Antibiotics in pregnancy and infancy do not raise autoimmune disease risk in children

By Vijay Kumar MalesuReviewed by Susha Cheriyedath, M.Sc.Aug 25 2025

Large-scale national data offer reassurance for families and clinicians that antibiotics used to treat infections in pregnancy or early infancy do not raise children’s autoimmune risk, though subtle signals in specific subgroups need ongoing attention.

Study: Exposure to antibiotics during pregnancy or early infancy and the risk of autoimmune disease in children: A nationwide cohort study in Korea. Image Credit: okskaz / Shutterstock

In a recent study published in the journal PLoS Medicine, researchers determined whether systemic antibiotic exposure during pregnancy or the first six months of life increases children’s risk of autoimmune diseases and examined subgroup-specific risks using confounding control.

Background

By age one, many infants and pregnant women receive antibiotics for urinary or respiratory infections. Families wonder whether these medicines, while vital, might subtly reshape immunity and raise the risk of autoimmune conditions such as type 1 diabetes mellitus, juvenile idiopathic arthritis, and inflammatory bowel disease (IBD). Early-life microbiome disruption, genetic susceptibility, and social determinants like access to care complicate the picture, while untreated infections also threaten mothers and babies. Headlines amplify fear without clarifying causation in practice. Clinicians need clear guidance grounded in population-level evidence. To disentangle infection from treatment and inform safe, equitable care, further research is needed.

About the study

Using the National Health Insurance Service–National Health Insurance Database (NHIS–NHID), investigators assembled two nationwide, mother-child linked cohorts for births from April 2009 to December 2020. Linkage used family insurance identifiers and delivery dates. Both cohorts were restricted to dyads with documented infections to reduce confounding by indication. Antibiotic exposure came from physician-prescribed systemic agents (Anatomical Therapeutic Chemical J01): any prescription from 30 days before the last menstrual period (LMP) through delivery for pregnancy, and any prescription in the first six months of life for infancy. Outcomes were type 1 diabetes mellitus, Crohn’s disease, systemic lupus erythematosus, juvenile idiopathic arthritis, ulcerative colitis, and autoimmune thyroiditis (Hashimoto’s thyroiditis).

Covariates included maternal sociodemographics, comorbidities and medicines (for example, nonsteroidal anti-inflammatory drugs (NSAIDs), systemic corticosteroids, chronic obstructive pulmonary disease (COPD) drugs), healthcare use (outpatient, inpatient, emergency room (ER) visits), obstetric factors, infant sex and perinatal factors, and body mass index (BMI) and smoking. Propensity Scores (PS) supported stabilized Inverse Probability of Treatment Weighting (IPTW), with balance assessed by standardized mean difference (SMD); absolute SMD (aSMD) < 0.1. Hazard Ratios (HR) with 95% Confidence Intervals (CI) were estimated via Cox proportional hazards models; sibling-matched analyses compared exposure-discordant pairs. Reporting followed Strengthening the Reporting of Observational Studies in Epidemiology (STROBE).

Study results

Across the pregnancy cohort, 1,516,574 children were exposed in utero and 1,186,516 were unexposed; in the infancy cohort, 1,925,585 received antibiotics in the first six months and 1,421,464 did not. Median follow-up was 7.6 years for pregnancy analyses and 7.4 years for infancy analyses. After IPTW weighting, the analytic samples were smaller (for example, ~1.34M exposed vs 1.04M unexposed in pregnancy, ~1.40M vs 1.28M in infancy), but well balanced on covariates.

After restricting to families with infections and applying stabilized IPTW, prenatal exposure showed no association with any autoimmune outcome: type 1 diabetes mellitus (HR 1.14, 95% CI 0.96–1.35), Crohn’s disease (HR 1.16, 95% CI 0.98–1.36, p = 0.076), juvenile idiopathic arthritis (HR 1.02, 95% CI 0.85–1.22), ulcerative colitis (HR 1.02, 95% CI 0.76–1.37), systemic lupus erythematosus (HR 0.70, 95% CI 0.49–1.01), and autoimmune thyroiditis (Hashimoto’s thyroiditis) (HR 1.06, 95% CI 0.91–1.23).

Early-infancy exposure was likewise not associated with increased risk: type 1 diabetes mellitus (HR 1.05, 95% CI 0.88–1.26), juvenile idiopathic arthritis (HR 1.11, 95% CI 0.93–1.33), ulcerative colitis (HR 0.95, 95% CI 0.67–1.36), Crohn’s disease (HR 1.07, 95% CI 0.91–1.25), systemic lupus erythematosus (HR 1.46, 95% CI 0.95–2.26), and autoimmune thyroiditis (Hashimoto’s thyroiditis) (HR 1.14, 95% CI 0.97–1.33).

Sibling-matched analyses, comparing exposure-discordant brothers and sisters, also yielded null associations across all outcomes, reinforcing control for shared genetic and environmental factors. Notably, “crude” full-cohort contrasts before infection restriction and weighting appeared to show elevated risks for some outcomes, but these signals attenuated to the null once indication and covariates were addressed in IPTW and sibling designs.

Subgroup analyses suggested small, hypothesis-generating patterns. During pregnancy, exposure to broad-spectrum antibiotics, and specifically cephalosporins, as well as exposure in the first or second trimester, was associated with a modest increase in Crohn’s disease risk. These signals strengthened when exposure was redefined as two or more antibiotic prescriptions. During infancy, the risk of autoimmune thyroiditis was modestly higher among males and after exposure within the first two months of life.

Sensitivity analyses, testing alternative exposure definitions, restricting to singleton births, limiting to breastfed infants, and excluding children with maternal autoimmune disease, were broadly consistent with the primary findings.

For families, these estimates mean that treating bona fide infections during pregnancy or early infancy was not associated with an increased overall risk of autoimmune disease. However, the study cannot entirely exclude residual confounding. For clinicians, they emphasize careful indication, attention to antibiotic class and timing in special situations, and continued surveillance of subgroups flagged by the secondary analyses.

Conclusions

Taken together, this nationwide analysis offers reassurance for families and clinicians: when antibiotics are prescribed for clear infections in pregnancy or early infancy, the long-term risk of autoimmune disease in children appears minimal. The careful restriction to infected populations, proper confounder control with stabilized IPTW, and sibling-matched comparisons together reduce bias that has clouded prior work. Nonetheless, signals for Crohn’s disease with certain prenatal exposures and for autoimmune thyroiditis in male infants after very early exposure warrant cautious follow-up.

The authors also highlight residual confounding by unmeasured factors as a key limitation. Judicious prescribing remains essential, treat infections promptly, avoid unnecessary courses, and continue monitoring specific subgroups over time.