Research in context
Evidence before this study
Extensive and regular searches of the literature written in English were conducted and we identified a small number of trials of probiotic in preterm infants, none with sepsis, and only one with necrotising enterocolitis as a primary outcome in a trial with multiple exclusions and very low incidence. The first Cochrane review of the topic was published in 2008, including a total of 11 trials with a meta-analysis suggesting reduced necrotising enterocolitis and death, but with no effect on sepsis. The authors commented on the heterogeneity of the trial participants and interventions, the difficulty of extracting outcome data for the high risk group below 1000 g birthweight, and the need for large trials.
Added value of this study
To the best of our knowledge, this study is the largest and the first statistically powerful published trial of the efficacy of a probiotic to reduce necrotising enterocolitis and sepsis in the preterm population. The population is likely to be more representative than the total population in recent meta-analyses. Furthermore, this is the first trial systematically to study stool colonisation in both groups of the trial and to emphasise both the incomplete colonisation in the active and the high cross colonisation in the placebo group. This trial reinforces the message from others that, in the short term, probiotic interventions are safe.
Implications of all the available evidence
The current results support the view that strains should be assessed separately and challenge the validity of combining trials of different interventions in meta-analyses. It is not plausible that cross colonisation with administered probiotic is confined to this trial; routine use of probiotic is likely to modify the gut microbiota of infants other than those for whom it is prescribed. We conclude that at the present time the evidence from clinical trials does not support the routine use of probiotics to prevent necrotising enterocolitis and sepsis in the preterm infant.