Pulmonary hypertension in congenital diaphragmatic hernia: Antenatal prediction and impact on neonatal mortality.

OBJECTIVE: To determine the prevalence of pulmonary hypertension (PAH) in left-sided congenital diaphragmatic hernia (CDH); how we could predict it; and how PAH contributed to the model for mortality prediction. STUDY DESIGN: Retrospective analysis in three European centers. The primary outcome was the presence of PAH on postnatal day (d) 1, 7, and at discharge. Studied predictors of PAH were: observed/expected-lung/head-ratio (o/e LHR), liver-herniation, fetoscopic endoluminal tracheal occlusion (FETO), and gestational age (GA) at delivery. The combined effect of pre- and postnatal variables on mortality was modeled by Cox regression. RESULTS: Of the 197 neonates, 56 (28.4%) died. At d1, 67.5% (133/197) had PAH and 61.9% (101/163) by d7. Overall, 6.4% (9/141) had PAH at discharge. At d1, o/e LHR (odds ratio (OR) 0.96) and FETO (OR 2.99) independently correlated to PAH (areas under the curve [AUC]: 0.74). At d7, PAH significantly correlated only with the use of FETO (OR 3.9; AUC: 0.65). None were significant for PAH at discharge. Combining the occurrence of PAH with antenatal biomarkers improved mortality prediction (p = 0.02), in a model including o/e LHR (HR: 0.94), FETO (HR: 0.35), liver herniation (HR: 16.78), and PAH (HR: 15.95). CONCLUSIONS: Antenatal prediction of PAH was only moderate. The postnatal occurrence of PAH further increases the risk of death. Whereas this may be used to counsel parents in the postnatal period, our study demonstrates there is a need to find more accurate antenatal predictors for PAH.