摘要
目的:分析急性胰腺炎(AP)并腹腔感染患者的影响因素,并构建预测模型,为 AP 并腹腔感染的早期防治提供参考.方法:搜集 2021 年 1 月至 2023 年 12 月我院收治的 103 例AP 患者为对象,依据患者有无腹腔感染分为感染组及无感染组.明确AP 并腹腔感染的独立影响因素,同时构建预测模型,ROC曲线分析预测模型的对腹腔感染的预测价值.结果:103 例 AP 患者,腹腔感染占比 32.04%(33/103).AP 并腹腔感染患者致病菌以革兰阴性菌为主,主要致病菌大肠埃希菌、肺炎克雷伯对常用抗菌药物耐药率较高.Logistic 回归结果示,年龄、住院时间、入院时 APACHEⅡ评分、CRP、PCT、NK细胞总量是AP 并腹腔感染的独立影响因素(P＜0.05).ROC 曲线分析结果示模型 ROC 曲线的AUC为0.843(95%CI为0.792～0.896)、敏感性为61.36%、特异性为92.06%、约登指数为0.534,Hom-ser-Lemeshow检验P 值为0.267,一致率为96.12%,结论:AP 并腹腔感染风险较高,AP 并腹腔感染与患者年龄、住院时间、入院时APACHEⅡ评分、CRP、PCT、NK细胞总量有关,Logistic 预测模型可有效预测AP 并发腹腔感染的风险.
Abstract
Objective:To analyze the influencing factors of patients with acute pancreatitis(AP)com-plicated by abdominal infection and to construct a prediction model,providing a reference for early prevention and treatment of AP with abdominal infection.Methods:Data from 103 AP patients admitted to Hefei Second People's Hospital between January 2021 and December 2023 were collected.Patients were divided into infec-ted and non-infected groups based on the presence of abdominal infection.Independent influencing factors for AP with abdominal infection were identified,and a prediction model was constructed.The predictive value of the model for abdominal infection was assessed using ROC curve analysis.Results:Among the 103 AP pa-tients,32.04%(33/103)had abdominal infections.The predominant pathogens were Gram-negative bacteri-a,with Escherichia coli and Klebsiella pneumoniae showing high resistance to common antibiotics.Logistic re-gression identified age,length of hospital stay,APACHE Ⅱ score at admission,CRP,PCT,and total NK cell count as independent influencing factors for AP with abdominal infection(P＜0.05).ROC curve analysis showed the model's AUC was 0.843(95%CI0.792～0.896),with a sensitivity of 61.36%,specificity of 92.06%,Youden index of 0.534,Homser-Lemeshow test P value of 0.267,and concordance rate of 96.12%.Conclusion:The risk of abdominal infection in AP patients is relatively high and is associated with age,length of hospital stay,APACHE Ⅱ score at admission,CRP,PCT,and total NK cell count.The logis-tic prediction model effectively predicts the risk of abdominal infection in AP patients.
基金项目
2022年安徽省重点研究与开发计划项目(2022e07020058)
国家科技期刊平台
NSTL
万方数据