Increased antibiotic resistance in preterm neonates under early antibiotic use

Abstract

The standard use of antibiotics in newborns to empirically treat early-onset sepsis can adversely affect the neonatal gut microbiome, with potential long-term health impacts. Research into the escalating issue of antimicrobial resistance in preterm infants and antibiotic practices in neonatal intensive care units is limited. A deeper understanding of the effects of early antibiotic intervention on antibiotic resistance in preterm infants is crucial. This retrospective study employed metagenomic sequencing to evaluate antibiotic resistance genes (ARGs) in the meconium and subsequent stool samples of preterm infants enrolled in the Routine Early Antibiotic Use in Symptomatic Preterm Neonates study. Microbial metagenomics was conducted using a subset of fecal samples from 30 preterm infants for taxonomic profiling and ARG identification. All preterm infants exhibited ARGs, with 175 unique ARGs identified, predominantly associated with beta-lactam, tetracycline, and aminoglycoside resistance. Notably, 23% of ARGs was found in preterm infants without direct or intrapartum antibiotic exposure. Post-natal antibiotic exposure increases beta-lactam/tetracycline resistance while altering mechanisms that aid bacteria in withstanding antibiotic pressure. Microbial profiling revealed 774 bacterial species, with antibiotic-naive infants showing higher alpha diversity (P = 0.005) in their microbiota and resistome compared with treated infants, suggesting a more complex ecosystem. High ARG prevalence in preterm infants was observed irrespective of direct antibiotic exposure and intensifies with age. Prolonged membrane ruptures and maternal antibiotic use during gestation and delivery are linked to alterations in the preterm infant resistome and microbiome, which are pivotal in shaping the ARG profiles in the neonatal gut.This study is registered with ClinicalTrials.gov as NCT02784821.

Importance: A high burden of antibiotic resistance in preterm infants poses significant challenges to neonatal health. The presence of antibiotic resistance genes, along with alterations in signaling, energy production, and metabolic mechanisms, complicates treatment strategies for preterm infants, heightening the risk of ineffective therapy and exacerbating outcomes for these vulnerable neonates. Despite not receiving direct antibiotic treatment, preterm infants exhibit a concerning prevalence of antibiotic-resistant bacteria. This underscores the complex interplay of broader influences, including maternal antibiotic exposure during and beyond pregnancy and gestational complications like prolonged membrane ruptures. Urgent action, including cautious antibiotic practices and enhanced antenatal care, is imperative to protect neonatal health and counter the escalating threat of antimicrobial resistance in this vulnerable population.

Keywords: meconium; microbiome; preterm infants; resistome; stool.

Fig 1

Flowchart illustrating the analysis methods for the REASON longitudinal cohort and antibiotic resistance study, including the investigation of potential exposures linked to elevated detection of ARGs. The left side of the figure highlights the types of samples and metadata collected from the REASON study (n = 91 preterm neonates). In this study, we evaluated a subset of samples from 30 preterm neonates across two sampling time points, including meconium and stool samples collected before discharge for the presence of ARGs. Potential determinants driving ARG detection and neonatal health outcomes were analyzed. This figure was made using Canva and BioRender.

Fig 2

Sampling schematic and participant demographics. A total of 30 infants were included, and stool samples were collected at two time points: immediately after birth (base stool) and just before hospital discharge (end stool). Among the 30 neonates, 16 received antibiotic treatment at birth (yellow line), while 14 did not (gray line). Neonates who did not receive antibiotics at birth remained free from antibiotic exposure throughout the study. The gestational age of the infants ranged from 24 to 32 weeks, with a median age of 29 weeks and a standard deviation (SD) of 2.55. Additional information included mode of delivery, antibiotic administration to mothers (with 21 mothers receiving antibiotics during delivery), and the timing of stool sample collection (marked as a blue circle for the first stool after birth and a red square prior to discharge). PEI stands for preterm infant. Infant ID refers to a unique identifier assigned to each participant throughout this study.

Fig 3

ARG diversity and abundance in preterm infant meconium and stool. (A) The number of unique ARGs identified in each infant based on time (meconium and stool) and treatment using the Wilcoxon signed-rank test. Antibiotic (+) represents those who received treatment compared with PEIs, who did not receive antibiotic (−). (B) Alpha diversity for the PEI resistome is depicted based on treatment and time. (C) Relative abundance of ARGs based on treatment and sample timeline. The top 25 most abundant ARGs are reported. PEI stands for preterm infant.

Fig 4

Variance in preterm infant resistance abundance at the ARG level. (A) Comparison of ARGs over time using meconium and stool samples. Panels B and C display the differential abundance of ARGs based on neonatal antibiotic treatment in meconium (n = 26 samples/PEIs) and stool samples (n = 30 samples/PEIs), respectively. PEI stands for preterm infant.

Fig 5

Functional profiling of infant gut post-antibiotic exposure. Functional analysis using output from MEGAN. Differential abundance was performed using meconium (A–D) and stool samples (E–H) (n = 60 samples total).

Fig 6

Comparative chord diagrams of bacterial species and ARG abundance in neonatal meconium and stool samples. This figure presents two pairs of chord diagrams, each illustrating the interplay between the top 50 most prevalent ARGs categorized by drug class and the top 10 bacterial species associated with the ARG. Diagrams A and B represent preterm infants who did not receive direct exposure to antibiotics, whereas diagrams C and D correspond to the antibiotic-treated group. Specifically, diagrams A and C highlight the relationships within meconium samples, and diagrams B and D show the connections within stool samples collected just before hospital discharge.