A Maternal Serum Metabolite Ratio Predicts Large for Gestational Age Infants at Term: A Prospective Cohort Study

Context Excessive birth weight is associated with maternal and neonatal complications. However, ultrasonically estimated large for gestational age (LGA; >90th percentile) predicts these complications poorly. Objective To determine whether a maternal serum metabolite ratio developed for fetal growth restriction (FGR) is predictive of birth weight across the whole range, including LGA at birth. Methods Metabolites were measured using ultrahigh performance liquid chromatography-tandem mass spectroscopy. The 4-metabolite ratio was previously derived from an analysis of FGR cases and a random subcohort from the Pregnancy Outcome Prediction study. Here, we evaluated its relationship at 36 weeks of gestational age (wkGA) with birth weight in the subcohort (n = 281). External validation in the Born in Bradford (BiB) study (n = 2366) used the metabolite ratio at 24 to 28 wkGA. Results The inverse of the metabolite ratio at 36 wkGA predicted LGA at term [the area under the receiver operating characteristic curve (AUROCC) = 0.82, 95% CI 0.73 to 0.91, P = 6.7 x 10(-5)]. The ratio was also inversely associated with birth weight z score (linear regression, beta = -0.29 SD, P = 2.1 x 10(-8)). Analysis in the BiB cohort confirmed that the ratio at 24 to 28 wkGA predicted LGA (AUROCC = 0.60, 95% CI 0.54 to 0.67, P = 8.6 x 10(-5)) and was inversely associated with birth weight z score (beta = -0.12 SD, P = 1.3 x 10(-9)). Conclusions A metabolite ratio which is strongly predictive of FGR is equally predictive of LGA birth weight and is inversely associated with birth weight across the whole range.