Effects of antenatal indomethacin on ductus arteriosus early closure and on adverse outcomes in preterm neonates

Introduction: Premature birth is a leading cause of neonatal morbidity and mortality. Since gestational age at birth is the most important predictive factor of adverse neonatal outcomes, strategies to postpone premature labor are of major importance. Studies on tocolytic drugs show that COX inhibitors such as indomethacin are superior to others in terms of efficiency in delaying birth, but results concerning neonatal outcomes associated with prenatal exposure to these drugs show controversial results. Indomethacin is also used in the postnatal age for pharmacologic treatment of patent ductus arteriosus (PDA), but no data concerning the effects of antenatal exposure on postnatal ductal patency are available.Methods: In this study, we focused primarily on the association between antenatal indomethacin (AI) and postnatal patency of ductus arteriosus while our secondary aim was to highlight any possible influence of AI exposure on adverse neonatal outcomes. We performed a retrospective analysis of 241 medical records of newborns born before 33 weeks' gestation and exposed to antenatal tocolysis. Obstetrical data and neonatal outcomes of newborns exposed to AI were compared to those of neonates exposed to other tocolytic drugs. Early ductal closure (EDC) was defined when functional echocardiography performed within 24 hours of life showed a closed duct. Occurrence of intraventricular hemorrhage (IVH), periventricular leukomalacia (PVL), respiratory distress syndrome (RDS), chronic lung disease (CLD), necrotizing enterocolitis (NEC), sepsis, and PDA were compared between the groups and the diagnosis of at least one of III-IV grade IVH, PVL, CLD, sepsis, surgical NEC, or death was defined as a severe outcome.Results: The univariate analysis showed that infants in the AI group were at a higher risk of IVH, CLD, RDS, sepsis, and PDA. The incidence of severe outcomes also appeared to be higher in this group, while no effect of AI on PDA was observed. Since we noticed that infants exposed to AI had a lower gestational age and worse clinical conditions at birth when compared to the controls, we considered this as a confounding factor. To overcome this bias, we performed a multivariate analysis that evidenced no significant role of AI on the occurrence of severe outcomes. On the other hand, a possible association was confirmed for all degrees of IVH (OR: 3.16, 95% CI : [1.41; 7.05]) and sepsis (OR: 2.81, 95% CI: [1.24; 6,28]).Conclusions: The unexpected result shown by the multivariate analysis was the association between AI exposure and EDC (OR: 2.52, 95% CI: [1.02; 6.21]). This result, which has never been evidenced in previous studies, has great clinical importance. It is well known that PDA is more frequent at lower gestational ages, thus reducing the incidence of PDA could lead to an improvement of overall outcomes in extremely preterm newborns.