1993-LB: Subgroups of Type 2 Diabetes (T2D) by Clinical Features and Islet Autoimmunity

Introduction: Adults with T2D may have phenotypic subgroups; prior clustering analyses had limited measures of beta-cell function, islet autoimmunity, or both. We aimed to characterize T2D subgroups in Look AHEAD (Action for Health in Diabetes), a randomized trial of adults with clinically-diagnosed T2D and overweight/obesity. Methods: In 3,952 participants, we measured islet autoantibodies (GADA, ZnT8A) and performed homeostatic model assessment v2 of beta cell function (HOMA2-B) and insulin resistance (HOMA2-IR) using fasting C-peptide. We applied k-means clustering using these measures plus age at T2D diagnosis, BMI, and HbA1c. Results: We identified four subgroups (Figure): lower adiposity with good glycemic control (LA/GC, 42% of the sample), high adiposity with good glycemic control (HA/GC, 26%), poor glycemic control (PC, 30%) and islet autoimmunity (IA, 3%). The LA/GC subgroup had high HOMA2-B (median 73%) and low HOMA2-IR (median 2.4). HA/GC had the highest HOMA2-B (median 113%) and HOMA2-IR (median 3.8). The PC subgroup had low HOMA2-B (median 46%) and youngest T2D diagnosis. Only the IA subgroup had frequent islet autoimmunity (95% GADA+, 18% ZnT8A+) and had low HOMA2-B (median 55%). Insulin use was infrequent in LA/GC (9%) and HA/GC (4%), and common in PC (32%) and IA (41%). Conclusions: Using clinically-available measures, we identified four distinct T2D subgroups with important physiologic differences. Disclosure S.J. Pilla: None. W.C. Knowler: None. A. Balasubramanyam: None. C.S. Hampe: None. S. Pietropaolo: None. D. Lin: None. A.M. Anderson: None. N. Zhao: None. Å. Lernmark: Advisory Panel; DiaMyd Medical AB, Stockholm, Sweden. M. Bancks: None. J. Clark: Advisory Panel; Boehringer-Ingelheim. N.M. Maruthur: Other Relationship; Johns Hopkins Health Care Solutions. Funding National Institute of Diabetes and Digestive and Kidney Diseases (R01DK126825, K23DK128572)