Presented at the Neonatal Society 2015 Summer Meeting.
Costeloe K1,2, Wilks M1,3, Hardy P4, Millar M1,3 for The Probiotics in Neonates Study Collaborative Group
1 Queen Mary, University of London, UK
2 Homerton University Hospital Foundation NHS Trust, UK
3 Barts Health NHS Trust, UK
4 National Perinatal Epidemiology Unit, University of Oxford, UK
Background: Necrotising enterocolitis (NEC) and sepsis are important causes of death and morbidity in preterm infants. Administration of commensal bacteria to modify the bowel flora may strengthen intestinal barrier function and prevent NEC and some cases of severe sepsis. Meta-analysis of over 20 published trials of probiotics (1) suggests that probiotics prevent NEC and death, however use is limited. This may be because of lack of robust design and concern about the generalisability of some of the trials (2). The aim of the PiPS trial was to test a single bacterial strain probiotic product, Bifidobacterium breve BBG-001 to reduce NEC, sepsis and death in an unselected population of preterm babies large enough to give clear results.
Methods: The trial was funded by the NIHR through the Health Technology Assessment (HTA) and approved by South Central Oxford A Research Ethics Committee and performed to ICH-GCP. PiPS is a multi-centre double blind randomised placebo controlled trial of Bifidobacterium breveBBG-001, 2.1 to 5.3 x 108 CFU daily in infants below 31 weeks of gestation, randomised within 48 hours of birth. Colonisation with B breve was monitored by culture and PCR of stools at 2w postnatal and 36w postmenstrual age. Primary outcomes were: NEC ≥Bell stage 2, late onset sepsis and death. Results are presented by intention to treat adjusted for sex, gestation, randomisation within 24 hours and allowing for clustering of multiples.
Results: 1315 infants were randomised, 5 withdrew, median gestation 28.0w, birthweight 1010g, age starting the intervention 44 hours. No adverse events related to the intervention were reported.
Analyses of the primary outcomes by subgroups defined by gestational age and birthweight were not suggestive of differential effects in more mature babies and analysis of subgroups defined by colonisation status did not suggest that efficacy was impacted by cross colonisation. Nor were there any differences in a range of clinical and microbiological secondary outcomes.
Conclusion: This probiotic intervention shows no evidence of benefit in this population. This result supports the view that different probiotic strains should be assessed separately and challenges the validity of combining trials using different interventions in meta-analyses.
Corresponding author: firstname.lastname@example.org
1. AlFaleh K, Anabrees J. Cochrane Database of Systematic Reviews 2014, Issue 4.
2. Mihatsch WA et al. Clinical Nutrition 2012;31:e15