The composition of an infant's gut microbiome during the first year of life may carry consequences that extend well beyond infancy — shaping immune programming, metabolic trajectories, and disease susceptibility for decades. Understanding which early-life exposures most reliably alter that microbial architecture is therefore a meaningful clinical and public health question, and this systematic review with meta-analysis attempts to untangle two of the most common: cesarean delivery and perinatal antibiotic use.
Drawing on 11 studies enrolling 5,309 healthy term infants across Europe, North America, Africa, and Asia — predominantly prospective cohorts with one randomized controlled trial — the analysis found no statistically significant difference in overall bacterial abundance between vaginally delivered and cesarean-born infants (mean difference 3.44%, 95% CI −2.00 to 8.89). Critically, however, the pooled heterogeneity was extreme (I² = 99%), and the evidence certainty was rated very low, complicating any firm conclusions. Among infants unexposed to perinatal antibiotics, a directional trend toward higher bacterial abundance following vaginal delivery emerged, suggesting that antibiotic exposure may be confounding or masking a birth-mode signal. Narrative synthesis across included studies consistently showed reductions in microbial diversity and depletion of Bacteroides species in antibiotic-exposed and cesarean-born infants.
This meta-analysis sits within a now-substantial body of literature linking cesarean delivery and early antibiotic exposure to altered microbiome colonization patterns, particularly deficits in anaerobes like Bacteroides and Bifidobacterium that are transmitted via the maternal birth canal and fecal microbiota. The interaction between these two exposures — which frequently co-occur, since cesarean deliveries routinely involve prophylactic antibiotics — has been underexplored, and this paper's attempt to disentangle them is methodologically valuable. The extreme heterogeneity and very low certainty rating are significant caveats: pooled abundance metrics may obscure clinically meaningful strain-level or genus-level differences that sequencing studies are capable of detecting. Future trials with standardized microbiome readouts and longer follow-up will be needed before population-level guidance on microbiome restoration strategies, such as vaginal seeding or probiotic supplementation, can be evidence-grounded. For now, this analysis is best characterized as confirmatory of existing directional evidence, not paradigm-shifting.