Multidrug-Resistant Gut Colonization in Preterm Neonates

Evidence

Two studies published in 2026 examine multidrug-resistant bacterial colonization in hospitalized neonates from different angles: one follows mother-infant pairs through discharge, while the other characterizes bacteria recovered from preterm infants’ stool. Together, they describe colonization and acquisition, not proof that a colonizing organism caused infection. That distinction is central because bacteria can be present in the gut without invading tissue or producing culture-confirmed disease.

The newer NeoCol study was a prospective cohort of 189 mother-infant dyads. At delivery, 120 of 189 mothers (63%) were colonized with multidrug-resistant bacteria. By hospital discharge, 115 of 189 neonates (61%) were colonized. The pattern differed for carbapenem-resistant Enterobacterales, or CRE: maternal CRE colonization was detected in 6 of 189 dyads (3%), whereas 50 of 189 hospitalized neonates (26%) acquired CRE at a median of 13 days. The authors interpreted this contrast as evidence that horizontal acquisition in the hospital predominated over maternal transmission for CRE.

NeoCol also found several associations with neonatal MDR colonization. Before colonization, antibiotic exposure was recorded in 63% of colonized neonates versus 24% of neonates who were not colonized. Median hospital stay was 16 days in the colonized group and 2 days in the non-colonized group. Episodes of sepsis after colonization occurred in 62% versus 24%, respectively. These comparisons do not establish that antibiotics or colonization caused the later outcomes: infants requiring longer, more complex hospital care may have had more opportunities for both exposure and testing.

A result within NeoCol directly limits a simple colonization-to-infection story. Eleven neonates had culture-positive infections, but in 10 of those 11 cases (91%), the infecting bacterium was not one that had previously been identified as a colonizer. Colonization was therefore associated with later sepsis episodes at the group level, yet the cultured infection usually did not match the earlier colonizing organism.

The second study sampled preterm neonates in three neonatal intensive care units in Madinah, Saudi Arabia. Gestational ages ranged from 25 to 33 weeks. Researchers cultured stool on eosin methylene blue agar, identified isolates biochemically and with 16S ribosomal RNA sequencing, and measured antimicrobial susceptibility with the VITEK 2 system. Among 16 colonized infants aged at least two weeks, Klebsiella pneumoniae appeared in 9 (56.2%), Klebsiella quasipneumoniae in 1 (6.3%), and Escherichia coli in 8 (50%); some infants carried more than one genus.

Two K. pneumoniae isolates were extensively multidrug resistant. One was resistant to the carbapenems tested, and the other had intermediate susceptibility. Four K. pneumoniae and four E. coli isolates produced extended-spectrum beta-lactamases, while none of the E. coli isolates was carbapenem resistant. The study documented resistant organisms in fecal samples; it did not demonstrate mucosal colonization, transmission between infants, or an infection caused by those isolates.

Analysis — Hospital exposure shapes colonization

The cross-study pattern suggests that neonatal MDR colonization is better understood as a marker of accumulated exposure than as a direct diagnosis of infection. This is an analysis, not a causal conclusion. NeoCol supplies the stronger temporal evidence: CRE was uncommon in mothers, appeared later in hospitalized infants, and clustered with prior antibiotic exposure and longer stays. The smaller Madinah study adds organism-level detail, showing that K. pneumoniae—including resistant isolates—can become prominent in stool from antibiotic-exposed preterm infants after the first week of life. The convergence supports a hospital-acquisition hypothesis, but neither study identifies a specific transmission route. NeoCol also provides an important counterweight: most culture-confirmed infections did not match previously detected colonizers. Taken together, the studies indicate that colonization surveillance may reveal the ecological pressure under which resistance accumulates, while remaining an unreliable stand-alone explanation for which organism causes a later infection.

Limitations

NeoCol is available in the source pack only as a publisher abstract, so its detailed protocol, site structure, adjustment strategy, microbiological matching rules, and full outcome definitions could not be independently examined here. Its reported comparisons are associations and may be influenced by illness severity, duration of hospitalization, procedures, or frequency of sampling. The abstract also reports sepsis episodes separately from culture-positive infections, making it important not to treat every sepsis episode as microbiologically linked to prior colonization.

The Madinah study provides commercially reusable full text but is small. Sampling covered three hospitals during a short 2021 period, follow-up was feasible only for some infants, and clinical records were incomplete. Stool detection may not represent stable intestinal-mucosal colonization. The work used culture, 16S sequencing, and phenotypic susceptibility testing without whole-genome sequencing or resistance-gene analysis, so it could not establish relatedness, transmission chains, or the genetic basis of resistance. All sampled infants had suspected sepsis and received or had recently received antibiotics, limiting comparison with unexposed preterm infants. Neither study tested a clinical intervention or establishes that changing any single exposure would reduce infection or mortality.