Chen Fei, Wang Jinping, Chen Xiaoping, Yu Liping, An Yali, Gong Qiuhong, Chen Bo, Xie Shuo, Zhang Lihong, Shuai Ying, Zhao Fang, Chen Yanyan, Li Guangwei, Zhang Bo
Department of Endocrinology, China-Japan Friendship Hospital, Beijing, China.
Department of Cardiology, Da Qing First Hospital, Da Qing, China.
Diabetol Metab Syndr. 2023 Mar 30;15(1):62. doi: 10.1186/s13098-023-01039-4.
This study aimed to develop cardiovascular disease (CVD) risk equations for Chinese patients with newly diagnosed type 2 diabetes (T2D) to predict 10-, 20-, and 30-year of risk.
Risk equations for forecasting the occurrence of CVD were developed using data from 601 patients with newly diagnosed T2D from the Da Qing IGT and Diabetes Study with a 30-year follow-up. The data were randomly assigned to a training and test data set. In the training data set, Cox proportional hazard regression was used to develop risk equations to predict CVD. Calibration was assessed by the slope and intercept of the line between predicted and observed probabilities of outcomes by quintile of risk, and discrimination was examined using Harrell's C statistic in the test data set. Using the Sankey flow diagram to describe the change of CVD risk over time.
Over the 30-year follow-up, corresponding to a 10,395 person-year follow-up time, 355 of 601 (59%) patients developed incident CVD; the incidence of CVD in the participants was 34.2 per 1,000 person-years. Age, sex, smoking status, 2-h plasma glucose level of oral glucose tolerance test, and systolic blood pressure were independent predictors. The C statistics of discrimination for the risk equations were 0.748 (95%CI, 0.710-0.782), 0.696 (95%CI, 0.655-0.704), and 0.687 (95%CI, 0.651-0.694) for 10-, 20-, and 30- year CVDs, respectively. The calibration statistics for the CVD risk equations of slope were 0.88 (P = 0.002), 0.89 (P = 0.027), and 0.94 (P = 0.039) for 10-, 20-, and 30-year CVDs, respectively.
The risk equations forecast the long-term risk of CVD in patients with newly diagnosed T2D using variables readily available in routine clinical practice. By identifying patients at high risk for long-term CVD, clinicians were able to take the required primary prevention measures.
本研究旨在为新诊断的2型糖尿病（T2D）中国患者开发心血管疾病（CVD）风险方程，以预测10年、20年和30年的风险。
利用大庆IGT和糖尿病研究中601例新诊断T2D患者的30年随访数据，开发预测CVD发生的风险方程。数据被随机分配到训练和测试数据集。在训练数据集中，使用Cox比例风险回归开发预测CVD的风险方程。通过按风险五分位数划分的预测和观察到的结局概率之间的直线斜率和截距评估校准，并在测试数据集中使用Harrell's C统计量检查区分度。使用桑基流程图描述CVD风险随时间的变化。
在30年的随访中，对应10395人年的随访时间，601例患者中有355例（59%）发生了新发CVD；参与者中CVD的发病率为每1000人年34.2例。年龄、性别、吸烟状况、口服葡萄糖耐量试验的2小时血浆葡萄糖水平和收缩压是独立预测因素。10年、20年和30年CVD风险方程的区分度C统计量分别为0.748（95%CI，0.710 - 0.782）、0.696（95%CI，0.655 - 0.704）和0.687（95%CI，0.651 - 0.694）。10年、20年和30年CVD风险方程斜率的校准统计量分别为0.88（P = 0.002）、0.89（P = 0.027）和0.94（P = 0.039）。
风险方程使用常规临床实践中容易获得的变量预测新诊断T2D患者的CVD长期风险。通过识别长期CVD高风险患者，临床医生能够采取所需的一级预防措施。
