Sepsis and Necrotising Enterocolitis

  • Brocklehurst P, Farrell B, King A, Juszczak E, Darlow B, Haque K, Salt A, Stenson B, Tarnow-Mordi W: Treatment of neonatal sepsis with intravenous immune globulin. The New England journal of medicine 2011, 365(13):1201-1211. http://www.ncbi.nlm.nih.gov/pubmed/21962214

This very large multi-center RCT (n=3500) showed no effect of intravenous immune globulin as a treatment for suspected sepsis, and subgroup analysis of the 1461 with proven sepsis was apparently also negative, as noted in the discussion, but not as far as I can see in the results. Previous small trials had suggested benefit, this trial puts that to rest. Mortality was identical, and disability at 2 years of age was unaffected.

In one sense this large multicenter RCT from Australia and New Zealand was a negative trial. The primary outcome (late-onset sepsis) was no different among the 550 VLBW babies under 32 weeks randomized to placebo compared to probiotics (ABCDophilus containing a lactobacillus, a bifidobacterium and a streptococcus). However there were 4.4% of controls who developed NEC and 2 % of the probiotic babies, a significant reduction.

One of several meta-analyses of probiotics to prevent NEC. A clear reduction in NEC (RR0.35) and in mortality (RR 0.42). This meta-analysis included a trial sequential analysis, confirming the results, and showing that we already knew this a while ago!

The meta-analysis was criticized by a number of individuals, including one comment that we couldn’t apply the results in the US because none of the studies were american. Another criticized the methodology, but it was co-authored by a statistician who thinks that we should not weight trials in meta-analysis, which I think no-one else agrees with.