Establishment and value analysis of a clinical predictive model for patients with secondary hemophagocytic lymphohistiocytosis
10.3760/cma.j.cn114656-20250121-00045
- VernacularTitle:继发性嗜血综合征临床不良预后预警模型建立及价值分析
- Author:
Wuchao WANG
1
;
Siqi LIU
;
Hao GONG
;
Yuanyuan PEI
;
Jihong ZHU
Author Information
1. 北京大学人民医院急诊科，北京 100044
- Keywords:
Secondary hemophagocytic lymphohistiocytosis;
In-hospital mortality, Early warning model;
Predictive value
- From:
Chinese Journal of Emergency Medicine
2025;34(9):1251-1257
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To establish a clinical predictive model for poor clinical outcomes in patients with secondary hemophagocytic lymphohistiocytosis (sHLH) and to evaluate its clinical application value.Methods:Patients diagnosed with sHLH who met the study criteria and were initially admitted to the Emergency Department of Peking University People’s Hospital between September 2017 and December 2024 were enrolled. Clinical data were collected, and patients were categorized into a death group or a survival group based on clinical outcomes as the observational endpoint. Differences in clinical data between the two groups were compared. Univariate and multivariate logistic regression analyses were conducted to screen significant variables, and a predictive model nomogram was developed using the R programming language. The discriminative ability, calibration, and clinical utility of the predictive model were assessed using the receiver operating characteristic curve, net reclassification improvement index, calibration curve, and decision curve analysis. K-fold cross-validation was employed to evaluate the model's performance. The model was compared with the Acute Physiology and Chronic Health Evaluation Ⅱ (APACHE Ⅱ) score and the Sequential Organ Failure Assessment (SOFA) score.Results:A total of 116 cases were enrolled in the study, comprising 36 cases in the mortality group and 80 cases in the survival group. Multivariate logistic analysis identified age, platelet count, prothrombin time, total bilirubin, altered mental status, and cardiac involvement as factors significantly associated with clinical outcomes. Based on these factors, an early warning model for adverse clinical prognosis was established, and a corresponding nomogram was developed. The model demonstrated excellent discriminative ability, calibration, and clinical utility (AUC=0.950; Hosmer-Lemeshow test: χ2=2.5476, P=0.980; calibration curve: R 2=0.649, P=0.906), outperforming both the APACHE Ⅱ and SOFA scores in predicting adverse outcomes (both P<0.01). Conclusions:This study established an early warning model for adverse clinical prognosis in sHLH based on objective clinical data. The model aids in the clinical assessment of sHLH patients, facilitates early warning, and supports clinical decision-making for treatment.
