2型糖尿病患者下肢血管病变介入治疗后再狭窄影响因素分析及列线图预测模型构建

Analysis of influencing factors of restenosis after interventional therapy for lower extremity vascular disease in patients with type 2 diabetes mellitus and construction of nomogram prediction model

目的分析影响2型糖尿病患者下肢血管病变介入治疗后再狭窄的因素,并构建列线图预测模型。方法本研究为前瞻性研究,选取2021年9月至2022年9月在新疆医科大学第一附属医院收治的316例行下肢血管病变介入治疗的2型糖尿病患者,按照7∶3将其随机分为建模组(n=221)和验证组(n=95)。收集患者资料,对患者进行半年随访,根据建模组患者术后有无出现再狭窄分为狭窄组(n=31)和未狭窄组(n=190)。单因素及多因素Logistic回归分析影响2型糖尿病患者下肢血管病变介入治疗后发生再狭窄的因素,并根据结果构建列线图预测模型。以受试者工作特征(ROC)曲线评估模型的区分度,以H-L拟合度曲线评估模型的预测有效性。结果单因素分析显示,年龄、总胆固醇、糖化血红蛋白(HbA1c)、空腹血糖水平为影响2型糖尿病患者下肢血管病变介入治疗后发生再狭窄的因素(P<0.05)。多因素Logistic回归分析显示,高水平总胆固醇、HbA1c、空腹血糖为影响2型糖尿病患者下肢血管病变介入治疗后发生再狭窄的独立危险因素(P<0.05)。验证模型显示,建模组预测有效性较好(χ^(2)=6.758,P=0.437),ROC曲线下面积(AUC)为0.964(95%CI:0.930~0.984),敏感性及特异性分别为83.87%、99.47%;验证组预测有效性较好(χ^(2)=6.112,P=0.315),AUC为0.916(95%CI:0.872~0.949),敏感性及特异性分别为93.55%、84.74%。结论高水平总胆固醇、HbA1c、空腹血糖为影响2型糖尿病患者下肢血管病变介入术后再狭窄的影响因素,而本研究基于影响因素构建的列线图模型具有较高的实用性,可能用于临床预测再狭窄发生的风险。

Objective To analyze the factors affecting restenosis after interventional therapy for lower extremity vascular disease in patients with type 2 diabetes mellitus,and to construct a nomogram prediction model.Methods This study was a prospective study.From September 2021 to September 2022,316 patients with type 2 diabetes who were treated with interventional therapy for lower extremity vascular diseases in the First Affiliated Hospital of Xinjiang Medical University were selected and divided into modeling group(n=221)and verification group(n=95)according to 7∶3.The patients'data were collected and followed up for half a year.According to whether the patients in the modeling group had restenosis after operation,they were divided into stenosis group(n=31)and non-stenosis group(n=190).The factors influencing restenosis after interventional therapy of lower extremity vascular diseases in patients with type 2 diabetes mellitus were analyzed by univariate and multivariate Logistic regression,and a nomogram prediction model was constructed according to the results.The discrimination of the model was evaluated by receiver operating characteristic curve,and the prediction effectiveness of was evaluated by the model H-L fitting curve.Results Univariate analysis showed that age,total cholesterol,glycosylated hemoglobin(HbA1c)and fasting blood glucose were the influencing factors of restenosis in patients with type 2 diabetes after interventional therapy(P<0.05).Multivariate Logistic regression analysis showed that high levels of total cholesterol,HbA1c and FBG were independent risk factors for restenosis in patients with type 2 diabetes after interventional therapy(P<0.05).The validation model showed that the modeling group had good predictive effectiveness(χ^(2)=6.758,P=0.437),the AUC under ROC curve was 0.964(95%CI:0.930-0.984),and the sensitivity and specificity were 83.87%and 99.47%respectively.The validation group had good predictive effectiveness(χ^(2)=6.112,P=0.315),AUC was 0.916(95%CI:0.872-0.949),and the sensitivity and specificity were 93.55%and 84.74%respectively.Conclusion High levels of total cholesterol,HbA1c and fasting blood glucose are the influencing factors for restenosis after interventional operation of lower extremity vascular diseases in type 2 diabetes patients.In this study,the column graph model constructed based on the influencing factors has high practicability and may be used for clinical prediction of the risk of restenosis.