摘要
目的 探索影响增殖性IgA肾病患者预后的因素,并构建疾病预测模型。方法 回顾性分析2020年1月至2022年12月期间在中国科学技术大学附属第一医院(安徽省立医院)肾脏内科确诊为增殖性IgA肾病患者的临床资料,所有患者均随访一年,根据是否出现联合终点事件判断预后情况。采用最小绝对收缩和选择算子(least absolute shrinkage and selection operator,LASSO)回归对增殖性IgA肾病患者出现联合终点事件的危险因素进行筛选,LASSO回归中的调节参数λ采用10折交叉验证方法进行验证,选取偏差最小时的λ值对应的回归系数不为0的变量纳入多因素Logistic回归模型,建立增殖性IgA肾病患者出现联合终点事件危险因素预测模型。基于自助法进行1000次重复抽样验证模型,绘制受试者工作特征(receiver operating characteristic,ROC)曲线,并计算曲线下面积(area under the curve,AUC),绘制校准曲线进行模型稳定性的评估。结果 根据纳入及排除标准,研究纳入患者215例,其中出现终点事件52例,无终点事件163例。单因素分析提示了10个差异有统计学意义的指标,使用LASSO回归降维处理,提示了6项最佳建模指标,将其进行Logistic回归分析,最终得到了3项指标:24 h尿蛋白定量、血红蛋白、血尿酸。构建危险因素模型:Logi(P)=-1.017+0.198×24 h尿蛋白定量(mg)-0.089×血红蛋白(g/L)+0.435×血尿酸(μmol/L)。校准曲线提示模型的预测概率与实际概率具有较高的吻合度,C指数为0.915。ROC曲线提示,AUC为0.904(95%CI:0.874~0.936),模型预测能力较好。结论 本研究基于LASSO-Logistic建立了增殖性IgA肾病患者预后的预测模型,临床中有助于帮助患者判断疾病预后,具有较好的应用价值。
Objective To explore the factors affecting the prognosis of proliferative IgA ne-phropathy,and to construct a disease prediction model.Methods Clinical data of patients diagnosed with proliferative IgA nephropathy in the Department of Nephrology,the First Affiliated Hospital of Uni-versity of Science and Technology of China(Anhui Provincial Hospital)from January 2020 to December 2022 were retrospectively analyzed.All patients were followed up for one year,and the prognosis was de-termined according to the occurrence of joint endpoint events.The least absolute shrinkage and selection operator(LASSO)regression was used to screen the risk factors for joint endpoint events in patients with proliferative IgA nephropathy.The adjustment parameterλin LASSO regression was verified by 10-fold cross-validation method.The variables with the regression coefficient not equal to 0 corresponding to theλvalue with the smallest deviation were included in the multivariate Logistic regression model.Risk fac-tor prediction model for joint endpoint events in patients with proliferative IgA nephropathy was con-structed.Based on Bootstrap method,the model was verified with 1000 repeated samples,and validated by the receiver operating characteristic(ROC)curve,and the area under the curve(AUC).Calibration curves were drawn to evaluate the stability of the model.Results According to the inclusion and exclu-sion criteria,215 patients were included in the study,of which 52 had endpoint events and 163 had no endpoint events.Univariate analysis indicated ten indicators with statistically significant differences.LASSO regression was used to reduce dimensionality,and 6 optimal modeling indicators were sug-gested.Logistic regression analysis was performed on them,and finally 3 indicators were obtained:24-hour urinary protein quantification,hemoglobin and blood uric acid.The risk factor model was estab-lished:Logit(P)=-1.017+0.198×24-hour urinary protein quantity(mg)‒0.089×hemoglobin(g/L)+0.435×blood uric acid(μmol/L).The calibration curve indicated that the prediction probability of the model was highly consistent with the actual probability,with a C-index of 0.915.The ROC curve indi-cated that the AUC was 0.904(95%CI:0.874-0.936),and the model had good prediction ability.Con-clusion A prediction model for the prognosis of proliferative IgA nephropathy is established based on LASSO-Logistic,which is helpful to judge the prognosis of the disease in clinical practice and has good application value.
作者
朱峰博
姜俊
任伟
Zhu Feng-bo;J iang Jun;Ren Wei(Anhui Medical University Provincial Clinical College,Hefei 230032,China;Department of Nephrol-ogy,the First Affiliated Hospital of University of Science and Technology of China(Anhui Provincial Hospital),Hefei 230001,China)
出处
《临床肾脏病杂志》
2025年第7期537-543,共7页
Journal Of Clinical Nephrology
关键词
增殖性IgA肾病
预后
最小绝对收缩和选择算子
预测模型
Proliferative IgA nephropathy
Prognosis
Least absolute shrinkage and selection operator
Forecasting model
作者简介
通信作者:任伟,Email:renweisn@163.com。
