The Liability Threshold Model for Predicting the Risk of Cardiovascular Disease in Patients with Type 2 Diabetes: A Multi-Cohort Study of Korean Adults

Abstract

Personalized risk prediction for diabetic cardiovascular disease (DCVD) is at the core of precision medicine in type 2 diabetes (T2D). We first identified three marker sets consisting of 15, 47, and 231 tagging single nucleotide polymorphisms (tSNPs) associated with DCVD using a linear mixed model in 2378 T2D patients obtained from four population-based Korean cohorts. Using the genetic variants with even modest effects on phenotypic variance, we observed improved risk stratification accuracy beyond traditional risk factors (AUC, 0.63 to 0.97). With a cutoff point of 0.21, the discrete genetic liability threshold model consisting of 231 SNPs (GLT231) correctly classified 87.7% of 2378 T2D patients as high or low risk of DCVD. For the same set of SNP markers, the GLT and polygenic risk score (PRS) models showed similar predictive performance, and we observed consistency between the GLT and PRS models in that the model based on a larger number of SNP markers showed much-improved predictability. In silico gene expression analysis, additional information was provided on the functional role of the genes identified in this study. In particular, HDAC4, CDKN2B, CELSR2, and MRAS appear to be major hubs in the functional gene network for DCVD. The proposed risk prediction approach based on the liability threshold model may help identify T2D patients at high CVD risk in East Asian populations with further external validations.

Genomic risk prediction of cardiovascular diseases among type 2 diabetes patients in the UK Biobank

“…Previously, the use of PRS to predict cardiovascular events among T2D patients was not clear. Hong et al (Hong et al, 2020) constructed PRS for CVD on 2,378 T2D patients and explored its prediction performance in the same dataset. One great limitation of this study was the lack of an independent validation dataset for PRS performance analysis, while in our study, we built the subtype PRSs using large-scale GWASs and tested the performance through cross-validation, which was more appropriate.…”

“…With much larger studies and improved PRS methods available now, we set out to build a new PRS for CVD onset among T2D patients. Since several studies have shown that utilizing a meta-analytic strategy to build PRS can help better capture the genetic risk information ( Inouye et al, 2018 ); also considering that CAD, IS, and HF are the major subtypes of CVD events with similar clinical implications and management ( Hong et al, 2020 ; Ma et al, 2022 ) and there are strong genetic correlations among these three subtypes to potentially boost power of PRS ( Dichgans et al, 2014 ; Verweij et al, 2017 ; Hong et al, 2020 ; Koyama et al, 2020 ), here we build a new PRS by combining the three “optimal” PRSs trained for each of these three CVD subtypes. With the newly-built meta-PRS CVD , we then comprehensively evaluate its prediction performance for CVD events among T2D patients.…”

“…Polygenic risk scores (PRSs), the sum of risk alleles weighted by the effect sizes inferred from GWAS summary statistics, were subsequently constructed and have shown promise in predicting the onset of CVD as well as its subtypes (Khera et al, 2018). A few attempts have then been made to associate the PRS trained from CVD events occurrence among T2D patients, but with some limitations (Hong et al, 2020). These initial efforts were mainly hampered by three challenges: 1) the small sample size (less than 5,000) to train and test PRS; 2) less optimal risk prediction models for creating PRS; and 3) the same dataset used to construct PRS and test for prediction performance without validations.…”

Genetic modifiers of rare variants in monogenic developmental disorder loci

“…The presence of such rare, putatively damaging variants in healthy population cohorts 3 can provide a lower boundary for estimates of penetrance, and individuals in both clinical and population cohorts display a spectrum of phenotypic variability caused by similar or identical variants 1,4 . Previous research has suggested that common genetic variants can modify the penetrance or expressivity of phenotypes caused by rare genetic variants [5][6][7] , potentially through the liability threshold model, which posits that a certain threshold of disease susceptibility needs to be crossed before clinically-diagnosable disease manifests [8][9][10][11] . Some damaging rare variants may reach this threshold alone, resulting in a monogenic disease phenotype with 100% penetrance, while other variants may need additional genetic, environmental, or other modifiers to reach this threshold 8 .…”

“…Previous research has suggested that common genetic variants can modify the penetrance or expressivity of phenotypes caused by rare genetic variants [5][6][7] , potentially through the liability threshold model, which posits that a certain threshold of disease susceptibility needs to be crossed before clinically-diagnosable disease manifests [8][9][10][11] . Some damaging rare variants may reach this threshold alone, resulting in a monogenic disease phenotype with 100% penetrance, while other variants may need additional genetic, environmental, or other modifiers to reach this threshold 8 . In certain diseases, common variant burden has been shown to confer a risk similar to that of a deleterious monogenic variant, where the highest polygenic risk may be equivalent to that conferred by a monogenic variant 12,13 .…”