论文部分内容阅读
目的:分析感染性心内膜炎（IE）住院患者发生急性肾损伤（AKI）的危险因素，构建预测模型，并探讨该预测模型的预测价值。方法:回顾分析2010年1月至2020年1月在青岛大学附属医院确诊为IE的402例成人住院患者的临床资料，根据是否发生AKI分为AKI组和非AKI组。比较两组患者的临床资料〔包括性别、年龄、有无糖尿病、基础估算肾小球滤过率（eGFR）、入院时实验室指标、累及瓣膜数量和位置、是否并发脓毒症、住院用药以及是否行心脏手术〕和相关预后；采用多因素Logistic回归分析筛选IE住院患者发生AKI的危险因素，构建预测模型，并采用受试者工作特征曲线（ROC）分析该模型的预测价值。结果:最终共有290例IE患者纳入分析，其中非AKI患者198例，AKI患者92例，AKI发生率为31.7%；92例AKI患者中，AKI 1期46例（占50.0%），AKI 2期、3期46例（占50.0%）。与非AKI组相比，AKI组患者年龄更大〔岁：64（55，71）比55（46，63）〕，基础eGFR更低（mL·minn -1·1.73 mn -2：64.6±13.6比82.9±19.5），糖尿病患者比例和脓毒症发生率更高（16.3%比8.6%，38.0%比13.1%），需要更频繁地使用血管紧张素转换酶抑制剂/血管紧张素Ⅱ受体拮抗剂（ACEI/ARB）、利尿剂和非甾体抗炎药（NSAIDs；25.0%比15.2%，82.6%比63.1%，58.7%比24.2%），尿常规异常（血尿或蛋白尿）的比例更常见（35.9%比22.7%），血培养阴性率更高（73.9%比51.5%），革兰阳性（Gn +）球菌感染率和心脏手术率更低（22.8%比40.4%，60.9%比81.8%），差异均有统计学意义（均n P<0.05）；而两组性别、累及瓣膜的数量和位置、入院时实验室指标差异均无统计学意义。与非AKI组相比，AKI组患者住院病死率明显升高（30.4%比8.6%，n P<0.01）；且AKI 2期、3期患者住院病死率明显高于AKI 1期患者（43.5%比17.4%，n P<0.01）。多因素Logistic回归分析显示，基础eGFR低〔风险比（n HR）＝0.136，95%可信区间（95%n CI）为0.066～0.280〕、并发脓毒症（n HR＝6.100，95%n CI为2.394～15.543）、频繁使用NSAIDs（n HR＝2.990，95%n CI为1.184～7.546）和造影剂（n HR＝3.153，95%n CI为1.207～8.238）是IE住院患者发生AKI的独立危险因素（均n P<0.05）。以上述危险因素构建预测模型，并进行ROC曲线分析，结果显示，该模型预测AKI的ROC曲线下面积（AUC）为0.888（95%n CI为0.833～0.943，n P<0.01），敏感度为86.4%，特异度为80.9%。n 结论:基础eGFR低、并发脓毒症、频繁使用NSAIDs和造影剂是IE住院患者发生AKI的独立危险因素；由上述危险因素构建的预测模型对IE住院患者发生AKI有一定预测价值。“,”Objective:To analyze the risk factors of acute kidney injury (AKI) in hospitalized patients with infective endocarditis (IE), construct prediction model, and discuss its predictive value.Methods:The clinical data of 402 adult inpatients diagnosed with IE admitted to the Affiliated Hospital of Qingdao University from January 2010 to January 2020 were retrospectively analyzed. The patients were divided into the AKI group and the non-AKI group. The clinical data, such as gender, age, presence of diabetes, basic estimated glomerular filtration rate (eGFR), laboratory indexes at admission, involvement of valves, presence of sepsis, medication during hospitalization, surgery and outcome of the two groups were compared. Multivariate Logistic regression analysis was used to screen the risk factors of AKI in IE inpatients. A predictive model was constructed, and receiver operating characteristic (ROC) curve was used to analyze the predictive value of the model.Results:A total of 290 patients with IE were enrolled, including 198 non-AKI patients and 92 AKI patients. The incidence of AKI was 31.7%. Among the 92 AKI patients, 46 patients were at AKI stage 1 (50.0%), while 46 patients were at AKI stage 2 and stage 3 (50.0%). Compared with the non-AKI group, patients in the AKI group were older [years old: 64 (55, 71) vs. 55 (46, 63)], and had lower basic eGFR (mL·minn -1·1.73 mn -2: 64.6±13.6 vs. 82.9±19.5), higher proportion of diabetic and incidence of sepsis (16.3% vs. 8.6%, 38.0% vs. 13.1%), more frequent use of angiotensin converting enzyme inhibitors/angiotensin Ⅱ receptor antagonists (ACEI/ARB), diuretics and non-steroidal anti-inflammatory drugs (NSAIDs; 25.0% vs. 15.2%, 82.6% vs. 63.1%, 58.7% vs. 24.2%), more abnormal urine test results (hematuria or proteinuria, 35.9% vs. 22.7%), higher pathogen culture negative rate (73.9% vs. 51.5%), lower Gram positive (Gn +) cocci infection rate and surgery rate (22.8% vs. 40.4%, 60.9% vs. 81.8 %), with significant differences (all n P < 0.05). There were no significant differences in the gender, number and location of involved valves, and laboratory indexes at admission between the two groups. Compared with the non-AKI group, the inpatient mortality rate of the AKI group was higher (30.4% vs. 8.6%, n P < 0.01), and the inpatient mortality rate of patients with AKI stage 2 and stage 3 was significantly higher than that of patients with AKI stage 1 (43.5% vs. 17.4%, n P < 0.01). In multivariate Logistic regression analysis, the lower basic eGFR [hazard ratio ( n HR) = 0.136, 95% confidence interval (95%n CI) was 0.066-0.280], sepsis (n HR = 6.100, 95%n CI was 2.394-15.543), demand for NSAIDs (n HR = 2.990, 95%n CI was 1.184-7.546) and radiocontrast agent (n HR = 3.153, 95%n CI was 1.207-8.238) were independent risk factors for AKI in hospitalized patients with IE (all n P < 0.05). A prediction model was constructed based on the above risk factors, and ROC curve analysis showed that the area under the ROC curve (AUC) of prediction model for AKI was 0.888 (95% n CI was 0.833-0.943, n P < 0.01) with sensitivity of 86.4% and specificity of 80.9%.n Conclusions:In the IE-susceptible population, low basic eGFR, sepsis, the need for NSAIDs and contrast agent are independent risk factors to AKI. The predictive model constructed by the above risk factors has certain predictive value for the occurrence of AKI in the IE inpatients.