105: Metabolic profiles at birth identify first 7 day mortality risk among extremely premature infants

Abstract

OBJECTIVE: To evaluate the ability of early metabolic profiles combined with infant characteristics to predict mortality ≤7 days of life (DOL) in infants with gestational age ≤28 weeks. STUDY DESIGN: This study included 3,508 singleton live births in California between 2005 and 2011 with gestations from 22 to 28 weeks. All infants had newborn metabolic screening data available. Data included linked birth certificate and mother and infant hospital discharge records. Mortality was derived from linked death certificates and death discharge information. Each mortality was matched to four controls by gestational age week and birthweight Z-score category, which left a final sample of 143 cases and 572 controls. The association between predictors and mortality ≤7 DOL was modeled using stepwise logistic regression. Model performance was assessed using area under the curve (AUC) and Akaike information criterion (AIC). Model fit was examined through cross validation. Variables considered in the model included infant characteristics, 42 natural log transformed metabolites and 15 metabolite ratios. Model parameter estimates, odds ratios (OR) with 95% confidence intervals (CI), and p-values were calculated. RESULTS: Within our sample, 150 (4.3%) infants died within 7 days of life. The final multivariable model included 3 infant characteristics, 8 metabolites, and 1 metabolite ratio (Table). This model was strongly predictive of infant mortality (AUC 0.848, 95% CI 0.813- 0.883), and was significantly more predictive than a model with just the selected metabolites (AUC 0.830, 95% CI 0.792-0.867) or a model with just characteristics (AUC 0.703, 95% CI 0.655-0.751) (Figure). The full model remained strongly predictive after cross validation (AUC 0.828, 95% CI 0.790-0.865). CONCLUSION: Extremely premature infants likely to survive are metabolically different from those with a high risk of mortality within the first 7 DOL. Physicians can use metabolic profiles measured at birth to refine mortality risks and inform postnatal counseling for extremely premature infants.(Figure Presented).