Dr Claire Caldwell | Dr Deena Patel

For references, click here

The collective cortical scar on neonatologists that runs deepest is surely the use of postnatal steroids in premature babies. It’s sometimes difficult for newer members of the team to fully understand our reluctance to use steroids to extubate babies (or, in the case of the PREMILOC study, prophylactically). The purpose of this piece is to pull together the history of post-natal steroid use and explore possible future strategies.

Neonatal intensive care has seen huge progress since the development of the first ventilators which were able to give time-cycled, pressure-limited ventilation to the smallest prem babies. The Cochrane review of antenatal maternal steroids1 showed the benefits of steroids for in-utero lung maturation, and this has led to fewer babies needing intubation.

After pulmonary surfactant was developed, we saw a decreased risk of pneumothorax, pulmonary interstitial emphysema, bronchopulmonary dysplasia; intra-ventricular haemorrhage and mortality.2 However, there were still babies who ‘failed’ extubation, and after a number of attempts were still ‘stuck’ on the ventilator.

During the 1990s, we discovered the underlying inflammatory process going on within the lungs. So we started using potent courses of steroids, which allowed successful extubation in a lot of cases. The treatment schedules were varied – from 3 days of dexamethasone started within 12 hours of age3, up to a 42 day tapering course started after 15 days on ventilated babies4.

The long term cost of steroids would unfortunately not become obvious until much later. Meta-analysis of longer term follow up5 showing a relative risk of cerebral palsy of 2.2 in surviving children who had been treated with steroids. A moratorium on postnatal steroid use began, with the Cochrane review6 concluding that “The benefits of early postnatal corticosteroid treatment (<96 hours) may not outweigh the known or potential adverse effects of this treatment” and the American Academy of Pediatrics7 limiting use of steroids to ‘exceptional clinical circumstances’.

Slowly, however, the pendulum began to swing back the other way. Clinicians were exploring a potential middle ground for using smaller doses/ duration of steroids in a carefully selected group of babies. The DART trial8 looked at a low dose, 10 day course of dexamethasone for babies <28 weeks or <1000g, unable to be extubated by one week of age. Unfortunately, they couldn’t recruit enough patients to adequately power the trial to study neurodevelopmental outcomes at 18-24 months.

A more pragmatic approach came out of a meta-analysis of 20 RCTs,9 which concluded that if the baseline risk of bronchopulmonary dysplasia (BPD) was below 35%, corticosteroid treatment significantly increased the chance of death or cerebral palsy – whereas on the other hand, if the risks for BPD were above 65%, it reduced this chance.

Many neonatal units are now much more cautious with their use of steroids, and parents are told about the higher risk of cerebral palsy. Since there are no other published protocols for low dose dexamethasone, the DART protocol is still used widely as a template for treatment.

At the same time as lower doses of dexamethasone were being introduced, attention also shifted to more ‘physiological’ treatment with hydrocortisone. This was driven by the finding that babies <1500g birth weight had varying levels of baseline cortisol, and response of cortisol levels to ACTH stimulation. Significantly lower cortisol secretion in response to ACTH was found in babies that later developed BPD (p=0.006).10  These investigators set up a multi-centre RCT of hydrocortisone versus placebo in ventilated babies between 500-999g birth weight, with treatment started less than 48 hours after birth11. Unfortunately, this had to be stopped early due to increased gastro-intestinal perforations in the treatment arm of the trial.

The PREMILOC trial12 randomised 523 babies born <28 weeks to either 10 days of hydrocortisone or placebo (started by 24 hours of age). The number of babies needed to power the study was 786, but sadly the trial was stopped early due to funding issues. The outcome was 60% survival without BPD in the treatment arm, versus 51% in the placebo group (p=0.04). There was no difference in gastro-intestinal perforation rates between the two groups (p=0.56). Two year follow up showed no difference in mild, moderate or severe neurodevelopmental delay between the control group and the treated group, as well as no difference in the incidence of CP between the two groups13 – however the trial was not powered well enough to confirm this. The authors decided that more randomised studies would be needed to look at the “neurodevelopmental safety of hydrocortisone in extremely preterm infants.” It will also be interesting to see the longer term outcomes of the PREMILOC babies.

There are still so many unanswered questions for neonatologists: do we take a prophylactic approach to BPD, treating all babies born before 28 weeks with hydrocortisone, knowing that not all of these babies will develop BPD? (Worryingly, there was a higher rate of sepsis in the sub-group of 24-25 weeks premature babies).

…So, should we test for low cortisol levels before we use this treatment? Is hydrocortisone really safe or do we need more RCTs and longer follow up time before we’re sure?

…Or do we wait until other all options are exhausted and then offer dexamethasone, with its risks for worse neurodevelopmental outcomes?

In 2018, a meta-analysis of 47 RCTs ranked treatments in order of preference. Both high and low dose dexamethasone were ranked higher than hydrocortisone at preventing BPD, and low dose dexamethasone was ranked first (i.e. less risk of CP) compared to both hydrocortisone and high dose dexamethasone for CP risk. These finding might suggest that low dose dexamethasone continues to have a role to play.14

The pendulum has not yet settled into place, but hopefully with this background knowledge it is possible to understand more fully just how far it has swung.

Dr Claire Caldwell, Neonatal GRID trainee, with Dr Deena Patel, Consultant Neonatologist, Chelsea and Westminster Hospital

For references, click here

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.