Study shows probiotics help preterm babies fight antibiotic-resistant bacteria

By Vijay Kumar MalesuReviewed by Susha Cheriyedath, M.Sc.Aug 17 2025

A new study reveals that probiotics curb drug-resistant bacteria in premature babies, but antibiotics still drive resistance genes to spread, underscoring the urgent need for safer treatment strategies in NICUs.

Study: Impact of early life antibiotic and probiotic treatment on gut microbiome and resistome of very-low-birth-weight preterm infants. Image Credit: Ratchat / Shutterstock

In a recent study published in the journal Nature Communications, researchers determined how early antibiotic exposure and routine probiotic supplementation influence the gut microbiome, antibiotic resistance genes (ARGs), and multidrug-resistant (MDR) pathogen dynamics in very-low-birth-weight (VLBW) preterm infants.

Background

About one in ten babies is born early, and VLBW infants under 1500 g begin life fragile. In Neonatal Intensive Care Units (NICUs), broad-spectrum antibiotics can be lifesaving, yet they may disrupt the gut community that trains immunity and blocks harmful microbes. The World Health Organization (WHO) recommends probiotics containing specific bacterial strains for very preterm, exclusively human milk-fed infants, prompting a practical question: can routine probiotics offset antibiotic risks? Hospitals also face MDR infections that endanger fragile newborns. Understanding how probiotics and antibiotics shape the resistome (the collection of ARGs) can guide safer care, and hence, further research is needed.

About the Study

This controlled sub-study from the Baby-Associated Microbiota of the Intestine (BAMBI) observational cohort followed 34 VLBW preterm infants under 33 weeks’ gestation, all exclusively fed human milk or donor breastmilk, split into probiotic-supplemented (“PS”, receiving Bifidobacterium bifidum and Lactobacillus acidophilus [Infloran®]) and non-supplemented cohorts. Within each cohort, some infants received short empirical antibiotics (benzylpenicillin and/or gentamicin for a median of 3 days), while others did not. Fecal samples were collected weekly during weeks one to three. Shotgun metagenomic sequencing generated paired-end reads for assembly, and genome-resolved analyses reconstructed metagenome-assembled genomes alongside cultured isolates, enabling strain-level tracking.

Taxonomic and functional profiles were derived with established pipelines, and ARGs were identified against curated databases. Average nucleotide identity (ANI) and multi-locus sequence typing (MLST) established relatedness and sequence types. To probe horizontal gene transfer (HGT), the team quantified plasmid replicons, modeled candidate transfer events, and performed an ex vivo infant-gut experiment using preterm-derived Enterococcus faecium donor strains and a plasmid-free, gentamicin-sensitive laboratory recipient. Infant fecal slurry lacking Enterococcus served as matrix; selective plating on gentamicin and confirmatory whole-genome sequencing (WGS) verified acquired resistance. Ordination, group-difference testing, and correlation analyses compared cohorts and antibiotic exposure groups, emphasizing meaningful contrasts. Culture work targeted Bifidobacterium, Enterococcus, Staphylococcus, Klebsiella, and Escherichia species to complement sequencing.

Study Results

Across three weeks, the probiotic-supplemented cohort was dominated by Bifidobacterium, driven by the administered Bifidobacterium bifidum, with active replication (Index of Replication (iRep) > 1.5), while the non-supplemented cohort showed higher representation of early-life pathobionts such as Klebsiella, Escherichia, Enterococcus, and Staphylococcus. Beneficial infant-associated species, including Bifidobacterium breve and Bifidobacterium longum, appeared earlier and at higher relative abundance with probiotics, consistent with human milk oligosaccharide use and colonization resistance. Functional pathway profiles diverged between cohorts beginning in week two, tracking these taxonomic differences.

The resistome showed that compared with non-supplemented infants, probiotic-supplemented infants carried fewer ARGs and fewer resistance classes. Genes conferring resistance to aminoglycosides, macrolides-lincosamides-streptogramins, beta-lactams, trimethoprim, and tetracyclines were common, but fluoroquinolone and colistin resistance signatures were confined to non-supplemented infants. Notably, a colistin resistance gene (mcr-9.1) appeared in one non-supplemented sample, predating its formal discovery and underscoring hidden circulation of last-resort determinants. Short empirical antibiotics did not produce large alpha-diversity shifts, yet early tilts toward Klebsiella or Enterococcus were observed, especially in week one.

Correlation analyses linked higher Bifidobacterium abundance with fewer ARGs, whereas Enterococcus and Staphylococcus tracked with higher ARG loads. At the strain level, Enterococcus, Escherichia, Klebsiella, and Staphylococcus carried the densest resistance portfolios. None of the probiotic-cohort Klebsiella or Escherichia genomes met the definition of MDR (resistance to three or more classes), whereas 47.6% of non-supplemented Escherichia did. MLST highlighted clinically relevant sequence types (STs), including Escherichia coli ST1193 and Klebsiella pneumoniae ST432; Enterococcus faecalis and coagulase-negative staphylococci (e.g., S. epidermidis, S. haemolyticus) also included STs previously linked to neonatal units. Plasmid analyses showed higher plasmid burdens in non-supplemented infants and a weak association between plasmid counts and ARG counts.

Predicted HGT events were more frequent after antibiotic exposure, showing that even short courses can favor the movement of mobile elements. Related Enterococcus strains circulated among unrelated infants within the same hospitals, consistent with nosocomial transmission. The ex vivo infant-gut model provided mechanistic support: a 137-kb plasmid carrying an aminoglycoside resistance gene transferred from Enterococcus faecium donors to a plasmid-free recipient, conferring gentamicin resistance, confirmed via WGS. Probiotics tilted the gut toward beneficial microbes and away from ARG-rich pathobionts, though they did not eliminate HGT risk, while antibiotic stewardship remained critical to curb plasmid-mediated spread of resistance in the nursery.

Conclusions

To summarize, in VLBW preterm infants, probiotic supplementation fostered Bifidobacterium-rich communities, reduced ARG burden, and limited MDR features compared with no probiotics. Yet Enterococcus persisted as a key reservoir, and short antibiotic courses increased the likelihood of HGT, particularly plasmid-mediated movement of resistance genes. Practically, pairing evidence-based probiotics with antibiotic stewardship and infection control can help protect this population. Longer follow-up and multi-site trials are needed to refine dosing, duration, and strain choice to maximize benefits while minimizing resistance risks. For families and NICUs, aligning feeding, hygiene, and prescribing is essential.