基于MIMIC⁃Ⅳ数据库构建和验证预测脓毒症相关肝损伤患者住院死亡风险的列线图

Development and validation of a nomogram for predicting in⁃hospital mortality in patients with sepsis⁃associated liver injury: Based on MIMIC⁃Ⅳ database

摘要

目的：构建首次入住重症监护病房（ICU）的脓毒症相关肝损伤（SALI）患者住院死亡率的预测模型。方法:从重症监护医疗信息集市（MIMIC）Ⅳ 2.2数据库中收集SALI患者信息，按7∶3的比例随机分为训练组和验证组，采用单因素COX回归、最小绝对收缩和选择算子（LASSO）回归及多因素COX回归分析确定SALI患者住院死亡的独立影响因素,并构建列线图模型。采用受试者工作特征曲线(ROC)、校准曲线和决策曲线分析(DCA)评价列线图模型的预测能力、校准能力和临床效度。结果：本研究共纳入1 095例患者，随机分为训练组和验证组。多因素COX回归分析结果显示：体质量、红细胞计数、乳酸水平、急性生理评分Ⅲ（APSⅢ）、Charlson合并症指数（CCI）、合并周围血管疾病及活化部分凝血活酶时间（APTT）是ICU SALI患者住院死亡的独立危险因素，基于这些变量构建列线图模型。ROC曲线、校准曲线和DCA曲线结果表明，该预测模型具有较好的准确性、一致性及临床应用价值。结论:本研究构建并验证了一个良好预测SALI患者住院死亡率的列线图模型，这对优化SALI患者的临床决策和管理策略具有重要的潜力。

Abstract

Objective: To construct a nomogram to predict in‐hospital mortality in patients with sepsis‐associated liver injury (SALI) who are first admitted to the intensive care unit. Methods: We extracted data from the MIMIC‐Ⅳ 2.2 database and randomly split the cohort into training and validation sets in a 7∶3 ratio. Univariate Cox regression, least absolute shrinkage and selection operator (LASSO) regression, and multivariable Cox regression were performed to identify independent predictors of in‐hospital mortality. A nomogram was developed using these predictors and further validated in the validation set. ROC analysis, calibration plots, and decision curve analysis (DCA) were used to assess the model's discrimination, calibration, and clinical utility. Results: A total of 1 095 patients were included and randomly assigned to training and validation groups. Multivariable Cox regression identified seven independent predictors of in‐hospital mortality in ICU patients with SALI: body weight, red blood cell count, lactate, acute physiology score Ⅲ (APS Ⅲ), Charlson comorbidity index (CCI), peripheral vascular disease, and activated partial thromboplastin time (APTT). A nomogram was constructed from these variables. ROC analysis, calibration plots, and DCA demonstrated good discrimination, calibration, and clinical utility of the model. Conclusion: We developed and validated a nomogram that accurately predicts in‐hospital mortality in patients with SALI. The model showed good predictive performance and has significant potential to optimize clinical decision‐making and management strategies for this population.

关键词

脓毒症相关肝损伤 / 住院死亡风险 / 列线图 / 预测模型

Key words

Sepsis‐Associated Liver Injury / In‐Hospital Mortality / Nomogram / Prediction Model

参考文献

[1] SINGER M, DEUTSCHMAN C S, SEYMOUR C W, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis‐3)[J]. JAMA, 2016, 315(8): 801‐810.

[2] BEYER D, HOFF J, SOMMERFELD O, et al. The liver in sepsis: Molecular mechanism of liver failure and their potential for clinical translation[J]. Mol Med, 2022, 28(1): 84.

[3] YAN J, LI S, LI S L. The role of the liver in sepsis[J]. Int Rev Immunol, 2014, 33(6): 498‐510.

[4] DELLINGER R, LEVY M, RHODES A, et al. Surviving sepsis campaign: International guidelines for management of severe sepsis and septic shock, 2012[J]. Intensive Care Med, 2013, 39: 165‐228.

[5] World Health Organization. Global Report on the Epidemiology and Burden of Sepsis: Current Evidence, Identifying Gaps and Future Directions[R/OL]. (2020‐09‐09). Geneva: World Health Organization, 2020. https://www.who.int/publications/i/item/9789240010789#:~:text=In%20this%20report%20we%20highlight%20the%20public%20health,future%20directions%20and%20priorities%20in%20sepsis%20epidemiology%20research.

[6] LEI J C, ZHAI J, ZHANG Y, et al. Supervised machine learning models for predicting sepsis‐associated liver injury in patients with sepsis: Development and validation study based on a multicenter cohort study[J]. J Med Internet Res, 2025, 27: e66733.

[7] JOUFFROY R, PARFAIT P A, GILBERT B, et al. Relationship between prehospital modified Charlson Comorbidity Index and septic shock 30‐day mortality[J]. Am J Emerg Med, 2022, 60: 128‐133.

[8] HWANG S Y, KIM I K, JEONG D, et al. Prognostic performance of sequential organ failure assessment, acute physiology and chronic health evaluation Ⅲ, and simplified acute physiology score Ⅱ scores in patients with suspected infection according to intensive care unit type[J]. J Clin Med, 2023, 12(19): 6402.

[9] YEO H J, KIM T H, JANG J H, et al. Obesity paradox and functional outcomes in sepsis: A multicenter prospective study[J]. Crit Care Med, 2023, 51(6): 742‐752.

[10] GAO L, LIU J J, FAN Q C, et al. Association of obesity and mortality in sepsis patients: A meta‐analysis from observational evidence[J]. Heliyon, 2023, 9(9): e19556.

[11] ZHANG C C, REN J J, XU X, et al. Causal relationships between obesity‐related anthropometric indicators and sepsis risk: A Mendelian‐randomization study[J]. Front Nutr, 2024, 11: 1433754.

[12] GALLO G, DESIDERI G, SAVOIA C. Update on obesity and cardiovascular risk: From pathophysiology to clinical management[J]. Nutrients, 2024, 16(16): 2781.

[13] BELONCLE F M, RICHARD J C, MERDJI H, et al. Advanced respiratory mechanics assessment in mechanically ventilated obese and non‐obese patients with or without acute respiratory distress syndrome[J]. Crit Care, 2023, 27(1): 343.

[14] BUTT J H, PETRIE M C, JHUND P S, et al. Anthropometric measures and adverse outcomes in heart failure with reduced ejection fraction: Revisiting the obesity paradox[J]. Eur Heart J, 2023, 44(13): 1136‐1153.

基金资助

国家自然科学基金资助项目(82241039)