摘要目的:研究肠杆菌血行感染脓毒症患者的预后危险因素。方法:收集2017年6月至2019年5月青岛市市立医院肠杆菌血行感染脓毒症患者的病历资料，记录患者的性别、年龄、入院科室、基础疾病、感染部位、病原学检查及治疗方案等基线资料，记录患者入院24 h内C反应蛋白（C-reactive protein，CRP）、血清前白蛋白（serum prealbumin，PA）等实验室检查指标，根据患者的生存情况分为生存组和死亡组，计算CRP与PA比值，急性生理及慢性健康评分（APACHEⅡ评分）和Pitt菌血症评分（Pitt bacteremia score，PBS），并根据血培养药物敏感试验结果分析初始经验治疗是否得当。采用logistic回归分析影响患者预后的危险因素，并绘制受试者工作特征曲线（ROC曲线）预测脓毒症患者不良预后的发生。结果:Logistic回归分析结果表明CRP（ OR=1.021， P<0.01）、CRP/PA（ OR=34.638， P<0.01）、产超广谱β内酰胺酶（ OR=0.244， P<0.01）、不适当的经验性抗菌药物治疗（ OR=0.156， P<0.01）、APACHEⅡ评分（ OR=1.436， P<0.01）、PBS评分（ OR=8.622， P<0.01）是影响患者不良预后的危险因素。多因素回归分析结果示CRP/PA（ OR=25.420， P<0.05）、不适当的经验性治疗（ OR=0.077， P<0.05）、APACHEⅡ评分（ OR=1.476， P<0.01）、PBS评分（ OR=12.042， P<0.01）是脓毒症患者死亡的独立危险因素（ P<0.05）。CRP/PA水平越高，PBS评分、APACHEⅡ评分越高提示患者预后越差，当CRP/PA≥0.89，PBS评分≥3.5，APACHEⅡ评分≥17.5，患者的死亡风险显著增加。此外，不适当的经验性治疗也是造成患者不良预后的关键因素。 结论:CRP/PA、PBS评分、APACHEⅡ评分、不适当的经验性治疗是影响肠杆菌血行感染脓毒症患者死亡的独立危险因素。PBS评分与APACHEⅡ评分一样能较好地预测患者不良预后及死亡风险，与APACHEⅡ相比，PBS评分更加简易、实用，可以被广泛推广使用。

abstractsObjective:To study the prognostic risk factors of Enterobacteriaceae bloodstream infection in patients with sepsis.Methods:The medical records of patients with sepsis caused by Enterobacteriaceae bloodstream infection in our hospital from June 2017 to May 2019 were screened. The gender, age, admission department, basic disease, infection site, etiology examination and treatment plan were recorded in detail. The survival and death groups were divided according to the patient's survival status. The ratio of C-reactive protein (CRP) to serum prealbumin (PA) was recorded within 24 h after admission. The acute physiological and chronic health scores (APACHEⅡ score) and Pitt bacteremia score (PBS score) were calculated within 24 h, and based on the results of blood culture drug sensitivity test to analyze whether the initial experience treatment was appropriate. Logistic regression analysis was used to analyze the risk factors affecting the prognosis of patients, and the receiver operating characteristic curve (ROC curve) was drawn to predict the occurrence of poor prognosis in patients with sepsis.Results:Logistic regression analysis showed that CRP ( OR=1.021, P<0.01), CRP/PA ( OR=34.638, P<0.01), extended-spectrum β-lactamase production ( OR=0.244, P<0.01), inappropriate empirical antibacterial treatment ( OR=0.156, P<0.01), APACHE Ⅱ score ( OR=1.436, P<0.01), and PBS score ( OR=8.622, P<0.01) were risk factors affecting patient's prognosis. Multivariate regression analysis showed that CRP/PA ( OR=25.420, P<0.05), inappropriate empirical treatment ( OR=0.077, P<0.05), APACHEⅡ score ( OR=1.476, P<0.01), PBS score ( OR=12.042, P<0.01) were independent risk factors for death in patients with sepsis ( P<0.05). The higher the CRP/PA level, PBS score and APACHEⅡ score, the worse the prognosis. When CRP/PA ≥0.89, PBS score ≥3.5, APACHEⅡ score ≥17.5, the patient's risk of death increased significantly. In addition, inappropriate empirical treatment was also a key factor in patients with poor prognosis. Conclusions:CRP/PA, PBS score, APACHE Ⅱ score, and inappropriate empiricaltreatment are independent risk factors affecting the prognosis of patients with enterobacter hemorrhagic infection with sepsis. The PBS score and APACHEⅡ score can better predict the poor prognosis and risk of death. Compared with APACHEⅡscore, the former is simpler and practical and can be widely used.