【摘要】 目的 针对高血压合并糖尿病人群构建预测模型，以评估慢性肾脏病(chronic kidney disease, CKD)的患病风险，为制定针对性的CKD防控措施提供科学依据。方法 选取2022年南京市60岁及以上社区人群体检中高血压合并糖尿病10 221名患者作为研究对象。通过单因素分析筛选出与CKD患病相关的变量，并利用LASSO回归进一步筛选变量，最终基于logistic回归模型构建CKD风险预测模型。模型的性能通过ROC曲线和校准曲线进行评估。结果 在研究人群中，CKD的患病率为22.71%,平均年龄为71.66岁。LASSO回归筛选出7个CKD的相关变量，包括年龄、血尿素氮、血红蛋白、尿酸、三酰甘油-葡萄糖指数、尿蛋白/肌酐比值和医疗保险方式。最终的logistic回归模型纳入6个变量：年龄[OR=1.067(95%CI:1.058～1.076)]、血尿素氮[OR=1.377(95%CI:1.338～1.418)]、血红蛋白[OR=0.992(95%CI:0.989～0.995)]、尿酸[OR=1.004(95%CI:1.003～1.004)]、三酰甘油-葡萄糖指数[OR=1.445(95%CI:1.324～1.577)]和医疗保险方式为自费[OR=1.732(95%CI:1.542～1.945)]。模型的AUC值为0.759(95%CI:0.747～0.770),Brier评分为0.140(95%CI:0.136～0.145),表明具有良好的预测效能，校准曲线显示预测风险与实际观察值一致性较好。结论 构建的LASSO-logistic回归风险预测模型能够有效评估60岁及以上老年高血压合并糖尿病人群的CKD患病风险，为早期识别高风险个体和制定针对性的CKD防控措施提供了依据。更多还原

【Abstract】 Objective To construct a prediction model for the population with hypertension and diabetes to assess the risk of chronic kidney disease(CKD),and to provide a scientific basis for formulating targeted CKD prevention and control measures. Methods Based on physical examination data from community residents aged 60 years and above in Nanjing in 2022,10 221 patients with hypertension and diabetes were selected as the study subjects.Variables associated with CKD prevalence were screened using univariate analysis, and further variable selection was performed using LASSO regression.Finally, a CKD risk prediction model was constructed based on logistic regression.The model′s performance was evaluated using the ROC curve and calibration curve. Results The prevalence rate of CKD in the study population was 22.71%,with a mean age of 71.66 years.LASSO regression identified seven variables associated with CKD:age, blood urea nitrogen(BUN),hemoglobin, uric acid, triglyceride-glucose(TyG) index, urine protein-to-creatinine ratio(UPCR),and medical insurance type.The final logistic regression model incorporated six variables: age [OR=1.067(95% CI:1.058-1.076)],BUN [OR=1.377(95% CI:1.338-1.418)],hemoglobin [OR=0.992(95% CI:0.989-0.995)],uric acid [OR=1.004(95% CI:1.003-1.004)],TyG index [OR=1.445(95% CI:1.324-1.577)],and self-payment medical insurance [OR=1.732(95% CI:1.542-1.945)].The model had an AUC of 0.759(95% CI:0.747-0.770) and a Brier score of 0.140(95% CI:0.136-0.145),indicating good predictive performance.The calibration curve showed good agreement between the predicted risk and the observed value. Conclusion The constructed LASSO-logistic regression risk prediction model in this study can effectively assess the risk of CKD in elderly individuals aged 60 years and above with hypertension and diabetes, providing a basis for early identification of high-risk individuals and the formulation of targeted CKD prevention and control measures.更多还原