首页|期刊导航|国际老年医学杂志|格列美脲治疗老年糖尿病时发生低血糖风险的预测模型

格列美脲治疗老年糖尿病时发生低血糖风险的预测模型OA

Prediction model for hypoglycemia risk in the treatment of elderly diabetes with glimepiride

中文摘要英文摘要

目的 构建老年 2 型糖尿病患者服用格列美脲治疗期间发生低血糖风险的预测模型,为临床个体化用药提供依据.方法 选取2022 年10 月—2024 年 10 月南京市第一医院老年医学科收治的规律口服格列美脲(单药或联合其他降糖药物)的 241 例老年 2 型糖尿病患者(年龄≥60 岁)的临床资料进行回顾性分析.根据患者在服用格列美脲治疗期间是否发生有症状或无症状的低血糖事件,将患者分为低血糖组60 例(24.90%)和非低血糖组 181 例(75.10%).通过单因素分析及多因素 logistic 回归分析筛选使用格列美脲治疗的老年糖尿病患者低血糖发生的独立影响因素,构建预测模型并评估其效能.结果 单因素分析显示,两组患者在年龄、糖尿病病程、体质量指数、估算肾小球滤过率(eGFR)、血清白蛋白、格列美脲日剂量、既往低血糖史、是否联用胰岛素以及是否联用 β 受体阻滞剂方面差异均有统计学意义(P<0.05).多因素logistic回归分析结果表明,高龄、低eGFR、高格列美脲日剂量、既往低血糖史以及联用胰岛素均为老年2 型糖尿病患者口服格列美脲期间发生低血糖的独立危险因素(P<0.05).ROC曲线预测模型的曲线下面积为 0.882(95%CI:0.812~0.918,P<0.001),选取 ROC 曲线上约登指数最大点(0.579)对应的概率值 0.251 作为最佳临界点,此时模型的灵敏度为 78.33%,特异度为79.56%,阳性预测值为55.95%,阴性预测值为91.72%.结论 基于高龄、肾功能、药物剂量、低血糖史以及胰岛素联用的预测模型对老年糖尿病患者服用格列美脲期间低血糖发生的风险预测具有良好的临床价值.

Objective To construct a risk prediction model for hypoglycemia in elderly patients with type 2 diabetes treated with glimepiride,providing a basis for individualized clinical medication.Methods The clinical data of 241 elderly patients(aged≥60 years)with type 2 diabetes who were regularly treated with oral glimepiride(monotherapy or in combination with other hypoglycemic agents)were retrospectively collected from the Department of Geriatrics,Nanjing First Hospital,from October 2022 to October 2024.According to whether patients experienced symptomatic or asymptomatic hypoglycemic events during treatment with glimepiride,they were divided into a hypoglycemic group of 60 cases(24.90%)and a non hypoglycemic group of 181 cases(75.10%).The independ-ent influencing factors of hypoglycemia in elderly diabetes patients treated with glimepiride were screened through single factor analysis and multi factor logistic regression analysis,and the predictive model was constructed and its efficacy was evaluated.Results Univa-riate analysis showed that there were statistically significant differences between the hypoglycemia group and the non hypoglycemia group in terms of age,duration of diabetes,body mass index,estimated glomerular filtration rate(eGFR),serum albumin,daily dose of glimepiride,previous history of hypoglycemia,combined use of insulin,and combined use of β-blockers(P<0.05).Multivariate logistic regression analysis revealed that advanced age,lower eGFR,higher daily dose of glimepiride,a previous history of hypoglyce-mia,and combined use of insulin were independent risk factors for glimepiride-induced hypoglycemia in elderly patients with type 2 di-abetes(P<0.05).The area under the curve of the ROC curve prediction model is 0.882(95%CI:0.812-0.918,P<0.001).The probability value of 0.251 corresponding to the maximum point of the Jordan index on the ROC curve(0.579)is selected as the opti-mal critical point.At this point,the sensitivity of the model is 78.33%,the specificity is 79.56%,the positive predictive value is 55.95%,and the negative predictive value is 91.72%.Conclusion The prediction model based on advanced age,renal function,drug dose,history of hypoglycemia and insulin combination has good clinical value in predicting the risk of hypoglycemia in elderly pa-tients with diabetes during glimepiride administration.

陈文洁;李林;季梅丽;戴露

南京医科大学附属南京医院(南京市第一医院)老年医学科,江苏 南京 210000南京医科大学附属南京医院(南京市第一医院)老年医学科,江苏 南京 210000南京医科大学附属南京医院(南京市第一医院)老年医学科,江苏 南京 210000南京医科大学附属南京医院(南京市第一医院)内分泌科,江苏 南京 210000

格列美脲2 型糖尿病低血糖风险预测

GlimepirideType 2 diabetesHypoglycemiaRisk prediction

《国际老年医学杂志》 2026 (2)

171-176,6

江苏省干部保健科研课题(BJ24038)

10.3969/j.issn.1674-7593.2026.02.008

评论