Abstract/Summary

Objective We aimed to test the hypothesis that early diet programmes the metabolic profile of young adults born preterm. Design We analysed banked urine samples obtained at a 20-year follow-up visit from adults that had participated as neonates in controlled trials involving randomisation within 48 hours of birth to feeds of preterm formula (PTF), banked breast milk (BBM) or term formula (TF) for 1month postnatally. Main outcome measures We performed proton nuclear magnetic resonance spectroscopy, analysing spectra by dietary group and sex. Orthogonal projections to latent structure discriminant analyses was used to model class differences and identify metabolites contributing to the differences between groups. Additionally, spectra were correlated with birth weight, gestational age and weight z score at 2weeks of age. Results Of the original number of 926 trial participants, urine samples were available from 197 (21%) healthy young adults (42% men) born preterm (mean 30.7±2.8 weeks) and randomised to BBM (n=55; 28 men), TF (n=48; 14 men) and PTF (n=93; 40 men). We found no significant differences in urinary spectra between dietary groups including when stratified by sex. Correlation analysis revealed a weak association between metabolic profile and gestational age that was lost on controlling for ethanol excretion. Conclusions We found no evidence that dietary exposures in the neonatal period influence the metabolic phenotype in young adult life.