Presented at the Neonatal Society 2018 Autumn Meeting.
Robertson C1,2, Savva GM3, Jones J1, Maimouni H4, Clapuci R1, Minocha A1, Hall L2, Clarke P1,4
1 NICU, Norfolk & Norwich University Hospital, Norwich, UK
2 Hall Laboratory, Norwich, UK
3 Analytical Sciences Unit, Quadram Institute Bioscience, Norwich, UK
4 Norwich Medical School, University of East Anglia, Norwich, UK
Background: Necrotising enterocolitis (NEC) is a leading cause of death worldwide among preterm very low birth weight neonates. Meta-analyses show that prophylactic dual-strain probiotics are effective at reducing NEC. In January 2013 our NICU introduced dual-strain probiotics (Lactobacillus acidophilus and Bifidobacteria bifidum) routinely for all neonates born at <32 weeks’ gestation and/or of birth weight <1500 g. We have reviewed our centre’s NEC rates in the 5-year epochs before and after introduction of routine prophylactic probiotics.
Methods: Eligible high-risk neonates were inborn at <32 weeks’ gestation and/or of birth weight <1500 g in two 5-year epochs: pre-probiotic epoch 1/1/08 to 31/12/12; routine probiotic epoch 1/1/13 to 31/12/17. We included out-born neonates if transferred in within 72 h of birth, unless their transfer was due to abdominal concerns. Potentially confounding variables were collected on all eligible neonates. Source data were gathered from BadgerNet, casenotes, radiology, laboratory, histology, and autopsy reports. For all infants with a BadgerNet recorded diagnosis of NEC/?NEC (any grade), two clinicians reviewed medical records, x-rays, blood results, surgical notes, histopathology reports, and death certificates to determine definitive NEC cases using slightly-modified versions of three different NEC diagnostic classification systems (Bell’s, Vermont-Oxford, Battersby). The effect of epoch on definite NEC, defined as modified Bell’s stage ≥2a, was estimated using multiple logistic regression, controlling for gestational age, birthweight, and any other covariate observed to vary between epochs.
Results: Irrespective of NEC classification system used, there were significant drops in NEC rates between pre- and post-probiotic epochs. Numbers of NEC Bell’s stage ≥2a cases pre-probiotics were 35/469 (7.5%) versus post-probiotics: 17/513 (3.3%), p-value=0.0058, Chi-sq. test, (Fig). Both epochs were similar with regard to use of antenatal steroids, delivery mode, gender, birthweight, gestational age, small for gestational age, history of prolonged rupture of membranes, nonsteroidal anti-inflammatory drug use, and days to first enteral feeds. The odds ratio for NEC associated with the use of probiotics was 0.43 (95% CI: 0.23–0.76; p=0.0047). This association was not attenuated by adjusting for birthweight, gestational age, or milk type used for feeds.
Conclusion: Our centre has observed a >50% reduction in NEC rates since introducing dual-strain probiotics, with effect-size mirroring the RCT meta-analyses and corroborating our ongoing use of dual-strain probiotics to prevent NEC in high-risk neonates. While we considered many possible confounding variables and cannot rule out an alternative explanation for the observed drop in NEC rates between epochs, the most likely explanation for our markedly lower NEC rates would presently appear to be our routine use of dual-strain probiotics.
Corresponding author: firstname.lastname@example.org