Validation and comparison of SCORE2-Diabetes and ASCVD score in Japanese patients with diabetes without a history of cardiovascular disease

Background/Introduction Type 2 diabetes mellitus is a major risk factor for cardiovascular disease (CVD), and several risk prediction models have been proposed. However, few risk prediction models have been externally validated in Asian populations. Purpose We sought to validate the Systematic Coronary Risk Evaluation 2-Diabetes (SCORE2-Diabetes) model recommended by the European Society of Cardiology and the Atherosclerotic Cardiovascular Disease (ASCVD) score recommended by the American College of Cardiology and the American Heart Association, while also comparing their predictive performance in Japanese patients with diabetes mellitus without a history of CVD. Methods We retrospectively collected data from all patients admitted to our hospital for diabetes education between January 2011 and December 2022. All patients underwent laboratory tests and physical measurements, and their backgrounds and medications were recorded during hospitalization. Prognosis after discharge was retrospectively collected from electronic medical records. The primary endpoint is a composite of myocardial infarction and stroke. Patients who already have a history of CVD at the time of admission and who have not risk scores calculated due to missing data were excluded. The performances of SCORE2-Diabetes and ASCVD score was evaluated by the area under the receiver-operating-characteristics curve (AUROC) and compared using DeLong's test. Results We evaluated a total of 1,186 patients with diabetes (median age of 59 years, 65% male, mean glycohemoglobin 8.5%) without a history of CVD, requiring admission for diabetes education. During a median follow-up of 3.2 years, 30 patients (2.5%) experienced the primary endpoint. Both the SCORE2-Diabetes (hazard ratio (HR): 1.09, 95% confidence interval (CI): 1.05–1.14) and the ASCVD score (HR: 1.04, 95% CI: 1.01–1.06) were significantly associated with the primary endpoint. The discrimination ability was superior in SCORE2-Diabetes (AUROC: 0.688, 95% CI: 0.585-0.791) compared to ASCVD score (AUROC: 0.639, 95% CI: 0.534-0.744), with a significantly higher AUROC (p = 0.008). Conclusion Although both the SCORE2-Diabetes model and the ASCVD model were validated in Japanese patients with diabetes mellitus without a history of CVD, the SCORE2-Diabetes model demonstrated superior predictive performance compared to ASCVD score. However, the discriminative performance of the models was modest, and a modified or novel risk model is needed.ROC and AUROC of two risk models