2型糖尿病并发结直肠腺瘤性息肉的Lasso-Logistic回归分析及Nomogram预测模型构建  

Lasso-Logistic regression analysis of type 2 diabetes complicated by colorectal adenomatous polyps and Nomogram prediction model construction

作　　者：王洁 刘正娟 王红霞 WANG Jie;LIU Zhengjuan;WANG Hongxia(Department of Endocrinology and Hematology,Qingdao Jiaozhou Central Hospital,Qingdao,Shandong 266300,China;Department of Gastroenterology,Qingdao Jiaozhou Central Hospital,Qingdao,Shandong 266300,China;Digestive Endoscopy Center,Qingdao Jiaozhou Central Hospital,Qingdao,Shandong 266300,China)

机构地区：[1]山东省青岛市胶州中心医院内分泌血液科,山东青岛266300 [2]山东省青岛市胶州中心医院消化内科,山东青岛266300 [3]山东省青岛市胶州中心医院消化内镜中心,山东青岛266300

出　　处：《检验医学与临床》2025年第6期778-784,共7页Laboratory Medicine and Clinic

基　　金：2022-2024年山东省青岛市市级临床重点专科胃肠道肿瘤综合治疗专科资助项目。

摘　　要：目的探讨2型糖尿病(T2DM)并发结直肠腺瘤性息肉(CAP)的影响因素,并构建T2DM并发CAP的Nomogram预测模型。方法选取2022年1月至2023年12月在该院进行肠镜检查的448例T2DM患者作为研究对象,根据是否检出CAP分为CAP组与非CAP组。收集2组患者的糖尿病病程、性别、年龄、有无高血压史、有无冠心病史、有无吸烟史、体质量指数、有无饮酒史、文化程度、是否合并胆囊疾病、是否合并幽门螺杆菌感染、是否使用胰岛素、是否使用二甲双胍、是否使用α-糖苷酶抑制剂、是否使用阿司匹林、是否合并糖尿病视网膜病变、是否合并糖尿病周围神经病变、是否合并周围动脉狭窄或闭塞、膳食钙摄入不足等一般资料及血清总胆固醇(TC)、甘油三酯(TG)、低密度脂蛋白胆固醇(LDL-C)、γ-谷氨酰转移酶(GGT)、糖化血红蛋白(HbA1c)、胰岛素抵抗指数(HOMA-IR)、尿酸、维生素D等实验室指标水平,并进行对比分析。经Lasso回归分析和多因素Logistic回归分析T2DM并发CAP的影响因素,并根据影响因素构建T2DM并发CAP的Nomogram预测模型;绘制受试者工作特征(ROC)曲线、校准曲线对Nomogram预测模型进行验证。结果448例T2DM患者中,289例并发CAP患者纳入CAP组,其余159例患者纳入无CAP组,T2DM并发CAP的检出率为64.51%(289/448)。2组HbA1c水平、性别、GGT水平、有饮酒史占比、TG水平、糖尿病病程、合并胆囊疾病占比、合并幽门螺杆菌感染占比、使用胰岛素占比、膳食钙摄入不足占比、低维生素D水平占比及体质量指数比较,差异均有统计学意义(P<0.05)。Lasso回归分析和多因素Logistic回归分析结果显示,男性、有饮酒史、糖尿病病程≥5年、幽门螺杆菌感染、使用胰岛素、膳食钙摄入不足、低维生素D水平、高HbA1c水平、高GGT水平是T2DM患者并发CAP的危险因素(P<0.05);根据上述影响因素构建T2DM并发CAP的Nomogram预测模型并验证,结Objective To investigate the influencing factors of colorectal adenomatous polyp(CAP)complicated by type 2 diabetes mellitus(T2DM)and construct a Nomogram prediction model.Methods A total of 448 T2DM patients who underwent colonoscopy in the hospital from January 2022 to December 2023 were selected and divided into CAP group and no-CAP group according to whether CAP was detected or not.The course of diabetes,gender,age,history of hypertension,history of coronary heart disease,smoking history,body mass index,alcohol consumption history,education level,gallbladder disease,Helicobacter pylori infection,insulin use,metformin use,α-glycosidase inhibitor use,aspirin use,diabetic retinopathy,diabetic peripheral neuropathy,peripheral artery stenosis or occlusion,dietary calcium intake deficiency and other general information,as well as serum total cholesterol(TC),triglyceride(TG),low density lipoprotein cholesterol(LDL-C),γ-glutamyltransferase(GGT),glycosylated hemoglobin(HbA1c),insulin resistance index(HOMA-IR),uric acid,vitamin D levels and other laboratory indicators were collected.After multivariate and Lasso-Logistic regression analysis of the influencing factors of T2DM complicating by CAP,a T2DM concurrent CAP prediction model was constructed based on the influencing factors,and the Nomogram prediction model was validated by the receiver operating characteristic(ROC)curve and calibration curve.Results Among 448 patients with T2DM,289 patients with concurrent CAP were included in the CAP group,and the remaining 159 patients were included in the no-CAP group,with detection rate of concurrent CAP of 64.51%(289/448).The differences of HbA1c level,gender,GGT level,proportion of alcohol consumption history,TG level,course of diabetes,proportion of concurrent gallbladder disease,proportion of Helicobacter pylori infection,proportion of insulin use,proportion of inadequate dietary calcium intake,proportion of low vitamin D level and body mass index between the two groups were statistically significant(P<0.05).The result

关 键 词：2型糖尿病 结直肠腺瘤性息肉 Nomogram预测模型 幽门螺杆菌感染 膳食钙 

分 类 号：R735.3[医药卫生—肿瘤] R587.1[医药卫生—临床医学]