Building clinical risk score systems for predicting the all-cause and expanded cardiovascular-specific mortality of patients with type 2 diabetes

Aim To develop and validate risk score systems by examining the effects of glycaemic and blood pressure variabilities on the all-cause and expanded cardiovascular-specific mortality of people with type 2 diabetes. Materials and methods This retrospective cohort study consisted of 9692 patients aged 30-85 years, diagnosed with type 2 diabetes and enrolled in a managed care programme of a medical centre from 2002 to 2016. All the patients were randomly allocated into two groups, namely, training and validation sets (2:1 ratio), and followed up until death or August 2019. Cox's proportional hazard regression was performed to develop all-cause and expanded cardiovascular-specific mortality prediction models. The performance of the prediction model was assessed by using the area under the receiver operating characteristic curve (AUROC). Results Overall, 2036 deaths were identified after a mean of 8.6 years of follow-up. The AUROC-measured prediction accuracies of 3-, 5-, 10- and 15-year all-cause mortalities based on a model containing the identified traditional risk factors, biomarkers and variabilities in fasting plasma glucose, HbA1c and blood pressure in the validation set were 0.79 (0.76-0.83), 0.78 (0.76-0.81), 0.80 (0.78-0.82) and 0.80 (0.78-0.82), respectively. The corresponding values of the expanded cardiovascular-specific mortalities were 0.85 (0.80-0.90), 0.83 (0.79-0.86), 0.80 (0.77-0.83) and 0.79 (0.77-0.82), respectively. Conclusions Our prediction models considering glycaemic and blood pressure variabilities had good prediction accuracy for the expanded cardiovascular-specific and all-cause mortalities of patients with type 2 diabetes.