The relationship of maternal gestational mass spectrometry-derived metabolites with offspring congenital heart disease: results from multivariable and Mendelian randomization analyses

Abstract

AbstractBackgroundIt is plausible that maternal pregnancy metabolism influences risk of offspring congenital heart disease (CHD). We sought to explore this through a systematic approach using different methods and data.MethodsWe undertook multivariable logistic regression of the odds of CHD for 923 Mass Spectrometry (MS)-derived metabolites in a sub-sample of a UK birth cohort (Born in Bradford (BiB); N = 2,605, 46 CHD cases). We considered metabolites reaching a p-value threshold <0.05 to be suggestively associated with CHD. We sought validation of our findings, by repeating the multivariable regression analysis within the BiB cohort for any metabolite that was measured by nuclear magnetic resonance (NMR) or clinical chemistry (N = 7,296, 87 CHD cases), and by using genetic risk scores (GRS: weighted genetic risk scores of single nucleotide polymorphisms (SNPs) that were associated with each metabolite) in Mendelian randomization (MR) analyses. MR analyses were performed in BiB and two additional European birth cohorts (N = 38,662, 319 CHD cases).ResultsIn the main multivariable analyses, we identified 44 metabolites suggestively associated with CHD, including those from the following super pathways: amino acids, lipids, co-factors and vitamins, xenobiotics, nucleotides, energy, and several unknown molecules. Of these 44, isoleucine and leucine were available in the larger BiB cohort (NMR), and for these the results were validated. MR analyses were possible for 27/44 metabolites and for 11 there was consistency with multivariable regression results.ConclusionsIn summary, we have used complimentary data sources and statistical techniques to construct layers of evidence. We found that amino acid metabolism during pregnancy, several lipids (more specifically androgenic steroids), and levels of succinylcarnitine could be important contributing factors for CHD.