摘要
目的基于细胞因子、免疫因子构建老年脓毒症预后的Nomogram预测模型。方法选取2022年3月—2023年6月西安大兴医院350例老年脓毒症患者,统计住院期间预后情况,将死亡患者纳入死亡组,存活患者纳入存活组。比较死亡组、存活组临床资料,通过全变量Logistic回归、逐步Logistic回归和Lasso-Logistic回归模型分析老年脓毒症患者预后为死亡的因素,根据影响因素构建预后为死亡的Nomogram预测模型,并验证其预测效能和临床效用。结果不同预后患者年龄、机械通气、入院当天APACHEⅡ评分、SOFA评分、外周血TNF-α、CD3^(+)、IL-6、CD4^(+)、CRP、CD4^(+)/CD8^(+)、IL-10、CD19^(+)B细胞水平比较,差异有统计学意义(P<0.05);Lasso回归筛选表1中P<0.05的指标,得出预测因素:机械通气、年龄、入院当天APACHEⅡ评分、SOFA评分、外周血TNF-α、CD3^(+)、IL-6、CD4^(+)、CD4^(+)/CD8^(+)、CD19^(+)B细胞水平;全变量Logistic回归、逐步Logistic回归和Lasso-Logistic回归模型显示,机械通气、年龄、入院当天APACHEⅡ评分、SOFA评分、外周血TNF-α、CD3^(+)、IL-6、CD4^(+)、CD4^(+)/CD8^(+)、CD19^(+)B细胞水平因素均为老年脓毒症患者预后不良的因素(P<0.05),其中Lasso-Logistic回归模型拟合和预测效果较好,并构建患者发生预后不良的Nomogram预测模型,ROC曲线、DCA曲线证实该模型具有良好预测效能和正向净收益。结论老年脓毒症预后不良因素包括机械通气、年龄、入院当天APACHEⅡ评分、SOFA评分、外周血TNF-α、CD3^(+)、IL-6、CD4^(+)、CD4^(+)/CD8^(+)、CD19^(+)B细胞水平,绘制的Nomogram预测模型具有良好预测价值。
Objective To establish a Nomogram model to predict the prognosis of senile sepsis based on cytokines and immune factors.Methods 350 elderly patients with sepsis in our hospital from March 2022 to June 2023 were selected,and the prognosis during hospitalization was analyzed.Dead patients were included in the death group,while surviving patients were included in the survival group.Compare the clinical data of the death group and the survival group,analyze the factors that affect the prognosis of elderly sepsis patients as death through full variable logistic regression,stepwise logistic regression,and Lasso logistic regression models,and construct a Nomogram prediction model for prognosis as death based on the influencing factors,and verify its predictive efficacy and clinical utility.Results There were statistically significant differences in age,mechanical ventilation,APACHEⅡscore,SOFA score,peripheral blood TNF-α,CD3^(+),IL-6,CD4^(+),CRP,CD4^(+)/CD8^(+),IL-10,CD19^(+)B cells of patients with different prognosis(P<0.05).The index of P<0.05 in Lasso regression screening Table 1 was as follows:mechanical ventilation,age,APACHEⅡscore on admission day,SOFA score,peripheral blood TNF-α,CD3^(+),IL-6,CD4^(+),CD4^(+)/CD8^(+),CD19^(+)B cells.All variable logistic regression,stepwise logistic regression,and Lasso logistic regression models showed that mechanical ventilation,age,APACHEⅡscore on admission day,SOFA score,peripheral blood TNFα,CD3^(+),IL-6,CD4^(+),CD4^(+)/CD8^(+),and CD19^(+)B cell level factors are all factors contributing to poor prognosis in elderly sepsis patients(P<0.05).Among them,the Lasso Logistic regression model has good fitting and prediction effects,and a Nomogram prediction model for poor prognosis in patients has been constructed.The ROC curve and DCA curve confirm that the model has good predictive performance and positive net income.Conclusion Adverse prognostic factors for elderly sepsis include mechanical ventilation,age,APACHEⅡscore on admission day,SOFA score,and peripheral blood TNFα,The Nomogram prediction model drawn at the levels of CD3^(+),IL-6,CD4^(+),CD4^(+)/CD8^(+),and CD19^(+)B cells has good predictive value.
作者
王九菊
闫杜娟
曹楠婧
李娜
武红梅
吕海丽
张哲
WANG Jiuju;YAN Dujuan;CAO Nanjing;LI Na;WU Hongmei;LYU Haili;ZHANG Zhe(Department of Laboratory,Xi'an Daxing Hospital,Xi'an Shaanxi 710016,China)
出处
《中国急救复苏与灾害医学杂志》
2024年第5期607-611,627,共6页
China Journal of Emergency Resuscitation and Disaster Medicine
基金
急救与创伤研究教育部重点实验室(海南医学院)开放基金(编号:Grant.KLET-202211)。
