Perinatal antibiotic exposure and risk of childhood infections: a retrospective cohort study.

BACKGROUND

Epidemiological studies report associations between antibiotics given during pregnancy, childbirth, and infancy and subsequent risk for childhood infections. The specific role of intrapartum and neonatal antibiotic exposures is not well-described.

METHODS

Retrospective cohort study of healthy term infants through 6 years of age. The primary exposure was perinatal antibiotics, defined as intravenous intrapartum antibiotic administration during the admission for childbirth or administered to the infant ≤3 days after birth. The primary outcome was infection-related inpatient encounters. Adjusted multivariable-adjusted Cox proportional hazards and marginal means/rates models were used to investigate the association between exposure and outcome. A secondary analysis examined the association between early infant antibiotics administered during the first three months after birth, including perinatal antibiotics, and the outcome.

FINDINGS: Of 13,919 infants, 3936 (28%) had perinatal antibiotic exposure. 1294 (9.3%) children had 1619 inpatient encounters, of which 988 (61%) were infection-related. Infection-related inpatient encounters occurred in 265 (6.7%) children exposed to perinatal antibiotics, compared to 584 (5.8%) unexposed children (risk difference 1.2%, 95% CI 0.3-2.1%, p = 0.005). In multivariable-adjusted models, perinatal antibiotic exposure was not associated with infection-related inpatient encounters [Cox model: aHR 1.16, 95% CI 0.95, 1.51, p = 0.15; marginal rates/means model: aHR 1.22, 95% CI 0.98, 1.51, p = 0.08]. Early infant antibiotics were also not associated with the outcome.

INTERPRETATION: In this study of 13,919 newborns across 24 primary pediatric practices, perinatal or early infant antibiotic exposure was not associated with subsequent early childhood hospitalization for infectious diseases.

FUNDING: Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, funding (K23 HD088753) supported this study.