肺炎克雷伯菌感染脓毒症患者死亡风险 因素研究

Study on Risk Factors of Death in Patients with Pneumonia Caused by Klebsiella pneumoniae Infection and Sepsis

摘要: 目的：探讨肺炎克雷伯菌(Klebsiella pneumoniae, KP)感染脓毒症患者发生死亡的危险因素，并构建列线图预测模型。方法：回顾性分析2017年7月~2021年6月河北省秦皇岛市第一医院收治的KP感染脓毒症患者(n = 199)临床资料，根据患者28 d生存情况划分为存活组和死亡组，采用单因素、最小绝对收缩与选择算子(Least absolute shrinkage and selection operator, LASSO)回归法和多因素logistic回归分析引起KP感染脓毒症患者死亡的独立危险因素，并将199例病人按照随机抽样以7:3比例分为训练集(139例)和验证集(60例)，然后使用训练集根据独立危险因素建立KP感染脓毒症患者列线图风险预测模型。采用校准曲线和受试者操作特征曲线(ROC曲线)评估列线图预测模型的性能。结果：多因素logistic回归分析显示年龄、血红蛋白、血小板为KP感染脓毒症患者死亡的独立危险因素(OR = 1.0573、0.9608、0.9852, P < 0.05)。训练集AUC：0.874 (95% CI: 0.812~0.935)，验证集AUC：0.781 (95% CI: 0.664~0.897)。校准曲线和ROC曲线提示模型准确度、区分度良好。结论：基于年龄、血红蛋白、血小板、乳酸构建列线图模型对KP感染脓毒症患者的死亡风险的预测价值良好。

Abstract: Objective: To investigate the risk factors for mortality in patients with Klebsiella pneumoniae (KP) infection-associated sepsis and to construct a logistic regression model. Methods: A retrospective analysis was conducted on clinical data of 199 patients with KP infection-associated sepsis admitted to the First Hospital of Qinhuangdao, Hebei Province, from July 2017 to June 2021. Patients were divided into a survival group and a mortality group based on 28-day survival outcomes. Univariate, least absolute shrinkage and selection operator (LASSO) regression, and multivariate logistic regression were used to identify independent risk factors for mortality in KP infection-associated sepsis. The 199 patients were randomly divided into a training set (139 cases) and a validation set (60 cases) in a 7:3 ratio. A logistic regression model was established using the training set based on the identified risk factors. The performance of the model was evaluated using calibration curves and receiver operating characteristic (ROC) curves. Results: Multivariate logistic regression analysis revealed that age, hemoglobin, and platelet count were independent risk factors for mortality in patients with sepsis due to KP infection (OR = 1.0573, 0.9608, 0.9852, P < 0.05); training set AUC: 0.874 (95% CI: 0.812~0.935), validation set AUC: 0.781 (95% CI: 0.664~0.897). The calibration curve and ROC curve indicated that the model demonstrated good accuracy and discrimination. Conclusion: The linear regression model based on age, hemoglobin and platelet had good predictive value for the mortality risk of sepsis patients with KP infection.