Chen Jing, Yue Ling, Wang Ruonan, Shu Sunjing, Liu Jin, Yan Mingmin, Ye Changkong, Shuang Liu
Ultrasound Department, The Second Clinical Medical College, The First Affiliated Hospital, Shenzhen Peoples Hospital, Jinan University, Southern University of Science and Technology), Shenzhen, 518020, Guangdong, China.
Department of Ultrasound, The Fourth Affiliated Hospital of China Medical University, Chongshandong Road, Huanggu District 4, Shenyang City, Liaoning Province, China.
BMC Cardiovasc Disord. 2025 May 15;25(1):365. doi: 10.1186/s12872-025-04737-1.
The long-term prognosis of diabetic patients with coronary artery disease (CAD) is influenced by various clinical variables and biomarkers. This study aimed to develop and validate a prognostic model that integrates clinical, echocardiographic, and angiographic data to predict disease progression.
We retrospectively analyzed 396 diabetic CAD patients with a 3-year follow-up starting from their first coronary angiography. Outcome variables included recurrent myocardial infarction, unstable angina rehospitalization, heart failure, ischemic stroke, cardiovascular death, and all-cause death. Non-progression was defined as the absence of these events. Variables included clinical data, echocardiographic parameters, coronary angiography results, and biomarkers. A multivariate Cox regression model was developed, incorporating key factors (coronary lesion number, myocardial infarction history, ejection fraction, and creatinine).
Multivariate analysis identified the number of obstructed coronary arteries, history of myocardial infarction, ejection fraction, and creatinine level as independent predictors of disease progression. The model showed good predictive performance, with AUC values of 0.742, 0.782, and 0.816 at 3, 6, and 9 months, respectively. The C-index was 0.669 (95% CI: 0.5959-0.7196) in the training set and 0.695 (95% CI: 0.5781-0.7436) in the validation set, reflecting consistent predictive performance. Calibration curves showed excellent agreement between predicted and observed outcomes.
We developed and validated a practical nomogram integrating clinical, biochemical, and imaging data to predict short-term disease progression in diabetic patients with CAD. This tool may assist clinicians in early risk stratification and individualized management planning.
糖尿病合并冠状动脉疾病（CAD）患者的长期预后受多种临床变量和生物标志物影响。本研究旨在建立并验证一个整合临床、超声心动图和血管造影数据的预后模型，以预测疾病进展。
我们回顾性分析了396例糖尿病CAD患者，从其首次冠状动脉造影开始进行为期3年的随访。结局变量包括复发性心肌梗死、不稳定型心绞痛再入院、心力衰竭、缺血性卒中、心血管死亡和全因死亡。无进展定义为未发生这些事件。变量包括临床数据、超声心动图参数、冠状动脉造影结果和生物标志物。建立了一个多变量Cox回归模型，纳入关键因素（冠状动脉病变数量、心肌梗死病史、射血分数和肌酐）。
多变量分析确定冠状动脉阻塞数量、心肌梗死病史、射血分数和肌酐水平为疾病进展的独立预测因素。该模型显示出良好的预测性能，在3、6和9个月时的AUC值分别为0.742、0.782和0.816。训练集中的C指数为0.669（95%CI：0.5959 - 0.7196），验证集中为0.695（95%CI：0.5781 - 0.7436），反映出一致的预测性能。校准曲线显示预测结果与观察结果之间具有良好的一致性。
我们建立并验证了一个实用的列线图，整合临床、生化和影像数据以预测糖尿病CAD患者的短期疾病进展。该工具可协助临床医生进行早期风险分层和个体化管理规划。
