In the ALPS study, strict inclusion and exclusion criteria were utilized with the goal of identifying women who were likely to deliver a late preterm infant at least 24 hours after but within 7 days of betamethasone administration (but before term gestation). Despite this, our study shows administration was suboptimal in 50% of women who received AS, demonstrating the inherent difficulty of timing and predicting preterm delivery. Our study shows that mothers in the AS group delivered at an earlier mean GA compared to the no AS group. Considering that the median latency between the first dose to time of delivery was 0.78 days with IQ range of 0–6.36 days in the group who received AS, the lower mean GA at delivery within the AS group could reflect strict adherence to ALPS/SMFM criteria(1, 3), or a bias by providers towards not offering or administering AS the closer a pregnancy is to 37 weeks gestation.
In the study by McElwee, et al, the highest rate of inappropriate exposure was for PPROM without cervical change(10), which is in contrast to our data where the No AS group had a higher incidence of PPROM compared to those who received AS. Pregestational diabetes was an exclusion criteria for the ALPS study and SMFM recommendations, and our cohort expectedly showed a significantly larger proportion of diabetes in the No AS group compared to those given AS. The fact that some pregestational diabetics received AS however, indicates that the strict criteria is not being applied by all providers, or that AS was given under the protocol for those at risk of delivering prior to 34 weeks GA.
In contrast to the ALPS trial, the infants exposed to AS in our study were significantly smaller compared to the unexposed group, and the difference in birth weight remained significant even after adjusting for GA, maternal smoking status, maternal pre-eclampsia, and maternal diabetes (p < 0.001). However, we found no difference between the groups for those classified as SGA or LGA, although we were not powered for these outcomes.
We found no significant differences in short term neonatal respiratory outcomes, NICU admission, or overall hospital length of stay in the cohort exposed to AS even as we controlled for GA, but we are not powered to detect such difference in clinical outcomes. However, AS have been shown to increase passive respiratory compliance (Crs) and functional residual capacity (FRC) after birth in preterm infants exposed to AS within 7 days prior to delivery(12–15). We have demonstrated that in a single center cohort of 50 LPIs born between 34 0/7 to 34 6/7 weeks’ GA, a significant increase in Crs was noted for those treated with a single course of AS within 7 days of delivery versus matched infants who did not receive AS ( 1.33 vs 1.06 mL/cmH2O/kg, p = 0.018)(16). The infants treated with AS also had a higher functional residual capacity, and significantly less of the AS treated infants required continuous positive airway pressure treatment (16% vs 68%; p = 0.007)(16).
A specific harm associated with late preterm AS is neonatal hypoglycemia, which was identified as a significant outcome in the ALPS trial, occurring in 24% of those in the betamethasone arm (vs 15.0% in placebo with a relative risk of 1.60; 95% CI, 1.37 to 1.87; P < 0.001).(3) Although the definition remains controversial, transient and persistent neonatal hypoglycemia has been associated with negative long-term neurodevelopmental outcomes in preterm and term infants(17–23). Our study showed no difference in the incidence of hypoglycemia between the two groups using the same definition as the ALPS trial, which could be related to stricter adherence at our center to the exclusion criteria of any maternal diabetes. Only 4.5% of our AS group had pregestational diabetes, which was similar to the findings of the cohort by McElwee, et al, where only 3% of inappropriate AS exposures were in the setting of pregestational diabetes.(10) This is in contrast to the findings of Battarbee, et al, whose survey showed that greater than 50% of the providers reported administering betamethasone in the setting of poorly controlled diabetes or diabetes on insulin.(9)
Our study has several limitations. Due to the retrospective and non-randomized nature of our study, it was difficult to ascertain the precise indications or contraindications for AS administration unless it was documented by the provider. It was also difficult to determine other clinical variables that may have contributed to the decision for AS administration. Another limitation is that we did not examine term deliveries exposed to AS. In the ALPS study, 16.4% of patients went on to deliver at term, and this may be higher in a clinical setting leading to higher rates of non-optimal administration. This is concerning for a large body of obstetric and neonatal providers, given that long-term outcomes for AS exposure in LPIs remain unknown. Further, a recent population-based retrospective cohort study in Finland showed that AS exposure was significantly associated with higher risk of any mental and behavioral disorder in the entire cohort of term-born children (12.01% vs 6.45%; absolute difference 5.56% [95% CI, 5.04%-6.19%]; adjusted hazard ratio 1.33 [95%CI, 1.26–1.41])(7). In this study with 14,868 AS-exposed children, 6730 (45.27%) were born at term. These would be important populations and clinical outcomes for future prospective longitudinal studies to measure as risk and benefits are considered for the administration of corticosteroids during the period of late preterm gestation.