Authors
Neil J Goulding, Katherine A Birchenall, Jane West, John Wright, Deborah A Lawlor.
Institution(s)
1 MRC Integrative Epidemiology Unit, University of Bristol (UoB); 2 PHS, UoB; 3 Dept of Obstetrics and Gynaecology, St Michael’s Hospital, Bristol; 4 THS, UoB; 5 Bradford Institute for Health Research, Bradford Teaching Hospitals NIHS Foundation Trust; 6 Bristol NIHR Biomedical Research Centre
Introduction
For most metabolites, it is not known whether or how they cross the placenta. We are not aware of research exploring associations of multiple maternal metabolites with the same measures in cord blood. Our hypothesis is that examining these associations will provide some insights into how metabolites cross the placenta.
Methods
Data from the Born in Bradford cohort were used.1 6614 maternal gestational (24-28 weeks) and matching cord-blood samples were analysed for 84 NMR metabolite levels and ratios. For 12 measures there was a high proportion of cord-blood samples with levels below the limit of quantification (LOQ). Pearson correlations between mothers and cord-blood metabolites were calculated with (i) replacing with LOQ (N = 5726 to 6614) and (ii) complete case analyses (N = 3066 to 6611). We also stratified on ethnicity1 and sex.
Funded by MRC, ESRC, Wellcome, BHF, NIHR, NIH, ERC. Ethics approval form Bradford NREC.
Results
There was no/very weak correlation (-0.03 to 0.20) between metabolite levels of mothers and cord-bloods for 79 (94%) of the NMR metabolites analysed in the whole sample. This included the correlation of maternal to cord glucose (R= 0.04). Creatinine, valine, glutamine, histidine and % Omega-3 fatty acids showed weak-moderate positive correlation (0.20 to 0.34). Creatinine (a measure of renal function) was the only metabolite to show noticeable differences in correlations for samples segregated by sex and/or ethnicity, ranging from 0.10 for White British boys to 0.34 for South Asian boys. A large proportion of cord-bloods (22% – 54%) were < LOQ for total lipids and particle concentration for very large and extremely large VLDL, small, medium and large LDL and large HDL. Repeating analyses on complete data treating these as missing (method (ii)) produced extremely similar results.
Conclusions
In general, there is very low correlation between NMR metabolite levels of mothers and cord-bloods. There is evidence of a moderate positive correlation for creatinine and some amino acids.
References
1 Taylor K et al. Differences in pregnancy metabolic profiles and their determinants between White European and South Asian women: Findings from the Born in Bradford Cohort. Metabolites 2019; 9: 190