2型糖尿病合并高血压的危险因素及其列线图预测模型的构建

Risk factors of type 2 diabetes mellitus complicating hypertension and its nomogram prediction model construction

目的通过分析2型糖尿病合并高血压(DH)的危险因素,建立个体化预测发生DH风险的列线图模型,并验证其效能。方法选取2017年3月至2019年4月新疆医科大学第一附属医院收治的2型糖尿病患者3527例作为研究对象,采用简单随机化方法分为训练组(2646例,75.0%)与验证组(881例,25.0%),运用Lasso回归结合多因素logistic回归模型分析构建DH列线图预测模型。由验证组评估DH列线图预测模型的可行性。最后采用受试者工作特征曲线(ROC曲线)、校准曲线和决策曲线分析(DCA)评估DH列线图预测模型的鉴别能力、准确性和临床实用性。结果年龄、糖尿病病程、低密度脂蛋白、体重指数、尿蛋白均为DH的独立风险因素,差异均有统计学意义(P<0.05)。将上述5个变量纳入构建DH列线图预测模型,模型在训练组和验证组中预测进展为DH风险的ROC曲线下面积分别为0.815(95%可信区间:0.80~0.83)、0.809(95%可信区间:0.78~0.84),校准曲线和DCA同样显示模型具有良好的准确性和临床实用性。结论建立的DH列线图预测模型有助于临床早期甄别DH高危人群,值得临床推广应用。

Objective To analyze the risk factors of type 2 diabetes mellitus(T2DM)complicating hypertension(DH),and to establish an individualized nomogram model for predicting the risk occurrence of DH and verify its efficiency.Methods A total of 3527 patients with T2DM treated in the First Affiliated Hospital of Xinjiang Medical University from March 2017 to April 2019 were selected as the study subjects and divided into the training group(2646 cases,75%)and validation group(881 cases,25%)by the simple randomization method.The DH nomogram prediction model was constructed by Lsasp regression combined with multivariate regression model analysis.The feasibility of DH nomogram prediction model was evaluated by the validation group.Finally,the receiver operating characteristic(ROC)curve,calibration curve and decision curve analysis(DCA)were adopted to evaluate the discriminant ability,accuracy and clinical practicability of DH nomogram prediction model.Results The age,diabetes duration,low density lipoprotein,BMI and urinary protein were the independent risk factors for DH,and the differences were statistically significant(P<0.05).The above five variables were included into the DH nomogram prediction model.The areas under ROC curve for predicting the risk for progressing to DH risk in the modeling group and verification group were 0.815(95%CI:0.80-0.83)and 0.809(95%CI:0.78-0.84)respectively.The calibration curves and DCA similarly showed that the model had good accuracy and clinical applicability.Conclusion The established DH nomogram prediction model is helpful to identify the high-risk population of DH in the early clinical stage,which is worthy of clinical promotion and application.