Antenatal Steroids and Child Health

(with Jennifer Hutcheon and Erin Strumpf)

RATIONALE: Randomized clinical trials have consistently shown that administering antenatal corticosteroids to pregnant women at risk of preterm delivery can reduce neonatal mortality, respiratory morbidity, and other adverse birth outcomes. Yet, use of this proven treatment remains sub-optimal in Canadian clinical practice, with as many as 40% of eligible preterm births at 24-33 weeks’ gestation not receiving antenatal corticosteroids as recommended in Canadian and international practice guidelines.

One important factor contributing to low antenatal corticosteroids use is the lack of rigorous data on their longer-term safety for child health. There is no clear evidence of adverse effects, but signals of harm from selected animal and human studies have created concerns that antenatal corticosteroids may impair child neurodevelopment. However, all of these studies are low quality. Follow-up studies of randomized trials were hampered by low follow-up rates and small sample sizes, and observational studies were highly susceptible to confounding. High-quality evidence on the longer-term safety of antenatal corticosteroids for child neurodevelopment is critical for optimizing use of this medication.

When randomization is infeasible, regression discontinuity studies can provide evidence on the causal effects of a drug or exposure. The design exploits differences in clinical care among individuals immediately on either side of a clinical or policy cut-point as a pseudo-randomization tool. With antenatal corticosteroids, clinical practice guidelines recommend administration up to 33+6 weeks’ gestation (33 weeks, 6 days) – but not one day later, at 34+0 weeks. This (somewhat) arbitrary cut-off means that infants born as little as hours apart have substantially different chances of receiving this effective treatment. Using preliminary data from over 8500 preterm births in British Columbia, we have shown that regression discontinuity analysis can replicate the clinical trial findings on short-term neonatal respiratory morbidity. We propose to now link our cohort with child development outcomes, and use this novel design to generate rigorous evidence on the longer-term safety of antenatal steroids.

OBJECTIVE: To determine if antenatal corticosteroid administration decreases kindergarten child development scores (primary outcome) or increases risks of adverse neurodevelopmental outcomes such cerebral palsy or attention deficit hyperactivity disorder (ADHD).

APPROACH: We will use a population-based cohort of 12,653 preterm births from British Columbia, Canada, 2000-2013. Obstetrical and neonatal medical records contained in the BC Perinatal Data Registry will be linked with provincial school child development assessments, hospitalization, prescription, and physician visit data. Our primary outcome will be child development scores based on the Early Development Instrument, a validated tool routinely administered by kindergarten teachers across Canada and internationally. International classification of diseases codes and prescriptions will be used to identify our secondary outcomes of cerebral palsy and ADHD. We will use a regression discontinuity design to estimate the effect of antenatal corticosteroids on outcomes among infants born on either side of the 34+0 weeks’ cut-point within a narrow gestational age window.

PUBLIC HEALTH IMPORTANCE: Preterm birth is the single most important cause of neonatal mortality and severe morbidity in Canada, and costs the Canadian health care system over $8 billion per year. Antenatal corticosteroids can reduce this burden of disease by 30% for as low as $1 per dose, but the unknown potential for longer-term harm hinders their use in clinical practice. Our findings will inform evidence-based Canadian clinical practice guidelines that address the potential longer-term harms associated with antenatal corticosteroids. This will enable physicians to optimize use of this treatment with proven neonatal benefits that may currently be being avoided unnecessarily.

Funding: Canadian Institutes for Health Research