To evaluate B-type natriuretic peptide (BNP) as a longitudinal biomarker of clinical outcome in infants with congenital diaphragmatic hernia (CDH).

We conducted a retrospective study of 49 infants with CDH, classifying the cohort by respiratory status at 56 days, based on a proposed definition of bronchopulmonary dysplasia for infants ≥32 weeks’ gestation: good outcome (alive with no respiratory support) and poor outcome (ongoing respiratory support or death). BNP levels were available at age 1-5 weeks. Longitudinal BNP trends were assessed using mixed-effects modeling. Receiver operating characteristic curves were generated to identify BNP cutoffs maximizing correct outcome classification at each time point. The time to reach BNP cutoff by outcome was assessed using Kaplan-Meier curves for weeks 3-5.

Twenty-nine infants (59%) had a poor outcome. Infants with a poor outcome were more likely than those with a good outcome to have liver herniated into the thorax (90% vs 50%; P = .002) and to undergo nonprimary repair (93% vs 35%; P < .001). Mixed-effects modeling demonstrated a differing decline in BNP over time by outcome group (P = .003 for interaction). BNP accurately predicted outcome at 3-5 weeks (area under the curve, 0.81-0.82). BNP cutoffs that maximized correct outcome classification decreased over time from 285 pg/mL at 3 weeks to 100 pg/mL at 4 weeks and 48 pg/mL at 5 weeks. Time to reach the cutoffs of 100 pg/mL and 48 pg/mL were longer in the poor outcome group (log-rank P = .006 and <.0001, respectively).

Elevated BNP accurately predicts poor outcome in infants with CDH at age 3-5 weeks, with declining cutoffs over 3-5 weeks of age.