The role of suspected or confirmed neonatal sepsis in modifying the risk of atopic disease during childhood was assessed. Children with early-onset neonatal sepsis were identified from a cohort of neonates, hospitalized between 1990 and 1995. Of 196 individuals, 140 were recruited (71.4%). Pre- and postnatal history was ascertained from neonatal medical records. Based on clinical symptoms and a positive blood culture or at least three of initially defined laboratory or bacteriological criteria, they were stratified in either confirmed neonatal sepsis (CS) or suspected sepsis (SS) group. A control group (C) comprised children who were never hospitalized during infancy (n = 696). Primary end-point was the development of atopic dermatitis, bronchial asthma or allergic rhinitis during childhood (mean age 8.4 yr, range 5.7-12.4). CS and SS children had a higher prevalence of atopic dermatitis (CS 15.7%, SS 21.4%) compared with controls (C 5.2%, p < 0.001). Similarly, children with SS (7.1%), but not with CS (4.3%) had significantly more often a doctor's diagnosis of bronchial asthma compared to controls (1.9%, p = 0.02). No difference in the prevalence of allergic rhinitis was observed (CS 4.3%, SS 10%, C 8.3%). After adjusting for parental history of atopic disease and demographic factors, no significant difference for the risk to develop atopic dermatitis, asthma or allergic rhinitis among the groups was calculated in children with normal birth weight (>2500 g). Our data failed to show a possible link between hospital admission with SS and development of atopic disease.