1450-P: Profiles of Glycemic Control in People with Type 1 Diabetes Using CGM—A Clustering Approach in the SFDT1 Study

CGM-derived data are usually analyzed in silos, metric by metric, for diabetes care or therapeutic strategy. Simultaneous analysis of the various dimensions of glycemic control would allow a more comprehensive overview. We analyzed data of people with type 1 diabetes (PWT1D) in the SFDT1 cohort study. A K-means clustering model was developed with HbA1c, coefficient of variation (CV), time in range (TIR), time below 70 mg/dl (TBR), Gold score and Glycemic Risk Index (GRI). We included 618 participants (53% men, age 41±14 years) from 20 centres in France. The optimal model had 3 clusters. The “Euglycemia” cluster (n=280, 45%) was characterized by a high TIR (mean 69.3%); low TBR ​​(4.2%), TAR (time above 180 mg/dl: 26.5%), CV (35.7%), HbA1c (7.0%) and GRI (36.4) values. The “Hyperglycaemia" cluster (n=197, 32%) was characterized by high TAR (56.5%), GRI (71.8) and HbA1c (8.5%) and low TIR (40.3%), TBR (3.2%) and Gold score (2.3). The “Hypoglycemia” cluster (n=141, 23%) was characterized by high TBR (15.8%), CV (46.1%), GRI (69.2) and Gold score (3.0). Participants in the “Hyperglycemia” cluster were younger, socially vulnerable, and had more frequent hypertension than those in the “Euglycemia” cluster. To conclude, we identified three distinct, clinically relevant glycemic control profiles, enabling better characterization of PWT1D and, thus, more personalised management.