Presented at the Neonatal Society 2015 Summer Meeting.
Battersby C, Mandalia S, Fitz-Simon S, Costeloe K, Modi N
Imperial College London
Background: Necrotising enterocolitis (NEC) is a devastating gastrointestinal disease predominantly affecting very preterm infants. The risks of death and long-term morbidities are high, aetiology is uncertain and the variations in enteral feeding strategies are widely believed to influence susceptibility. Here we address the null hypothesis that there is no variation in the incidence of severe NEC (defined as NEC leading to surgery or death) between neonatal networks in England.
Methods: We conducted a two-year, national population surveillance study of severe NEC (UKNC-NEC study). We extracted information from the National Neonatal Research Database that contains detailed extracts from the Electronic Patient Records of all admissions to neonatal units in England. These data were verified with local study leads. We used multivariable logistic regression to identify patient characteristics associated with severe NEC. We calculated the unadjusted and adjusted incidence and the ratio of observed to predicted number of cases for each neonatal network and depict this graphically using funnel plots.
Results: All regional networks (n=23) and neonatal units (n=163) in England participated. During the two year period 2012-13, 118,071 infants were born and admitted to neonatal care; 529 died or received surgery for NEC (461 <32 weeks GA). Of infants that received surgery 33.0% (139/421) died. The incidence of severe NEC per 1000 admissions was highest for infants born 24 weeks (112.4, 95% CI 90.0 to 134.8). Low gestational age and growth restriction were significant independent predictors of severe disease. The unadjusted funnel plot for infants born <32 weeks illustrates that the network-level variation in severe NEC rates is consistent with the pattern we would expect if the population rate of NEC is 3.07%. Two of twenty three networks fall just outside the 95% control limits, which may be expected if all variation in NEC cases at network level is due to chance rather than systematic differences between networks. Adjusting NEC rates for birth weight SDS, gestational age and antenatal steroids using a logistic regression model did not affect this conclusion.
Conclusion: We have shown no unusual variation in the incidence of severe NEC at network level in England. NEC is now a cardinal cause of death among very immature neonates; the low incidence rates call for national and international collaboration to test preventive stratagems in adequately powered randomised trials.
Corresponding author: firstname.lastname@example.org