OBJECTIVE To determine if the presence of a neonatal surgical facility on site has any effect on mortality and morbidity of very premature infants with necrotising enterocolitis (NEC).

DESIGN AND SETTING Retrospective review of infants of less than 29 weeks gestation cared for in the seven perinatal centres in New South Wales.

PATIENTS Between 1992 and 1997, 605 infants were cared for in two centres with in house surgical facilities (group 1) and 1195 in five centres where transfers were required for surgical management (group 2).

RESULTS Although use of antenatal steroids was significantly lower in group 1 centres than group 2 centres (74% v 85% respectively), and the incidence of hyaline membrane disease, pneumothorax, and NEC was higher, mortality was identical (27%). Fifty two (9%) infants in group 1 and 72 (6%) in group 2 of comparable perinatal characteristics and CRIB (Clinical Risk Index for Babies) scores developed radiologically or pathologically proven NEC. The overall mortality of infants with NEC in group 1 was lower but this was not statistically significant (27% v 35%). Significantly more infants with NEC in group 1 had surgery (69%v 39%). There were fewer postoperative deaths in group 1 and more deaths without surgery in group 2. The duration of respiratory and nutritional support and hospital stay for the survivors were similar in the two groups. In a multivariate analysis, shorter gestation was the only factor associated with mortality in infants with NEC; the presence of in house surgical facilities was not.

CONCLUSIONS There were no significant differences in outcome of premature infants with NEC managed in perinatal centres with or without on site surgical facilities. Early transfers should be encouraged. This finding may have implications for future planning of facilities for neonatal care.