Abstract:

Objective　To construct a prediction model of vancomycin-related nephrotoxicity in patients with methicillin-resistant Staphylococcus aureus (MRSA) bloodstream infection. Methods　The clinical data of 128 patients with MRSA bloodstream infection treated with vancomycin from January 2019 to January 2023 in Tianjin Medical University General Hospital were retrospectively analyzed. There were 66 males and 62 females, aged (61.47±10.25) years. According to whether the patients developed vancomycin-related acute kidney injury (AKI), they were divided into an AKI group (32 cases) and a non-AKI group (96 cases). The gender, age, body mass index (BMI), shock, white blood cell count (WBC), procalcitonin, hypersensitive C-reactive protein (hs-CRP), Sequential Organ Failure Assessment (SOFA) score, Acute Physiology and Chronic Health Evaluation Ⅱ (APACHEⅡ) score, combined underlying diseases (diabetes mellitus, hypertension, and coronary heart disease), serum creatinine, use of nonsteroidal anti-inflammatory drugs, use of aminoglycosides, use of vasoactive drugs, types of nephrotoxic drugs used, glomerular filtration rate (GFR), treatment dose, administration interval, duration of treatment, cumulative dose, area under the curve (AUC) of vancomycin, and other data were compared between the two groups. Logistic regression equation was used to analyze the risk factors of vancomycin-related nephrotoxicity in patients with MRSA bloodstream infection. A risk nomogram model of vancomycin-related nephrotoxicity was constructed based on the risk factors, and the constructed model was validated and its predictive efficacy was evaluated. Independent sample t test, Mann-Whitney U test, χ² test, and Hosmer-Lemeshow test were used. Results　The age, serum creatinine level, vancomycin AUC, and proportions of number of nephrotoxic drugs ≥2 and GFR ≤60 ml/min in the AKI group were higher than those in the non-AKI group (all P<0.05). Logistic regression analysis showed that age ≥60 years old, serum creatinine ≥95.42 μmol/L, number of nephrotoxic drugs ≥2, GFR ≤60 ml/min, and vancomycin AUC ≥30 g/L were all independent risk factors for AKI in patients with MRSA bloodstream infection treated with vancomycin (all P<0.05). Hosmer-Lemeshow goodness of fit test showed that the nomogram model had a good agreement in predicting the risk of AKI in patients with MRSA bloodstream infection treated with vancomycin (χ²=3.571, P=0.672). The internal verification results of Bootstrap method showed that the C-index of the nomogram prediction model was 0.785 (95%CI 0.678-0.889), indicating that the model had good differentiation. The receiver operating characteristic curve (ROC) results showed that the AUC (95%CI), sensitivity, and specificity of the nomogram risk model for predicting AKI in patients with MRSA bloodstream infection treated with vancomycin were 0.859 (0.618-0.979), 94.50%, and 78.30%, respectively (all P<0.001). Conclusions　Age ≥60 years old, serum creatinine ≥95.42 μmol/L, number of nephrotoxic drugs ≥2, GFR ≤60 ml/min, and vancomycin AUC ≥30 g/L are all independent risk factors for AKI in patients with MRSA bloodstream infection treated with vancomycin. The risk nomogram model based on the above risk factors has a high predictive value for AKI in patients with MRSA bloodstream infection treated with vancomycin.

Key words:

Methicillin-resistant Staphylococcus aureus, Bloodstream infection, Vancomycin, Acute kidney injury, Nomogram

摘要：

目的　构建耐甲氧西林金黄色葡萄球菌（MRSA）血流感染患者万古霉素相关肾毒性预测模型。方法　回顾性分析2019年1月至2023年1月天津医科大学总医院收治的128例接受万古霉素治疗的MRSA血流感染患者临床资料。其中男66例，女62例；年龄（61.47±10.25）岁。根据患者是否发生万古霉素相关急性肾损伤（AKI）分为AKI组（32例）和非AKI组（96例），比较两组患者性别、年龄、体质量指数（BMI）、休克、白细胞计数（WBC）、降钙素原、超敏C反应蛋白（hs-CRP）、序贯器官功能衰竭评估（SOFA）评分、急性生理学与慢性健康状况评分系统Ⅱ（APACHEⅡ）评分、合并基础病（糖尿病、高血压、冠心病）、血肌酐、使用非甾体抗炎药、使用氨基糖苷类药物、使用血管活性药物、使用肾毒性药物数量、肾小球率滤过率（GFR）、治疗剂量、给药间隔、治疗时间、累积剂量、万古霉素曲线下面积（AUC）等资料。采用logistic回归方程分析MRSA血流感染患者万古霉素相关肾毒性的危险因素。基于危险因素构建万古霉素相关肾毒性的风险列线图模型，并对构建的模型进行验证及预测效能评估。采用独立样本t检验、Mann-Whitney U检验、χ²检验、Hosmer-Lemeshow检验。结果　AKI组年龄、血肌酐水平、万古霉素AUC及使用肾毒性药物数量≥2个、GFR≤60 ml/min占比均高于非AKI组（均P<0.05）。logistic回归分析结果显示，年龄≥60岁、血肌酐≥95.42 μmol/L、使用肾毒性药物数量≥2个、GFR≤60 ml/min、万古霉素AUC≥30 g/L均是影响接受万古霉素治疗的MRSA血流感染患者发生AKI的独立危险因素（均P<0.05）。Hosmer-Lemeshow拟合优度检验结果显示，列线图模型预测接受万古霉素治疗的MRSA血流感染患者发生AKI风险的一致性良好（χ²=3.571，P=0.672）。Bootstrap法内部验证结果显示，列线图预测模型C指数为0.785（95%CI 0.678～0.889），表明该模型具有较好的区分度。受试者操作特征曲线（ROC）结果显示，列线图风险模型预测接受万古霉素治疗的MRSA血流感染患者发生AKI的AUC（95%CI）、灵敏度、特异度分别为0.859（0.618～0.979）、94.50%、78.30%（均P<0.001）。结论　年龄≥60岁、血肌酐≥95.42 μmol/L、使用肾毒性药物数量≥2个、GFR≤60 ml/min、万古霉素AUC≥30 g/L均是影响接受万古霉素治疗的MRSA血流感染患者发生AKI的独立危险因素。基于上述危险因素构建的风险列线图模型对接受万古霉素治疗的MRSA血流感染患者发生AKI具有较高的预测价值。

关键词:

耐甲氧西林金黄色葡萄球菌, 血流感染, 万古霉素, 急性肾损伤, 列线图