SummaryThe effect of route of delivery and use of epidural anaesthesia on the mortality and morbidity of extremely premature babies was studied using retrospective data and multivariate analysis to correct for confounding variables. Women delivering babies less than 32 weeks after the last menstrual period or of birth weight less than 1500g were included and babies dying before the onset of labour, before admission to this hospital or with lethal congenital abnormalities were excluded. Survival to one month and absence of significant intraventricular haemorrhage were chosen as end points. The only group in which caesarean section appeared to confer a significant benefit was those with a gestational age of less than 27 weeks. Epidural anaesthesia conferred no significant benefit. There was no evidence that the effect of the route of delivery or of an epidural anaesthetic varied with the presentation.