| 陈家辉,万雨林,汪莉萍.基于 LASSO回归的肝硬化病人食管静脉曲张破裂出血后并发细菌感染的预测模型建立[J].安徽医药,2025,29(12):2431-2438. |
| 基于 LASSO回归的肝硬化病人食管静脉曲张破裂出血后并发细菌感染的预测模型建立 |
| Nomogram model construction for bacterial infection complicated with esophageal variceal hemorrhage in cirrhotic patients based on LASSO regression |
| |
| DOI：10.3969/j.issn.1009-6469.2025.12.021 |
| 中文关键词: 肝硬化 食管静脉曲张破裂出血 细菌感染 危险因素 列线图 |
| 英文关键词: Liver cirrhosis Esophageal variceal hemorrhage Bacterial infection Risk factor Nomogram |
| 基金项目: |
|
| 摘要点击次数: 546 |
| 全文下载次数: 278 |
| 中文摘要: |
| 目的探索肝硬化病人食管静脉曲张破裂出血后并发细菌感染的影响因素，并进行列线图预测模型的建立并评价。方法回顾性纳入 2017年 10月至 2023年 10月于徐州医科大学附属医院因食管静脉曲张破裂出血就诊住院的肝硬化病人共 373例，根据是否产生细菌感染分为感染组( 154例)与非感染组( 219例)，收集病人一般资料、实验室指标等数据，采用简单随机抽样法以 7∶3将病人分为训练组和验证组。在训练组中经单因素 logistic回归分析及 LASSO回归筛选变量，使用多因素 logistic回归分析细菌感染发生的独立预测因素及建立列线图，并进行两组验证及评价。结果纳入病人中 41.3%(154/ 373)产生感染，腹腔感染占主要部分，占比78.5%(121/154)。多因素 logistic回归分析显示 Child-Pugh分级 C级［ OR=6.90， 95%CI:(2.77，18.02)， P<0.001］、中性粒细胞与淋巴细胞比值( NLR)［OR=1.24，95%CI:(1.08，1.45)， P=0.004］、 C反应蛋白(CRP)［OR=1.06，95%CI:(1.02，1.11)P=0.008］是病人发生细菌感染的独立预测风险因子，白蛋白是其独立的保护性因素［OR =0.89，95%CI:(0.81，0.98)P=0.016］。以，4个独立预测因素构建受试者操作特征曲线( ROC曲线)，训练组中， ROC曲线下面积(AUC)为 0.91［95%CI:(0.87，0，.94)］，验证组 AUC为 0.88［95%CI:(0.82，0.94)］。同时，校准曲线和决策曲线结果表明模型具有较好一致性和精准度。结论根据病人的一般情况、实验室指标，基于 LASSO回归，利用 Child-Pugh分级、 NLR、CRP、白蛋白构建了列线图，可以预测肝硬化病人食管静脉曲张破裂出血后并发细菌感染的发生概率，可为病人的临床治疗提供重要的参考和理论支撑。 |
| 英文摘要: |
| Objective To explore the factors affecting bacterial infection complicated with esophageal variceal hemorrhage in pa-tients with liver cirrhosis, and to develop and evaluate the nomogram prediction model.Methods A total of 373 patients with liver cir-rhosis were retrospectively enrolled in the study, who were hospitalized at The Affiliated Hospital of Xuzhou Medical University fromOctober 2017 to October 2023 due to esophageal variceal hemorrhage. They were assigned to the infection group (n=154) or the non-in-fection group (n=219), data on the general condition of the patients, laboratory indicators and other data were collected, and the patientswere randomly divided into a training set and a verification set according to 7∶3 using simple random sampling. Univariate logistic re-gression and LASSO regression were used to screen variables in the training set, and multiple logistic regression was used to analyzethe independent predictors of bacterial infection and to establish a nomogram, and validation and evaluation on the both sets were con-ducted. Results Of the enrolled patients 41.3% (154/373) developed infection, and abdominal infection accounted for the majority,reaching 78.5% (121/154). Multiple logistic regression analysis results showed that Child-Pugh grade C [OR=6.90, 95%CI: (2.77, 18.02), P<0.001], neutrophil-to-lymphocyte ratio (NLR) [OR=1.24, 95%CI: (1.08, 1.45), P=0.004], C reactive protein (CRP) [OR=1.06, 95%CI: (1.02, 1.11), P=0.008] were independent predictive risk factors for bacterial infection in patients, and albumin (ALB) was an in-dependent protective factor [OR=0.89, 95%CI: (0.81, 0.98), P=0.016]. The receiver operating characteristic curve(ROC curve) was con-structed with 4 independent predictors. In the training set, the area under the ROC curve (AUC) was 0.91 [95%CI: (0.87, 0.94)], while in the validation set, the AUC was 0.88 [95%CI: (0.82,0.94)]. Meanwhile, the calibration curve and decision curve results showed that the model had good consistency and accuracy.Conclusion According to the patients' general condition and laboratory indicators, andbased on LASSO regression, a nomogram is constructed using Child-Pugh classification, NLR, CRP, and ALB, which can predict the probability of bacterial infection after esophageal variceal hemorrhage in patients with liver cirrhosis, and provide important referenceand theoretical support for patients' clinical treatment. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
