中文|English

Current issue
2024-5-25
Vol 32, issue 5

ISSUE

2023 年9 期 第31 卷

论著 HTML下载 PDF下载

阿尔茨海默病的影响因素及其预测模型构建

Influencing Factors and Prediction Model Construction of Alzheimer Disease

作者:吴天晨,杨卉,梁艳

单位:
1.210001江苏省南京市中医院脑病科 2.210023江苏省南京市南京中医药大学护理学院
Units:
1.Department of Neurology, Nanjing Hospital of T.C.M, Nanjing 210001, China,2.School of Nursing, Nanjing University of Chinese Medicine, Nanjing 210023, China
关键词:
阿尔茨海默病;影响因素分析;预测模型
Keywords:
Alzheimer disease; Root cause analysis; Prediction model
CLC:
R 745.7
DOI:
10.12114/j.issn.1008-5971.2023.00.191
Funds:
国家自然科学基金资助项目(81904112);江苏省自然科学基金项目(BK20190136);南京市中医药青年人才计划(ZYQ20047)

摘要:

 目的 探讨阿尔茨海默病(AD)的影响因素并构建其预测模型。方法 选取2019年1月至2021年1月就诊于南京市中医院的61例AD患者作为AD组,另选取同期于该院体检中心进行体检的健康者122例作为健康组。比较两组基线资料。采用LASSO回归和多因素Logistic回归分析探讨AD的影响因素并构建其预测模型。采用ROC曲线评估预测模型对AD的预测价值。结果 两组性别、年龄、载脂蛋白E(ApoE)基因分型、TC、载脂蛋白B、游离三碘甲状腺原氨酸(FT3)、总甲状腺素(TT4)比较,差异有统计学意义(P<0.05)。LASSO回归分析结果显示,性别、年龄、ApoE基因分型、FT3、总三碘甲状腺原氨酸(TT3)是5个系数不为零的因子。多因素Logistic回归分析结果显示,性别、年龄、ApoE基因分型、TT3是AD的独立影响因素(P<0.05)。根据上述影响因素构建的预测模型如下:P=ex/(1+ex),其中x=-5.170+1.267×男性+0.058×年龄+2.389×ApoE3(ε3/ε3、ε2/ε4)+4.572×ApoE4(ε3/ε4、ε4/ε4)-2.059×TT3。ROC曲线分析结果显示,预测模型预测AD发生的AUC为0.885〔95%CI(0.832,0.938)〕。结论 性别、年龄、ApoE基因分型、TT3是AD的影响因素,而根据上述影响因素构建的预测模型对AD具有一定预测能力。

Abstract:

 Objective To explore the influencing factors of Alzheimer disease (AD) and construct its predictionmodel. Methods A total of 61 patients with AD who were admitted to Nanjing Hospital of T.C.M from January 2019 to January2021 were selected as AD group, and 122 healthy subjects who underwent physical examination in the physical examinationcenter of the same hospital during the same period were selected as healthy group. The baseline data were compared between thetwo groups. LASSO regression and multivariate Logistic regression analysis was used to investigate the influencing factors of ADand to construct its prediction model. ROC curve was used to evaluate the predictive value of prediction model for AD. ResultsThere were significant differences in gender, age, apolipoprotein E (ApoE) genotype, TC, apolipoprotein B, free triiodothyronine(FT3) and total thyroxine (TT4) between the two groups (P < 0.05) . LASSO regression analysis showed that gender, age, ApoEgenotype, FT3 and TT3 were 5 factors with non-zero coefficients. Multivariate Logistic regression analysis showed that gender, age,ApoE genotype, and TT3 were the independently influencing factors of AD (P < 0.05) . The prediction model constructed accordingto the above influencing factors was as follows: P=ex/ (1+ex) , where x=-5.170+1.267×male+0.058×age+2.389×ApoE3 (ε3/ε3,ε2/ε4) +4.572×ApoE4 (ε3/ε4, ε4/ε4) -2.059×TT3. ROC curve analysis showed that the AUC of prediction model forpredicting AD was 0.885 [95%CI (0.832, 0.938) ] . Conclusion Gender, age, ApoE genotype, and TT3 are the influencing factorsof AD, and the prediction model constructed according to the above influencing factors has a certain ability to predict AD.

ReferenceList: