血液透析患者发生透析失衡综合征风险模型的构建

农梅.血液透析患者发生透析失衡综合征风险模型的构建[J].内科急危重症杂志,2026,32(1):57-60

DOI：10.11768/nkjwzzzz20260113

英文关键词:

基金项目:

作者	单位	E-mail
农梅		vndurk313@163.com

摘要点击次数: 49

全文下载次数: 62

中文摘要:

摘要目的：基于决策树构建维持性血液透析（MHD）患者发生透析失衡综合征的预测模型，为临床及时采取防控策略提供参考。方法：选取426例MHD患者，收集患者一般情况、病历资料，统计透析失衡综合征发生情况，利用决策树构建MHD患者发生透析失衡综合征早期预警模型，采用受试者工作特征（ROC）曲线评价其预测效能。结果：426例MHD患者在透析治疗期间发生透析失衡综合征49例（11.50%），其中发生于透析第1～4次者占61.22%（30/49）；决策树共3层，节点有11个，共提取出6条分类规则，超滤量、低蛋白血症、尿素氮（BUN）、S-100B蛋白、髓鞘碱性蛋白（MBP）是MHD患者透析失衡综合征风险的预测变量；决策树模型曲线下面积（AUC）为0.936（95%CI：0.908～0.957），灵敏度为95.92%，特异性为78.78%，模型预测效能良好。结论：MHD患者存在透析失衡综合征风险，S-100B蛋白、BUN、MBP、超滤量、低蛋白血症是其预测变量，而决策树模型可对高危人群进行有效识别和风险预测。

英文摘要:

Abstract Objective: To construct a prediction model for dialysis disequilibrium syndrome (DDS) in maintenance hemodialysis (MHD) patients based on a decision tree algorithm, providing a reference for timely clinical prevention and control strategies. Methods: A total of 426 MHD patients were selected. General information and medical records were collected, and the incidence of DDS was analyzed. An early warning model for DDS in MHD patients was constructed using a decision tree, and its predictive efficiency was evaluated by the receiver operating characteristic (ROC) curve. Results: Among the 426 MHD patients, 49 cases (11.50%) developed DDS during dialysis treatment, with 61.22% (30/49) occurring during the 1st to 4th dialysis sessions. The decision tree model consisted of 3 layers and 11 nodes, from which 6 classification rules were extracted. Ultrafiltration volume, hypoproteinemia, blood urea nitrogen (BUN), S-100B protein, and myelin basic protein (MBP) were identified as predictive variables for DDS risk in MHD patients. The area under the ROC curve (AUC) for the decision tree model was 0.936 (95% CI: 0.908~0.957), with a sensitivity of 95.92% and a specificity of 78.78%, indicating good predictive performance. Conclusion: MHD patients are at risk for DDS, with S-100B protein, BUN, MBP, ultrafiltration volume, and hypoproteinemia serving as key predictive variables. The decision tree model can effectively identify and predict the risk in high-risk populations.