No evidence of increased harm from steroid treatment in extremely preterm babies

A national study of neonatal care in England and Wales has found no evidence that giving extremely preterm babies low-dose hydrocortisone treatment to lung complications was associated with a higher risk of harm.

The study, carried out by researchers at Imperial College London and involving data from thousands of babies, was carried out to investigate potential risks linked to the treatment, which may be given to very premature babies in their first ten days of life to help prevent the development of long-term lung problems. It follows concerns of an increased risk of intestinal perforation in very preterm babies, which can be life threatening.

The researchers explain that while there was no evidence of increased risk of the bowel complication with the treatment, they did observe a higher rate of death before discharge among babies classified as receiving hydrocortisone. However, they caution that this was an observational study, it is not possible to establish whether hydrocortisone was a direct cause of that difference, or if it was due to other factors.

Their findings are published in the journal Archives of Disease in Childhood – Fetal and Neonatal Edition.

Dr Cheryl Battersby, senior author and Clinical Associate Professor in Neonatal Medicine at Imperial College London, said: “We cannot say that this treatment is the cause of infant deaths, as the evidence from this study does not support that. But what we can say, is that this is an important finding which requires further investigation. Extremely premature babies are a high-risk patient cohort, more likely to have health issues, so we need to take care interpreting these findings. Ultimately, we hope these findings help to inform further research and improve the clinical care and health of babies born premature.

“In this study we observed higher mortality among babies classified as receiving early hydrocortisone. This is an important finding, but the study cannot tell us whether hydrocortisone contributed to those deaths. We used routinely collected clinical data and could not be certain exactly why hydrocortisone was prescribed for each baby, as it can be used to treat a number of conditions. Babies who received it may therefore have differed from those who did not in ways our analysis could not fully account for. Some may have received it because they were already unwell. Our findings should be considered alongside evidence from randomised trials when assessing the potential benefits and risks of early hydrocortisone.”

In the study, researchers at Imperial and collaborators examined routinely collected records for 15,746 babies born before 28 weeks of pregnancy between 2016 and 2023. They compared outcomes in groups of babies matched on a range of recorded characteristics. In the matched analysis, 210 of 988 babies (21.3%) classified as receiving early hydrocortisone died before discharge, compared with 295 of 1,827 (16.2%) in the comparison group.

The study was designed primarily to investigate focal intestinal perforation, a serious bowel complication. It found no evidence of an increased rate in the hydrocortisone group. Death before discharge was a secondary outcome. The mortality finding warrants further investigation, but should not be interpreted as evidence that the treatment caused deaths.

Although the researchers matched babies on recorded characteristics, the clinical reason for giving hydrocortisone was not available in the dataset. Some babies may have received it because they were already becoming unwell. Differences in illness after treatment began, or in care between neonatal units, could also have contributed to the observed results. Earlier randomised trial evidence has not established an increase in mortality with this treatment.

‘Association between prophylactic hydrocortisone and focal intestinal perforations in preterm infants born before 28 weeks gestation: an observational propensity matched population study in England and Wales’ by Aveline AS et al. is published in Archives of Disease in Childhood – Fetal and Neonatal Edition. DOI: 10.1136/archdischild-2026-330846.