中国实用内科杂志2026，Vol.46Issue(6)：514-520,528,8.DOI:10.19538/j.nk2026060113
基于Lasso特征筛选的噬血细胞综合征合并急性肾损伤患者死亡风险logistic预测模型构建
Construction of a logistic prediction model for mortality risk in patients with hemophagocytic syndrome complicated by acute kidney injury based on Lasso feature selection
摘要
Abstract
Objective To identify 28-day mortality risk factors in patients with hemophagocytic syndrome(HPS)complicated by acute kidney injury(AKI)and to develop a predictive model.Methods Clinical data of 345 HPS-AKI patients admitted to Beijing Friendship Hospital,Capital Medical University,from January 2016 to December 2021 were retrospectively collected.Patients were grouped by 28-day outcome:survival group(225 cases)and death group(120 cases).Least absolute shrinkage and selection operator(LASSO)regression was used for dimensionality reduction,followed by multivariable logistic regression to build a nomogram.The 500-sample Bootstrap was used for internal validation,and calibration curves,area under the receiver-operating-characteristic curve(AUC),Brier score,and decision-curve analysis(DCA)were used to assess the performance.Results Among the 345 patients,the 28-day mortality rate was 34.7％.LASSO identified three key predictors:hemoglobin level,non-renal sequential organ failure assessment(SOFA)score,and Kidney Disease:Improving Global Outcomes(KDIGO)stage.The resulting nomogram demonstrated excellent discrimination(AUC=0.847,95％CI,0.804-0.890),which remained robust after Bootstrap correction(AUC=0.843).Calibration showed close agreement between predicted and observed risk(Dxy=0.693),with a Brier score of 0.147.DCA indicated meaningful clinical net benefit across threshold probabilities of 10％-80％.Conclusion Hemoglobin level,non-renal SOFA score,and KDIGO stage are risk factors for 28-day mortality in patients with HPS complicated by AKI.The predictive tool developed in this study offers a robust quantitative basis for prognosis assessment and risk stratification in HPS patients complicated by AKI,enabling clinicians to identify high-risk individuals early and optimize therapeutic strategies.关键词
噬血细胞综合征/急性肾损伤/预后/预测模型Key words
hemophagocytic syndrome/acute kidney injury/prognosis/predictive model分类
医药卫生引用本文复制引用
林瑾,赵梦雅,段美丽..基于Lasso特征筛选的噬血细胞综合征合并急性肾损伤患者死亡风险logistic预测模型构建[J].中国实用内科杂志,2026,46(6):514-520,528,8.基金项目
首都卫生发展专项(2024-2-1194) （2024-2-1194）
