The majority of preterm infants with very low birth weight who are admitted to neonatal intensive care units (NICUs) in the USA receive antimicrobial therapy, and any resulting antibiotic-induced perturbation of the gut microbiota was thought to resolve with time. Now, a study in Nature Microbiology reports metagenomic signatures of early-life antibiotic treatment and hospitalization in preterm infants that persist long-term.

Credit: Pulsar Imagens/Alamy

Whole-metagenome shotgun sequencing was performed on 437 stool samples, obtained from 58 infants over the first 21 months of life. The cohort included 17 antibiotic-naive, healthy infants born near term and 41 preterm infants who received a single course of antibiotics (n = 9) or prolonged antibiotics therapy (n = 32).

The species composition of the preterm infant gut microbiota was found to cluster by both gestational age at birth and treatment status. Early-life antibiotic treatment was associated with decreases in both species richness and species diversity. Antibiotics-treated infants exhibited younger predicted microbiota ‘ages’ than chronological age across several stages of development, which is indicative of a disruption in microbiota development. However, by 12–15 months of infant age, computed microbiota-for-age z-scores of antibiotics-treated preterm infants resembled those of antibiotic-naive, near-term infants, suggesting partial recovery of microbiota immaturity following hospital discharge.

The team determined the antibiotic resistome of 217 preterm and near-term infants by functional metagenomic analysis, sequencing resistance-conferring metagenomic inserts and assembling 874 unique antibiotic resistance genes (ARGs). Antibiotics-specific enrichment of ARGs was observed in preterm, treated infants, with Enterobacteriaceae and uncultured lineages identified as the likely source of ARGs. Multidrug-resistant Enterobacteriaceae lineages of on average >99.997% nucleotide identity were measured in cultured isolates of preterm infants’ stool samples taken 8–10 months apart, that is, before and after discharge from the NICU. This finding suggests that antibiotic treatment, hospitalization and/or premature birth have long-term effects on the gut microbiome.

Finally, the authors devised a machine learning model that was able to distinguish infants born preterm who received early-life antibiotics from unhospitalized, near-term infants based on the species and ARGs present in the gut microbiota following hospital discharge (or at concordant time points in untreated infants). The role this persistent metagenomic signature of early-life antibiotic treatment and hospitalization has in chronic diseases of preterm infants remains uncertain, but this study emphasizes the need for alternatives to broad-spectrum antibiotics in this vulnerable patient population.