526 - Does Extubation Mediate the Relationship Between Dexamethasone and Bronchopulmonary Dysplasia?

Publication Number: 3507.526

Megha Dasani, Yale School of Medicine, Paramus, NJ, United States; Samuel McCartney, Children's of Alabama, Irondale, AL, United States; Noa Fleiss, Yale School of Medicine, New Haven, CT 06510, CT, United States; Namasivayam Ambalavanan, University of Alabama School of Medicine, Birmingham, AL, United States; Waldemar Carlo, University of Alabama, Birmingham, AL, United States; Samuel Gentle, Yale School of Medicine, New Haven, CT, United States

Background: Dexamethasone is commonly administered to reduce the risk of bronchopulmonary dysplasia (BPD) or death; however, whether its benefit arises primarily through facilitating extubation and reducing ventilator dependence or through other anti-inflammatory mechanisms remains uncertain.

Objective: To evaluate whether early administration ( ≤21 days of life) versus late (>21 days of life) dexamethasone exposure reduces the composite outcome of grade 2-3 BPD or death and to distinguish indirect effects mediated by ventilator duration or extubation success from direct effects independent of respiratory parameters.

Design/Methods: Multicenter cohort study in preterm infants born at < 29 weeks' gestation without major congenital anomalies or genetic syndromes and concurrently exposed to both invasive respiratory support and postnatal dexamethasone. Comparison groups were early (≤21 days) and late (>21 days but < 32 weeks' postmenstrual age) dexamethasone exposure. The primary outcome was grade 2 or 3 BPD or death. As late dexamethasone may result from illness severity, propensity score matching (1:2 nearest-neighbor, caliper = 0.2) was used to match infants on the logit of the propensity score from gestational age, birth weight, and respiratory support on postnatal day 14. Two mediators were evaluated: (1) ventilator days after dexamethasone initiation, and (2) extubation success (defined as remaining extubated ≥72 hours) during dexamethasone exposure. The average causal mediation effect (ACME), average direct effect (ADE), and total effect were estimated using bootstrapping.

Results: Two hundred fifty-nine infants were available for analysis of whom 231 infants were successfully matched(Table 1). The median (interquartile range: IQR) for gestational age and birth weight were 24.3 (23-25w) and 620g (525-710g), respectively. Early dexamethasone exposure was associated with a reduction in the composite outcome of BPD or death (Total Effect: -0.13, 95% CI: -0.21 to -0.05, p = 0.002). Mediation analysis (Figure 1) through ventilator days showed a significant direct effect (ADE: -0.060, 95% CI: -0.108 to -0.003, SE = 0.025, p = 0.038) and borderline significant indirect effect (ACME: -0.073, 95% CI: -0.153 to 0.006, SE = 0.041, p = 0.078). In contrast, extubation success did not significantly mediate the outcome (ACME: -0.0013, 95% CI: -0.0054 to 0.0027, SE = 0.0021, p = 0.41), though a direct effect remained (ADE: -0.020, 95% CI: -0.040 to -0.0004, SE = 0.013, p = 0.046).

Conclusion(s): These findings suggest that dexamethasone's benefit is mediated by direct effects and not indirectly through ventilator liberation.

Table 1. Baseline Demographic, Clinical, and Respiratory Characteristics of Infants Receiving Early vs Late Dexamethasone
PDA: patent ductus arteriosus; IVH: intraventricular hemorrhage; BPD: bronchopulmonary dysplasia; FiO2: fraction of inspired oxygen; MAP: mean airway pressure. Values are presented as median(IQR) for continuous variables (Mann-Whitney U Test) and % for categorical variables (Chi-square). *P <0.05 considered statistically significant.

Figure 1. Mediation of the effect of early dexamethasone exposure on the composite outcome of bronchopulmonary dysplasia (BPD) or death.
Panel A illustrates mediation via ventilator days and Panel B by extubation success following dexamethasone exposure. Points represent effect estimates on the risk difference scale, with horizontal bars denoting 95% confidence intervals. Estimates shown include the indirect effect (ACME), the direct effect (ADE), and the total effect of early dexamethasone exposure on BPD or death. A vertical dashed line at 0 indicates the null effect.