevotella; Lactobacillaceae/Enterococcaceae; Enterobacteriaceae; Streptococcaceae; Staphylococcaceae and Enterococcaceae. Absolute counts of all bacteria and proportional identification for each bacterial group were calculated. DNA extraction followed by universal bacterial RT-PCR was performed on FISH-positive samples.

Results 15 babies met inclusion/exclusion criteria. All babies were 37–40 weeks gestation. 8/15 were male, mean birth weight was 3.36kg and mean maternal age was 31.9 years. 10/15 (66%) infants had evidence of bacteria on FISH. Of these, RT-PCR was positive in only 1. Positive FISH counts ranged from 2.2 to 41.8*10⁴ cells/g with the mean of positive samples being 15.4*10⁴ cells/g. (Limit of detection for automated counting is 10⁶ cells/g). Cell counts were too low to allow formal diversity analysis. Amplification by RT-PCR was not possible despite positive spiked samples demonstrating the feasibility of reaction. Three babies were dominated by a single family, either Enterobacteriaceae or Enterococcaceae. The others contained 2–5 genera. Bifidobacterium, Enterobacteriaceae and Bacteroides-Prevotella were the most dominant bacteria identified. There was no association between rupture of membrane duration, time to passage of meconium or time to lab with bacterial counts.

Conclusion Evidence of bacteria in first-pass meconium samples from healthy, vaginally-delivered, breastfed term infants is scant with only two-thirds having demonstrable bacteria at levels too low for automated counting. Bacterial RT-PCR failed to amplify 9/10 FISH-positive samples. This study suggests that gut bacterial diversity is extremely limited at birth and supports the hypothesis that the neonatal gut is sterile and colonised rapidly thereafter.