After Dutch NICUs made probiotics routine, the tiniest babies had less bowel disease

Necrotising enterocolitis (NEC) is one of the most feared complications of extreme prematurity: the bowel tissue inflames and can die, and the smallest babies are the most vulnerable. For years, trials have suggested that giving preterm infants probiotics — live bacteria to seed a healthier gut — lowers the risk of NEC, of bloodstream infection and of death. But the evidence has been thinnest, and the safety worries loudest, exactly where the stakes are highest: in infants born before 28 weeks. A multicentre study published in the European Journal of Pediatrics on October 31 looks at what happened when Dutch units made a probiotic the default for those babies [s1].

A natural experiment

This was not a randomised trial but a pre-post implementation study — a comparison of outcomes before and after a change in routine care [s1]. From 2021, Dutch neonatal intensive care units adopted a three-strain probiotic product (Bifidobacterium infantis Bb-02, Bifidobacterium lactis BB-12 and Streptococcus thermophilus TH-4) as standard of care [s1]. The authors compared extremely preterm infants — gestational age under 28 weeks — admitted to six Dutch NICUs between 2018 and 2023, splitting them into a pre-implementation cohort and a post-implementation cohort [s1].

In total 481 infants were included: 212 before implementation and 269 after [s1]. Clinical data were collected over the first 29 days of life [s1]. NEC was defined as modified Bell's stage IIA or worse; the other outcomes were late-onset sepsis, all-cause mortality, growth velocity and time to full enteral feeds [s1]. NEC was compared with logistic regression, growth velocity with linear regression, and time to full feeds with survival analysis [s1].

The biological rationale is that extremely preterm infants are born with a gut that has barely been colonised, then spend weeks on antibiotics and intravenous nutrition in intensive care — conditions that leave the microbiome dominated by potentially harmful organisms. Seeding it with defined strains of Bifidobacterium and Streptococcus is meant to tilt that balance before the inflammation of NEC can take hold.

What changed

After the switch, the adjusted odds ratio for NEC was 0.49 (95% confidence interval 0.25-0.99, p = 0.049) — roughly half the odds, with an interval whose upper bound just clears 1 [s1]. For late-onset sepsis the adjusted odds ratio was 0.87 (0.59-1.28, p = 0.49) and for all-cause mortality 0.67 (0.35-1.27, p = 0.22) — both pointing in the same direction but neither statistically distinguishable from no effect [s1].

Two feeding and growth measures moved more clearly. Time to full enteral feeds fell by two days (p < 0.001), and probiotic use was associated with a gain in growth velocity of 1.8 g/kg per day (p < 0.001) [s1]. Getting an extremely preterm baby onto full milk feeds sooner matters in its own right, because it shortens the time on intravenous nutrition and the central lines that carry infection risk.

The safety question

The worry with giving live bacteria to a baby whose immune system is barely formed is that the probiotic organism itself causes an infection. In this cohort, one infant (0.4%) developed probiotic-associated sepsis [s1]. That is a real event, not a theoretical one — but a single case among 269 treated infants is the kind of rare harm these programmes are designed to accept against a halving of NEC odds, provided units can identify and treat it.

How much to make of it

The design is the main caveat. A before-and-after comparison is not a randomised trial: neonatal care changes continuously, and some of the improvement could reflect other advances between 2018 and 2023 rather than the probiotic. The NEC and feeding results were adjusted for measured differences between the cohorts, but unmeasured drift cannot be ruled out [s1]. The mortality and sepsis odds ratios, while favourable, had wide intervals that include no benefit [s1].

What gives the finding weight is that it is consistent with a large existing body of randomised evidence. A Cochrane review of probiotics for preventing NEC in very preterm and very-low-birthweight infants has found a reduction in NEC overall, while flagging uncertainty about which strains work and how well the benefit extends to the most immature babies [s2]. This study's contribution is real-world data from exactly that hardest group, the under-28-week infants, where the trial evidence has been weakest [s1].

What to watch

Whether other national networks report the same NEC reduction after standardising on a defined multi-strain product; whether the growth and feeding benefits hold up; and whether surveillance for probiotic-associated sepsis keeps that rare harm rare as use widens.

This article describes a single observational implementation study and is informational only. It is not medical advice, and decisions about a preterm infant's care belong with the neonatal team.