The effect of protein levels of ABO on pregnancy related outcomes: a Mendelian randomization study

Abstract

Protein level of Histo-Blood Group ABO System Transferase (ABO) has been reported to be associated with cardiometabolic diseases. But its effect on pregnancy related outcomes was still unclear. Here we conducted a two-sample Mendelian randomization study to ascertain the putative causal roles of protein levels of ABO in pregnancy related outcomes. Cis-acting protein quantitative trait loci (pQTLs) robustly associated with protein level of ABO (P < 5x10-8) were used as instruments to proxy the ABO level (N = 35,559, data from deCODE), with two additional pQTL datasets from Fenland (N = 10,708) and INTERVAL (N = 3,301) used as validation exposures. Ten pregnancy related diseases and complications were selected as outcomes. We observed that a higher protein level of ABO showed a putative causal effect on venous complications and haemorrhoids in pregnancy (OR = 1.207, 95%CI = 1.107-1.316, colocalization probability = 91.3%), which was validated by using pQTLs from Fenland and INTERVAL. The Mendelian randomization results further showed effects of the ABO protein on gestational hypertension (OR = 0.974, 95%CI = 0.953-0.995), despite insignificance after multiple testing correction and little colocalization evidence. Sensitivity analyses, including proteome-wide Mendelian randomization of the cis-acting ABO pQTLs, showed little evidence of horizontal pleiotropy. Correctively, our study prioritised ABO as a putative causal protein for venous complications and haemorrhoids in pregnancy. Future epidemiology and clinical studies are needed to investigate whether ABO can be considered as a drug target to prevent adverse pregnancy outcomes.