Identifying subtypes of type 2 diabetes mellitus based on real-world electronic medical record data in China

Jiayu Wang,Bixia Gao,Jinwei Wang,Wenwen Liu,Weijia Yuan,Yangfan Chai,Jun Ma,Yangyang Ma,Guilan Kong,Minchao Liu
DOI: https://doi.org/10.1016/j.diabres.2024.111872
2024-09-26
Abstract:Aims: To replicate the European subtypes of type 2 diabetes mellitus (T2DM) in the Chinese diabetes population and investigate the risk of complications in different subtypes. Methods: A diabetes cohort using real-world patient data was constructed, and clustering was employed to subgroup the T2DM patients. Kaplan-Meier analysis and the Cox models were used to analyze the association between diabetes subtypes and the risk of complications. Results: A total of 2,652 T2DM patients with complete clustering data were extracted. Among them, 466 (17.57 %) were classified as severe insulin-deficient diabetes (SIDD), 502 (18.93 %) as severe insulin-resistant diabetes (SIRD), 672 (25.34 %) as mild obesity-related diabetes (MOD), and 1,012 (38.16 %) as mild age-related diabetes (MARD). The risk of chronic kidney disease (CKD) and diabetic retinopathy (DR) were different in the four subtypes. Compared with MARD, SIRD had a higher risk of CKD (HR 2.40 [1.16, 4.96]), and SIDD had a higher risk of DR (HR 2.16 [1.11, 4.20]). The risk of stroke and coronary events had no difference. Conclusions: The European T2DM subtypes can be replicated in the Chinese diabetes population. The risk of CKD and DR varied among different subtypes, indicating that proper interventions can be taken to prevent specific complications in different subtypes.
What problem does this paper attempt to address?