Background: Necrotising enterocolitis (NEC) is growing in global relevance. Aetiology and pathophysiology are uncertain; non-invasive diagnostic tests and case-definition lacking; population incidence data scant, and evidence to inform preventive feeding strategies are inadequate. The National Neonatal Research Database (NNRD) comprises data from infants admitted to neonatal units in England and offers opportunities to facilitate population studies.
Methods: I performed a systematic review of global NEC incidence; secured the participation of all English neonatal units in a prospective two-year study employing the NNRD as the source of data; quantified the burden of severe NEC (confirmed at laparotomy and/or leading to death); investigated the impact of early maternal breast milk (MBM), and avoidance of bovine products (formula/fortifier) on NEC, and developed a case-definition for NEC suitable for surveillance and research. Statistical methods included multivariate logistic regression, funnel plot and propensity analysis.
Results: 118,073 infants (14,678 <32 weeks (w) gestational age (GA)) were admitted to 163 English neonatal units (23 networks) in 2012-2013; 531 (462; 3.15% of infants <32w GA) developed severe NEC; 20% died without a laparotomy; a third who received a laparotomy died. Among infants <32w there was no strong evidence of variation in relation to the adjusted national incidence of 3.13% (95%CI 2.85, 3.42) despite variation in feeding practices. Early vs. late or no MBM (within 7 days) resulted in an Absolute Risk Difference (ARD) of -0.88% (95% CI -1.15, -0.61), Relative Risk (RR) 0.69 (95% CI 0.60, 0.78); equivalent figures for no vs. any bovine products within 14 postnatal days, were ARD -0.65% (95% CI -1.01, 0.29), RR 0.61 (95% CI 0.39, 0.83). The ordinal NEC score allocated three points to pneumatosis, two to blood in stools and one to abdominal tenderness; abdominal discolouration; the composite of increased and/or bilious aspirates AND abdominal distension; one or more of pneumoperitoneum, fixed loop and portal venous gas. Infants with a score of ≥2 (<30 w); ≥3 (30 to <37 w); ≥4 (≥37w GA) were considered to have NEC.
Conclusions: This thesis provides, for the first time, population incidence of severe NEC in England. The proposed NEC score and GA-specific case-definition offer opportunity to strengthen research efforts. Commencing MBM early and avoiding bovine products may reduce NEC but effect sizes appear small. The lack of network variation in severe NEC despite differences in feeding practices highlights the need to address ongoing uncertainties.