摘要
目的 探讨脓毒症患者急性肾损伤(AKI)的影响因素,并构建预测模型。方法 回顾性选取脓毒症患者200例,根据是否发生AKI将其分为AKI组(80例)和非AKI组(120例),收集所有患者一般资料、临床资料、入院时实验室检查结果及炎症因子水平并进行组间比较。采用多因素logistic回归分析探讨影响脓毒症患者发生AKI的因素。构建Nomogram预测模型,采用Bootstrap法对模型进行内部验证,绘制校准曲线及受试者工作特征(ROC)曲线评价模型,绘制决策曲线评价临床净获益。同时另选取172例脓毒症患者进行外部验证。结果 AKI组BMI、急性生理学和慢性健康状况评分系统(APACHE)Ⅱ评分、肿瘤坏死因子(TNF)-α、IL-6、C反应蛋白(CRP)、降钙素原(PCT)水平及合并糖尿病患者比例均高于非AKI组,PLT计数低于非AKI组(P<0.05)。多因素logistic分析结果显示,BMI、合并糖尿病、PLT、TNF-α、IL-6、CRP、PCT均为脓毒症患者发生AKI的影响因素(P<0.05)。经内部验证,Nomogram预测模型预测脓毒症患者发生AKI相关因素的一致性指数为0.902(95%CI 0.885~0.974),ROC曲线结果显示,该列线图模型预测ROC曲线下面积(AUC)为0.902(95%CI 0.875~0.969,P<0.05)。决策曲线分析结果显示,当含炎症因子的模型预测脓毒症患者发生AKI阈值在0.1~1.0区间时,可提供附加临床净获益。外部验证结果显示,该模型预测敏感度为84.06%,特异度为85.29%。结论 脓毒症患者AKI的发生受BMI、合并糖尿病、PLT、TNF-α、IL-6、CRP等因素影响,含炎症因子的预测模型可提高预测脓毒症患者发生AKI的准确性。
Objective To investigate the influencing factors of acute kidney injury(AKI) in sepsis patients,and establish a prediction model.Methods A total of 200 patients with sepsis were retrospectively selected and divided into AKI group(80 cases) and non-AKI group(120 cases) according to whether AKI occurred.General data,clinical data,laboratory examination results on admission and levels of inflammatory factors of all patients were collected and compared between groups.Multivariate logistic regression analysis was used to investigate the factors affecting the occurrence of AKI in patients with sepsis.The Nomogram prediction model was constructed,and the model was internally verified by Bootstrap method.Calibration curve and receiver operating characteristic(ROC) curve evaluation model were drawn,and decision curve was drawn to evaluate clinical net benefit.At the same time,172 patients with sepsis were selected for external verification.Results BMI,Acute Physiology and Chronic Health Evaluation(APACHE)Ⅱ score,tumor necrosis factor(TNF)-α,IL-6,C-reactive protein(CRP),procalcitonin(PCT) levels and the proportion of patients with diabetes in AKI group were higher than those in non-AKI group,and PLT count was lower than that in non-AKI group(P<0.05).Multivariate logistic analysis showed that BMI,diabetes mellitus,PLT,TNF-α,IL-6,CRP and PCT were all influencing factors for AKI in sepsis patients(P<0.05).After internal verification,the consistency index of Nomogram prediction model in predicting AKI occurrence in patients with sepsis was 0.902(95%CI 0.885-0.974),ROC curve results showed that the AUC predicted by the nomogram model was 0.902(95%CI 0.875-0.969,P<0.05).The results of decision curve analysis showed that when the model containing inflammatory factors predicted the AKI threshold in sepsis patients in the range of 0.1 to 1.0,it could provide additional clinical net benefit.The external verification results showed that the prediction sensitivity of this model was 84.06% and the specificity was 85.29%.Conclusion The occurrence of AKI in sepsis patients is affected by BMI,diabetes mellitus,PLT,TNF-α,IL-6,CRP and other factors.The prediction model containing inflammatory factors can improve the accuracy of predicting AKI in sepsis patients.
作者
李雅琳
李东风
孙振康
王静
Li Yaling;Li Dongfeng;Sun Zhenkang;Wang Jing(Department of critical care medicine,Fuyang People’s Hospital,Fuyang 236000,China)
出处
《临床内科杂志》
CAS
2023年第11期754-757,共4页
Journal of Clinical Internal Medicine
基金
阜阳市卫生健康委科研立项课题(FY2021-131)。
关键词
脓毒症
急性肾损伤
炎症因子
预测模型
Sepsis
Acutekidney injury
Inflammatory factors
Predictive model