摘要
目的 探讨重症患者万古霉素相关急性肾损伤(AKI)的危险因素,构建风险预测列线图模型并进行验证。方法 选取2021年1月至2023年1月我院接受万古霉素治疗的重症患者209例,根据是否发生万古霉素相关AKI分为AKI组(52)和非AKI组(157)。应用Logistic回归模型分析导致重症患者万古霉素相关AKI发生的独立危险因素,并根据影响因素构建重症患者万古霉素相关AKI发生风险预测模型,采用受试者工作特征曲线(ROC)及Hosmer-Lemeshow检验验证该模型的区分度和拟合度。结果 209例重症患者发生万古霉素相关AKI 52例(24.88%);AKI组中高尿酸血症、呼吸衰竭、休克、联合应用利尿剂、用药疗程>14 d、肾小球滤过率(GFR)≤60 ml/min的患者及血肌酐、万古霉素谷浓度均高于非AKI组(P<0.05);Logistic结果显示,万古霉素谷浓度、联合应用利尿剂、GFR≤60 ml/min、血肌酐增高、高尿酸血症是重症患者发生万古霉素相关AKI的独立危险因素(P<0.05);根据独立危险因素建立重症患者发生万古霉素AKI的预测模型方程,采用Bootstrap法对预测模型进行内部验证,Calibration curve显示,模型拟合度好(Hosmer-Lemeshow χ^(2)=0.227,P=0.172)。ROC曲线结果显示,该预测模型预测重症患者发生AKI的AUC为0.924(95%CI 0.871~0.960),敏感度93.55%,特异度83.85%,约登指数0.774。结论 重症患者中万古霉素谷浓度增高、联合应用利尿剂、GFR≤60 ml/min、血肌酐增高、高尿酸血症导致万古霉素相关AKI发生风险增加,基于上述因素构建的风险评估模型对重症患者发生AKI具有良好的区分度和拟合度。
Objective To analyze the risk factors of vancomycin-related acute kidney injury(AKI)in critically ill patients,constructing and validating a risk prediction nomogram model.Methods A total of 209 critically ill patients who received vancomycin from January 2021 to January 2023 were selected.They were divided into AKI group(52)and non-AKI group(157)according to whether vancomycin-related AKI occurred.Application of Logistic regression model to analyze independent risk factors for vancomycin associated AKI in critically ill patients.A risk prediction model was constructed based on these factors,and its discriminative ability and goodness-of-fit were evaluated using receiver operating characteristic(ROC)curves and the Hosmer-Lemeshow test.Results Among the 209 critically ill patients,52(24.88%)developed vancomycin-related AKI.The AKI group had significantly higher rates of hyperuricemia,respiratory failure,shock,combined diuretic use,treatment duration>14 days,and glomerular filtration rate(GFR)≤60 ml/min,as well as higher serum creatinine and vancomycin trough concentrations compared to the non-AKI group(P<0.05).Logistic regression results showed that trough concentration of vancomycin,combined use of diuretics,GFR≤60 ml/min,elevated blood creatinine,and hyperuricemia were independent risk factors for vancomycin related AKI in critically ill severe patients(P<0.05).The prediction model equation of vancomycin AKI in critically ill patients was established based on independent risk factors.Bootstrap method was used to verify the prediction model internally.The calibration curve indicated good model fit(Hosmer-Lemeshow χ^(2)=0.227,P=0.172).ROC curve results showed that the AUC of the predictive model for AKI in severe patients was 0.924(95%CI 0.871~0.960),the sensitivity was 93.55%,the specificity was 83.85%,and the Yodon index was 0.774.Conclusion The increased valley concentration of vancomycin,combined use of diuretics,GFR≤60 mL/min,elevated blood creatinine,and hyperuricemia in critically ill patients lead to an increased risk of vancomycin related AKI.The risk assessment model constructed based on these factors has good discrimination and fitting ability for the occurrence of AKI in critically ill patients.
作者
丁亮
李茹
晋鹏
Ding Liang;Li Ru;Jin Peng(Department of Pharmacy,Xi an Chang an Hospital,Xi an,Shaanxi 710018;Department of Pharmacy,Xi an Daxing Hospital,Xi an,Shaanxi 710003;Xianyang First People's Hospital,Xianyang,Shaanxi 712099,China.)
出处
《四川医学》
2025年第5期508-513,共6页
Sichuan Medical Journal
基金
陕西省教育厅专项科研计划项目(编号:20JK0215)。
