A case report has suggested that exposure to erythromycin through breast milk might cause infantile hypertrophic pyloric stenosis. This study therefore examined whether macrolides, transmitted via breast milk, increase the risk of infantile hypertrophic pyloric stenosis in neonates. A population-based cohort study was conducted, based on data from a prescription registry, the Danish Birth Registry and North Jutland County's hospital discharge registry, Denmark, and comprising 1166 pregnant women who had been prescribed macrolides from birth to 90 d postnatally, and 34,690-41,778 pregnant women as controls. The odds ratios for infantile hypertrophic pyloric stenosis varied between 2.3 and 3.0 according to different periods of postnatal exposure, and after stratification for gender they were 10.3 [95% confidence interval (95% CI) 1.2-92.3] for girls and 2.0 (95% CI 0.5-8.4) for boys. The use of macrolides during breast-feeding increases the risk of infantile hypertrophic pyloric stenosis.