急性胆源性胰腺炎患者内镜逆行胰胆管造影术后胰腺真菌感染的早期病原学证据及预测模型构建

Early etiological evidence and predictive model construction of pancreatic fungal infection after endoscopic retrograde cholangiopancreatography in acute biliary pancreatitis

  • 摘要:
    目的 探究急性胆源性胰腺炎患者行内镜逆行胰胆管造影术(ERCP)后胰腺真菌感染的菌种分布特征,并分析易感因素。
    方法 选取就诊于中国人民解放军空军军医大学第一附属医院的急性胆源性胰腺炎行ERCP后胰腺真菌感染的患者31例为感染组,另选取同期急性胆源性胰腺炎行ERCP后未感染患者155例为未感染组。收集2组患者临床资料,采用多因素Logistic回归模型分析急性胆源性胰腺炎患者行ERCP后胰腺真菌感染的影响因素; 采用广义估计方程(GEE)模型验证多因素Logistic回归分析结果; 采用受试者工作特征(ROC)曲线评估预测价值,并应用卡方自动交互检验(CHAID)算法基于影响因素构建决策树模型,构建胰腺真菌感染风险评分表,根据截断值进行低危和高危风险分层,并比较ERCP后胰腺真菌感染发生率。
    结果 感染组入院至胰腺感染确诊时间为11(8, 14) d, 共分离出38株真菌菌株,白色念珠菌为检出最多的菌株。186例患者术后送检胰液、胆汁标本各1份,胰腺感染确诊至胰液标本培养阳性时间为5(3, 6) d, 胰腺感染确诊至胆汁标本培养阳性时间为6(4, 7)d。感染组胰液和胆汁标本培养阳性占比均高于未感染组,差异有统计学意义(P < 0.05)。胰液标本培养结果阳性占比高(OR=10.413, 95%CI: 2.059~52.667)、胆汁标本培养结果阳性占比高(OR=7.468, 95%CI: 1.603~34.789)、患有糖尿病(OR=4.567, 95%CI: 1.090~19.131)、手术时间延长(OR=1.052, 95%CI: 1.012~1.093)、术后胆汁引流不充分(OR=4.951, 95%CI: 1.304~18.802)、术前总胆红素(TBil)高(OR=1.221, 95%CI: 1.100~1.354)是急性胆源性胰腺炎患者行ERCP后胰腺真菌感染的独立影响因素(P < 0.05); 术前白蛋白高(OR=0.860, 95%CI: 0.773~0.958)是急性胆源性胰腺炎患者行ERCP后胰腺真菌感染的独立保护因素(P < 0.05)。ROC曲线显示,联合检测的诊断效能、灵敏度和特异度高于单独检测。决策树模型选择胰液和胆汁标本培养结果、术后胆汁引流不充分3个风险因素作为模型的节点,其中胰液标本培养结果是最重要的预测因子,模型的分类准确率为83.3%。高危患者胰腺真菌感染发病率高于低危患者,差异有统计学意义(P < 0.05)。
    结论 急性胆源性胰腺炎行ERCP后胰腺感染主要检出白色念珠菌,胰液标本培养结果阳性比例高、胆汁标本培养结果阳性比例高、患有糖尿病、手术时间延长、术后胆汁引流不充分、术前TBil高是急性胆源性胰腺炎患者行ERCP后胰腺真菌感染的独立危险因素,术前白蛋白高是独立保护因素。

     

    Abstract:
    Objective To explore the fungal species distribution features in pancreatic infections in patients with acute biliary pancreatitis after endoscopic retrograde cholangiopancreatography (ERCP) and analyze the predisposing factors.
    Methods A total of 31 patients with pancreatic fungal infection after ERCP for acute biliary pancreatitis admitted to the First Affiliated Hospital of Air Force Medical University of Chinese People's Liberation Army were selected as infection group, and another 155 patients without infection after ERCP for acute biliary pancreatitis during the same period were selected as non-infection group. Clinical data of patients in both groups were collected. A multivariate Logistic regression model was used to analyze the influencing factors of pancreatic fungal infection after ERCP for acute biliary pancreatitis patients. The generalized estimating equation (GEE) model was used to validate the results of the multivariate Logistic regression analysis. The receiver operating characteristic (ROC) curve was used to evaluate the predictive value. The chi-square automatic interaction detector (CHAID) algorithm was applied to construct a decision tree model based on the influencing factors, and a risk score table for pancreatic fungal infection was established. Risk stratification (low-risk and high-risk groups) was performed according to the cut-off value, and the incidence of pancreatic fungal infection after ERCP was compared.
    Results In the infection group, the time from admission to the diagnosis of pancreatic infection was 11 (8, 14) days. A total of 38 fungal strains were isolated, with Candida albicans being the most frequently detected strain. Among the 186 patients, one pancreatic juice specimen and one bile specimen were collected for culture after the operation. The time from the diagnosis of pancreatic infection to a positive pancreatic juice specimen culture was 5 (3, 6) days, and the time from the diagnosis of pancreatic infection to a positive bile specimen culture was 6 (4, 7) days. The proportions of positive cultures in pancreatic juice and bile specimens in the infection group were higher than those in the non-infection group, with statistically significant differences (P < 0.05). A high proportion of positive pancreatic juice specimen cultures (OR=10.413, 95%CI, 2.059 to 52.667), a high proportion of positive bile specimen cultures (OR=7.468, 95%CI, 1.603 to 34.789), diabetes (OR=4.567, 95%CI, 1.090 to 19.131), prolonged operation time (OR=1.052, 95%CI, 1.012 to 1.093), inadequate postoperative bile drainage (OR=4.951, 95%CI, 1.304 to 18.802), and high preoperative total bilirubin (TBil) (OR=1.221, 95%CI, 1.100 to 1.354) were independent influencing factors for pancreatic fungal infection after ERCP for acute biliary pancreatitis patients (P < 0.05). High preoperative albumin (OR=0.860, 95%CI, 0.773 to 0.958) was an independent protective factor for pancreatic fungal infection after ERCP for acute biliary pancreatitis patients (P < 0.05). The ROC curve showed that the diagnostic efficacy, sensitivity, and specificity of combined detection were higher than those of single detection. The decision tree model selected three risk factors (results of pancreatic juice and bile specimens and inadequate postoperative bile drainage) as the nodes of the model. Among them, the culture result of the pancreatic juice specimen was the most important predictive factor, and the classification accuracy rate of the model was 83.3%. The incidence of pancreatic fungal infection in high-risk patients was higher than that in low-risk patients, with a statistically significant difference (P < 0.05).
    Conclusion Candida albicans is the main fungus detected in pancreatic infections after ERCP for acute biliary pancreatitis. A high proportion of positive pancreatic juice specimen cultures, a high proportion of positive bile specimen cultures, diabetes, prolonged operation time, inadequate postoperative bile drainage, and high preoperative TBil are independent risk factors for pancreatic fungal infection after ERCP for acute biliary pancreatitis patients, while high preoperative albumin is an independent protective factor.

     

/

返回文章
返回