The collaborative national survey on morbidity and mortality in preterm and small for gestational age infants in the Netherlands enrolled initially 1338 infants born in 1983. The relationship between maximal serum total bilirubin concentration in the neonatal period and neurodevelopmental outcome in the survivors of this cohort was studied. This relationship at the corrected age of 2 years was previously reported. A dose-response relationship between maximal serum total bilirubin concentration and risk of adverse outcome was observed in the 831 surviving children. The present study reassessed the relationship at the age of 5 years in 814 children. There was no significant difference in mean maximal serum total bilirubin concentration between the children with and without a handicap. This was confirmed by logistic regression analysis. After correction for seven suspected confounding factors (gestational age, birth weight, intracranial hemorrhage, ventriculomegaly, seizures, bronchopulmonary dysplasia, and socioeconomic status) the estimated odds ratio was 1.2 (confidence interval 0.89, 1.43) per 50 mumol/L increase of total bilirubin. However, in this analysis an interaction between bilirubin and intracranial hemorrhage was observed. Therefore, the cohort was divided into two groups according to the absence or presence of an intracranial hemorrhage. Logistic regression analysis including four suspected confounding factors (gestational age, ventriculomegaly, seizures, and socioeconomic status) was then again applied. In children who had suffered from an intracranial hemorrhage in the neonatal period the estimated odds ratio was 1.84 (confidence interval 1.08, 3.15) per 50 mumol/L increase of bilirubin. Similar results were obtained treating bilirubin as a categorized exposure. The odds ratio in children without a hemorrhage was 1.05 (confidence interval 0.80, 1.38), probably because of the small number of surviving handicapped children.