453 - Longitudinal Metabolic Profiling Predicts Neonatal Morbidity and Mortality in Very Preterm Neonates <32 Weeks Gestation

Poster Number: 453
Publication Number: 453.1883

Background: The very preterm neonate ( < 32 weeks gestation) is at the highest risk of developing complications that can result in death or significant life-long disability. Currently, the ability to identify newborns within the first week of life who are at high risk for mortality or developing a significant neonatal morbidity is limited. Our team has previously shown that dried blood spot metabolites measured at 24-72 hours after birth by routine newborn screening (NBS) are novel predictors of major neonatal morbidity and mortality in neonates born < 32 weeks gestation.

Objective: The goal of this study was to examine longitudinal metabolic profiles at five time points across the first week of life for prediction of mortality and major neonatal morbidity.

Design/Methods: Neonates born < 32 weeks gestation were enrolled into a prospective cohort at two clinical sites. A composite outcome of mortality before discharge and major neonatal morbidity including necrotizing enterocolitis (NEC), intraventricular hemorrhage (IVH) grades 3 or 4, and retinopathy of prematurity (ROP) stage 3 or higher was captured from the electronic medical record. Forty-nine metabolites were measured using standard protocols for NBS on prospectively collected samples at five time points throughout the first week: cord blood and blood at day of life 1, 3, 5, and 7. Functional principal components analysis was performed on the longitudinal logarithm transformed metabolite values to identify patterns associated with the composite outcome.

Results: A total of 188 neonates, 27 with the composite outcome, were included in the analysis. Ten metabolites were significantly longitudinally associated with the composite outcome: thyroid stimulating hormone, immunoreactive trypsinogen, alanine, argininosuccinic acid, leucine, methionine, and acylcarnitines C4-DC, C10:1, C12:1, and C16-OH. The discrimination (i.e., Area under the curve (AUC)) for the model including only metabolites was 0.92 (95% CI: 0.87, 0.97) compared to an AUC of 0.80 (95% CI: 0.71, 0.88) when examining a model with birth weight, sex, and gestational age. The AUC for longitudinal metabolites combined with clinical characteristics was 0.93 (95% CI: 0.87, 0.99).

Conclusion(s): The longitudinal metabolite models for predicting a composite measure of mortality and major morbidity performed comparable to models including a single metabolite value at 24-72 hours after birth. However, prospective monitoring of metabolites over the course of the first week of life may reveal nutritional targets for treatment and prevention of mortality and neonatal morbidity in very preterm neonates.