Neonatal antibiotic exposure impairs child growth during the first six years of life by perturbing intestinal microbial colonization

Newborns frequently receive empirical antibiotics due to suspected early-onset sepsis, although confirmed infection is relatively uncommon. Early-life antibiotics are known to acutely alter neonatal gut microbiome composition, but long-term clinical consequences are unclear. Prior epidemiological and experimental evidence has linked the intestinal microbiome to growth, obesity, and metabolic disease, and some studies have suggested early antibiotic exposure relates to altered growth trajectories. The authors hypothesized that antibiotic treatment during the first days of life disrupts normal gut colonization, leading to long-lasting effects on childhood growth, and that the timing of antibiotic exposure (neonatal versus later in childhood) differentially associates with growth outcomes.

Previous work indicates: (1) neonatal and infant antibiotic exposure perturbs gut microbiome composition and reduces diversity; (2) antibiotic use later in infancy/childhood has been linked to higher risk of overweight and obesity; (3) the gut microbiome influences energy harvest and metabolism, and reduced Bifidobacterium in infancy has been associated with later overweight/obesity; (4) some studies reported decreased growth in the first year of life following antibiotics in the first week, whereas others linked post-neonatal antibiotics to excessive weight gain, with possible sex-specific effects. These data informed the study’s hypothesis on timing-dependent effects and a microbiome-mediated mechanism.

Design and cohorts: The primary analysis used the Southwestern Finland Birth Cohort (SFBC), including 12,422 term singletons born 2008–2010 at Turku University Hospital, excluding genetic abnormalities, significant chronic disorders, long-term prophylactic antibiotics, and multiple births. Neonatal antibiotic exposure was defined as antibiotics administered within the first 14 days of life (predominantly intravenous benzylpenicillin and gentamicin). Exposure was categorized as brief empirical therapy (discontinued when infection ruled out) or treatment for confirmed/clinical infection. Growth outcomes (population-specific Z-scores for weight, height, BMI) were obtained from municipal well-baby clinics. Potential confounders included gestational age, birth weight Z-score, mode of delivery, maternal pre-pregnancy BMI, intrapartum antibiotic treatment, breastfeeding (where available), and child’s age at measurement. Boys and girls were analyzed separately due to sex differences in exposure frequency and suspected dimorphism. Replication cohort: The German PEACHES cohort (Programming of Enhanced Adiposity Risk in Childhood-Early Screening) included 535 singleton children with data through 5 years; 6.4% received neonatal antibiotics. Analyses adjusted for gestational age, birth weight Z-score, mode of delivery, maternal pre-pregnancy BMI, intrapartum antibiotics, breastfeeding during the first 6 months, and child’s age. Early-childhood antibiotic exposure: In SFBC, antibiotic purchases during the first six years were extracted from the national Drug Prescription Register. Cumulative antibiotic prescriptions were categorized (quartiles) at each time point to assess associations with BMI Z-scores. Microbiome sub-study: In an independent clinical trial cohort at Turku University Hospital, fecal samples were collected at 1, 6, 12, and 24 months from full-term infants exposed (n≈13–20 across timepoints) or not exposed (controls) to neonatal antibiotics (benzylpenicillin + gentamicin). 16S rRNA gene sequencing (QIIME2; DADA2 denoising; taxonomy assignment using a pretrained classifier) and subset metagenomics were performed. Diversity metrics included alpha diversity (Phylogenetic Diversity, PD whole tree) and beta diversity (weighted and unweighted UniFrac), analyzed with Kruskal–Wallis and PERMANOVA, respectively. Dimensionality reduction via PCoA and Hotelling tests assessed persistent exposure effects. Genome-resolved metagenomics on 27 samples identified Bifidobacterium species presence/abundance. Germ-free mouse FMT: Germ-free Swiss Webster mice received fecal microbiota transplants (FMT) from 1-month-old infants exposed or unexposed to neonatal antibiotics. Mice were monitored for up to 43 days with periodic weight measurements and fecal sampling for microbiome analyses. Weight changes were normalized to day 0; group differences assessed with appropriate statistical tests (e.g., unpaired two-tailed tests; FDR where applicable).

- In SFBC, after adjustment, boys exposed to neonatal antibiotics (brief empirical or for infection) had significantly lower weight Z-scores across 0.5–6 years compared to unexposed; p = 0.007 (infected) and p = 0.031 (non-infected). Boys treated for infection also had significantly lower height Z-scores (p = 0.011). A detailed model (0.5–5 years) showed β (95% CI) and p-values for boys: weight Z-score −0.43 (−0.86, −0.002), p = 0.0488; height Z-score −0.62 (−1.08, −0.16), p = 0.0080; BMI Z-score −0.11 (−0.68, 0.47), p = 0.743. No significant associations were observed in girls for weight, height, or BMI after adjustment (e.g., weight β 0.28 [−0.21, 0.77], p = 0.2576; height β −0.26 [−0.77, 0.26], p = 0.3331; BMI β 0.62 [−0.19, 1.44], p = 0.1315). - Replication in PEACHES confirmed lower weight and height Z-scores through 5 years in children with neonatal infection treated with antibiotics versus non-exposed, particularly in boys; no differences in BMI Z-scores. - Antibiotic use after the neonatal period (first six years) was associated with increased BMI Z-scores in both boys and girls in SFBC; cumulative prescriptions showed a dose-dependent association with higher BMI Z-scores across the first six years (p < 0.001 in boys; p < 0.001 in girls, mixed models adjusted as above). - Microbiome: Neonatal antibiotic exposure led to persistent alterations in gut microbiome composition up to 24 months. Significant group differences by Hotelling tests on PCoA projections were observed at 1 month (p < 0.002) and 6 months (p < 0.003). Alpha diversity differences (PD) between exposed and controls were significant at 1 month (p = 0.014), 12 months (p = 0.014), and 24 months (p = 0.001). Richness was lower at 1 month in exposed infants, converged by 6 months, and exceeded controls at 12 and 24 months. - Bifidobacterium: Controls had high relative abundance of Bifidobacterium (~50% at 6 and 12 months; ~25–30% at 24 months), whereas exposed infants showed consistently lower relative abundance at all time points except a transient increase at 6 months. Exposed infants exhibited reduced diversity of fecal Bifidobacterium features across all four time points. Genome-resolved metagenomics showed at 1 month Bifidobacterium species were detected almost exclusively in controls (13 species across 7 control samples vs 1 species across 5 exposed samples; p = 5.3 × 10⁻⁶), notably B. bifidum (4/7 controls vs 0/5 exposed). - FMT experiment: Male germ-free mice receiving FMT from antibiotic-exposed infants gained significantly less weight than those receiving FMT from unexposed infants starting day 7 (p < 0.0001), with persistent differences and distinct microbiome trajectories; in female mice, growth effects were transient or not significant. Mouse fecal microbial richness differed between groups at multiple time points (e.g., day 21 p = 0.013; day 35 p = 0.004; day 43 p = 0.001).

The findings support a timing- and sex-dependent association between early-life antibiotic exposure and growth. Neonatal antibiotics were linked to reduced weight and height gain in boys but not girls through early childhood, while antibiotics later in infancy/childhood correlated with higher BMI in both sexes. The data suggest a microbiome-mediated mechanism: neonatal antibiotics disrupt the orderly maturation of the infant gut microbiome, including a sustained reduction in Bifidobacterium abundance and diversity and altered community structure persisting to 24 months. The FMT experiments provide functional support that antibiotic-altered infant microbiota can impair growth in male hosts, aligning with the observed human sex-specific associations. These results highlight the importance of preserving normal microbial colonization in the neonatal period and suggest that the consequences of antibiotic exposure depend critically on developmental timing, with potential implications for later body composition and metabolic risk.

Neonatal antibiotic exposure is associated with long-term perturbations of gut microbiome development and with reduced weight and height gain in boys during the first six years of life, whereas antibiotic use later in childhood is associated with increased BMI. The work integrates large cohort analyses, an independent replication cohort, longitudinal microbiome profiling, and germ-free mouse FMT to support a causal role of microbiome disruption. Clinically, judicious use of neonatal antibiotics and improved rapid diagnostics for neonatal sepsis are warranted to minimize unnecessary exposure. Future research should delineate mechanisms underlying sex-specific effects, assess the influence of antibiotic class and duration, evaluate interventions (e.g., targeted probiotics or microbiome restoration) to mitigate adverse effects, and extend follow-up into later childhood and adolescence to assess long-term metabolic outcomes.

- Observational cohort design limits causal inference and is susceptible to residual confounding despite adjustment for multiple covariates. - Breastfeeding data were unavailable in one cohort/data set, which may confound microbiome and growth associations; although in PEACHES, breastfeeding did not differ by exposure, differential duration could still influence outcomes. - Potential misclassification of infection versus empirical exposure and heterogeneity in antibiotic indications/durations may affect estimates. - Microbiome sub-study sample size was limited, reducing power to detect taxa-level differences and to generalize findings; stool sampling times varied and some analyses used subsets. - Generalizability may be limited to term infants in predominantly Finnish and German populations. - FMT experiments were performed in mice and showed sex-specific effects predominantly in males; translational extrapolation to humans requires caution and further mechanistic work.