Subtyping of Type 2 Diabetes from a large Middle Eastern Biobank: Implications for Precision Medicine

BackgroundType 2 diabetes (T2D) can be classified into Severe Insulin-Deficient Diabetes (SIDD), Severe Insulin-Resistant Diabetes (SIRD), Mild Obesity-related Diabetes (MOD), and Mild Age-related Diabetes (MARD). This classification predicts disease complications and determines the best treatment for individuals. However, the classifications applicability to non-European populations and sensitivity to confounding factors remain unclear. MethodsWe applied k-means clustering to a large Middle Eastern biobank cohort (Qatar Biobank; QBB, comprising 13,808 individuals; 2,687 with T2D). We evaluated the efficacy of the European cluster coordinates and analyzed the impact of using actual age on clustering outcomes. We examined sex differences, analyzed insulin treatment frequency, investigated the clustering of maturity-onset diabetes of the young (MODY), and evaluated the incidence of chronic kidney disease (CKD) among T2D subtypes. ResultsWe identified the four T2D subtypes within a large Arab cohort. Data-derived centers outperformed European coordinates in classifying T2D. The use of actual age, as opposed to age of diagnosis, impacted MOD and MARD classification. Obesity prevalence was significantly higher in females, however, that did not translate to worse disease severity, as indicated by comparable levels of HbA1C and HOMA2-IR. Insulin was predominantly prescribed for individuals in SIDD and SIRD, which also displayed the highest risk of CKD, followed by MOD. Interestingly, most MODY individuals were clustered within MARD, further highlighting the need for precise classification and tailored interventions. ConclusionThe observed sex differences underscore the importance of tailoring treatment plans for females compared to males. For SIDD and SIRD individuals, who are at a higher risk of CKD, insulin therapy requires closer monitoring and physician oversight. Additionally, in populations without access to genetic testing, likely MODY individuals can be identified within the MARD cluster. These findings strongly support the need for a transition to more personalized, data-driven treatment approaches to minimize diabetes-related complications and improve patient outcomes.