Assessing the clinical utility of a genetic risk score associated with type 2 diabetes in a southern European population. (14th October 2021)

• Record Type: Journal Article
• Authors: Santos, M; Mendonca, M I; Temtem, M; Sousa, J A; Mendonca, F; Sousa, A C; Freitas, S; Henriques, E; Rodrigues, M; Borges, S; Guerra, G; Drumond, A; Palma Dos Reis, R
• Is Part Of: European heart journal. Volume 42(2021)Supplement 1
• Publication Date: 2021-10-14
• Subjects: Scores; Cardiology -- Periodicals; Heart -- Diseases -- Periodicals; 616.12005
• DOI: 10.1093/eurheartj/ehab724.2486
• Languages: English
• ISSNs: 0195-668X

Abstract

Abstract: Background: The development of personalized susceptibility profiles based on genetic information to aid prediction, early detection and prevention of type 2 diabetes (T2D) with potential clinical application, begins to awaken interest in the scientific community. However, its clinical translation is controversial. Objective: Evaluate the clinical utility of a genetic risk score (GRS) created with the GWAS-derived genetic variants associated to T2D to predict and discriminate the susceptibility to Type 2 diabetes, in a Southern European population with and without T2D. Methods and results: We studied through a case-control with 3, 139 subjects (772 with T2D and 2, 367 without) the usefulness of implementing a GRS in clinical practice. We constructed a multiplicative GRS (mGRS) calculated using 10 SNPs of genetic loci robustly associated to T2D (HNF4A rs1884613, IGF2BP2 rs4402960, PPARG rs1801282, TCF7L2 rs7903146, SLC30A8 rs1326634, MC4R rs17782313, ADIPOQ rs266729, FTO rs8050136, TAS2R50 rs1376251 and APO E rs7412 and rs429358), to evaluate the prediction and discrimination of T2D. Two logistic regression models were performed the first with age, sex and BMI. The second with these three risk factors plus hypertension, LDL >130mg/dl and physical inactivity. Logistic regression models, receiver operating characteristic analyses (ROC curve) were used. Each model was analysed individually and added with mGRS to calculate the area under the ROC curve (AUC), which may be considered a global estimate of each model's predictive power. The inclusion of GRS in the first model increased the discriminative power of T2D (AUC=0.669 to 0.692; p<0.0001. In the second model, the increase was AUC=0.712 to 0.729; p<0.0001. Conclusions: Adding genomic information to traditional models improves the ability to predict and discriminate type 2 Diabetes slightly, compared to traditional models alone. Nevertheless, this increase is not sufficiently robust for translation in clinical practice. However, clinicians should be conscious that T2D genetic research is experiencing a dramatic revolution and stay optimistic that these innovative studies translate into improved care for diabetic patients. FUNDunding Acknowledgement: Type of funding sources: None.