Genetic risk score as a predictor of gestational diabetes in Central European Caucasians

Gestational diabetes mellitus (GDM) is a common complication of pregnancy associated with perinatal complications and increased future risk of T2DM. GDM has a polygenic background partly overlapping with T2DM, and using a genetic risk score (GRS) that captures complex genetics may improve early detection of individuals at risk. SNPs previously associated with T2DM in the Czech population (21 loci) have been examined in a study comprising Caucasian (i) pregnant women with GDM (N = 416), (ii) non-diabetic population controls (N = 1170) and (iii) T2DM patients (N = 359) to analyse (di)similarities of T2DM and GDM genetic architecture using both weighted (wGRS) and unweighted (uGRS) GRS constructed based on single locus analyses. Fourteen of 21 SNPs had no influence on the GDM development. The most significant SNPs associated with GDM were within the ARAP1 (rs1552224; OR, 95% CI 2.98, 2.18-4.08) and MNTR1B (rs10830963; OR, 95% CI 1.78, 1.41-2.24) genes. Both uGRS (P < 0.0001) and wGRS (P < 0.0001), comprising alleles of the seven most strongly associated variants, were associated with GDM, with wGRS being a better discriminator-AUC 0.669 vs. 0.628 (P < 0.001). Carriers of the population-derived Q5 of wGRS (in comparison with Q1) have a significantly increased risk of GDM with OR (95% CI) 6.8 (4.3-10.9); P < 1.10^-6. The same scores derived from the ARAP1, NOTCH2, and MTNR1B genes discriminate between the GDM and T2DM groups (P < 0.0001). Identification of statistically significant association of the population-specific wGRS with GDM susceptibility supports its potential clinical utility for the risk stratification of pregnant women (beyond established non-genetic risk factors). Yet, the actual health care adoption of GRS in the near future is unlikely, since additional validation, regulatory and administrative steps, thorough economic analyses, ethical discussion, etc. are certainly required.