A Weighted Genetic Risk Score Using Known Susceptibility Variants to Predict Graves Disease Risk

CONTEXT: Graves disease (GD) is a common thyroid-specific autoimmune disease and one of the most heritable diseases in the population. We present a risk-prediction model, including confirmed, known genetic variants associated with GD. DESIGN: To construct a stable-prediction model, we used known GD susceptibility single nucleotide polymorphisms (SNPs) as markers and trained and tested our model in a cohort of 4897 patients with GD and 5098 healthy controls. We weighted the contribution of each SNP to the disease to calculate the weighted genetic risk score (wGRS) for each individual. The efficiency of this model can be estimated by the area under the curve (AUC) receiver operator characteristic curve and the specificity and sensitivity of each wGRS. RESULTS: With the 20 confirmed GD risk-related SNPs, our wGRS-prediction model could predict patients with GD from the general population (AUC 0.70 [95% CI: 0.69 to 0.71]) and did especially well in predicting patients with GD with persisting thyroid-stimulating hormone receptor antibody positive [pTRAb+; AUC 0.74 (95% CI: 0.72 to 0.76)]. We also evaluated how the four pTRAb+ specific risk SNPs predicted patients with GD with pTRAb+ among all patients with GD [AUC 0.62 (95% CI: 0.61 to 0.63)]. For clinical use, we partitioned subjects in each set into different risk categories to generate the wGRS cutoff of high risk for reference. CONCLUSIONS: Our study provides an approach to predict GD risk in the general population by the calculation of the wGRS of 20 known GD susceptibility variants. The wGRS-prediction model was more stable and convenient, whereas the prediction performance was still modest.