Introduction:

Antenatal betamethasone reduces neonatal respiratory morbidity; but the benefit of one vs two doses of betamethasone remains in question.¹ Furthermore, little research has been done to characterize the interval dose-exposure for betamethasone and delivery. The goal of this study is to characterize the effect of second betamethasone dose on neonatal outcomes as time-dependent with distinct dose-timing phenotypes.

 

Methods:

This was a retrospective cohort study on pregnancies at the Mount Sinai Hospital receiving antenatal betamethasone. Patients were included based on chart review, betamethasone administration before delivery (one or two doses), and preterm birth from January 1, 2013 through December 31, 2023 who received 2 doses of betamethasone prior to delivery and delivered at < 37 weeks. Patients were excluded if they received > 2 doses, did not deliver at our institution, received betamethasone at > 37 weeks, or experienced an intrauterine fetal demise. After data extraction, we also eliminated those who received betamethasone over 14 days prior to delivery Patients were excluded if they did not deliver at MSH, were not in the gestational interval, or delivered after 2 weeks of the last betamethasone exposure.  Primary outcome was defined as a composite of neonatal respiratory morbidity including (transient tachypnea of newborn, respiratory distress syndrome, CPAP>12hrs, supplemental oxygen>24hrs, intubation, surfactant). Secondary outcome was defined as a composite severe neonatal morbidity (intraventricular hemorrhage, periventricular leukomalacia, retinopathy of prematurity, necrotizing enterocolitis, bronchopulmonary dysplasia, sepsis).

Exposure was defined as time from the most recent betamethasone dose (one or two) to delivery. Logistic regression with restricted cubic spline modeled nonlinear exposure-response relationships including dose and timing – adjusting for gestational age, birthweight, neonatal sex, premature labor, premature rupture of membranes, and chorioamnionitis. Marginal standard risk and risk differences (RDs) comparing one vs two doses at clinically meaningful timepoints (12h, 24h, 48h, 72h, and >72h) were assessed. Uncertainty was assessed with bootstrap resampling.

 

Results:

A total of 2,086 mothers and neonates were analyzed, 873 received one dose and 1,266 received two doses of betamethasone. Of the mothers who received two doses, 53 of them received betamethasone <= 2hours before delivery.

For the primary outcome, inclusion of dose-timing interaction significantly improved model fit (p<0.0001). The effect of a second dose varied nonlinearly by exposure timing, captured by the cubic spline. Compared to one dose, two doses were associated with lower respiratory morbidity when delivery occurred shortly after exposure (RD_12h = -11.3%, RD_24h = -7.3%, RD_48h = +1.1%, RD_72h = +6.6% and RD_>72h +10.2%, p < 0.05).  For secondary severe neonatal morbidity outcome, no significant dose-timing interaction was observed, and risk differences were small across all exposure times (p = 0.68).

Upon gestational age (GA) subgroup analysis, no significant primary outcome difference was observed until 32 weeks with no difference after 36 weeks. Thus, a 32-36 week sub cohort of 1,493 deliveries (722 received one dose, 771 received two doses of betamethasone) was created and analyzed, demonstrating the two-dose group having significant lower risk of the primary composite outcome across all refit-bootstrap estimates. Absolute risk reduction approximately 14-16% lower than the one dose group from 12 to > 72 hours (p<0.001). However, no significant association was observed between dose number and secondary composite between one and two doses (ARR < 0.6% across all times).

 

Discussion:

This is one of the largest studies demonstrating the benefit of betamethasone as a dose-timing phenotype. Globally across this patient population, a second dose confers benefit only if delivery occurs soon after exposure, with diminishing and potentially adverse associations as intervals increase beyond 48 hours. However, when selecting for the 32-36 week population based on the landmark analysis, two doses of betamethasone were associated with a clinically and statistically significant reduction in the primary composite outcome, but not the secondary across all time points. Although the initial analysis suggested a dose-dependent response, with the 32-36 week subgroup, the observed association is not simply a function of longer time exposure persisting across the flexible spline, but does suggest additional pulmonary maturation benefits in this population. This could possibly be explained by late-preterm fetuses having sufficient baseline lung maturity that partial corticosteroid exposure (1 dose) is necessary but insufficient to achieve maximal benefit, whereas a two-dose regimen may provide additional pulmonary maturity. Conversely more mature neonates approaching term have marginal benefits. Very early neonates globally have high respiratory morbidity despite full dosing.