急性白血病患者化疗后血流感染的病原菌分布及不良事件发生的危险因素分析和预测模型构建  

Distribution of pathogenic bacteria of bloodstream infection after chemotherapy in patients with acute leukemia and risk factors analysis of the occurrence of adverse events and prediction model construction

作　　者：李旺阳 付玉 杨艳萍[1] 林海[1] 樊红琼 刘秋菊[1] 高素君[1] 谭业辉[1] Li Wangyang;Fu Yu;Yang Yanping;Lin Hai;Fan Hongqiong;Liu Qiuju;Gao Sujun;Tan Yehui(Department of Hematology,the First Bethune Hospital of Jilin University,Changchun 130021,China)

机构地区：[1]吉林大学白求恩第一医院血液科,长春130021

出　　处：《白血病．淋巴瘤》2023年第7期394-399,共6页Journal of Leukemia & Lymphoma

基　　金：吉林省直卫生专项项目(JLSWSRCZX2020-0048)。

摘　　要：目的探讨急性白血病患者(AL)化疗后血流感染的病原菌分布情况、分析不良事件发生的危险因素,并构建预测不良事件发生的列线图模型。方法回顾性分析吉林大学白求恩第一医院2018年1月至2020年12月收治的313例AL合并血流感染患者的临床资料。分析AL患者化疗后血流感染的发生率、病死率及病原菌分布特点;比较不同临床病理特征的患者发生不良事件(死亡或感染性休克)的情况。采用非条件logistic二元回归模型多因素分析筛选AL化疗后合并血流感染患者发生不良事件的独立危险因素;使用R软件,建立预测不良事件发生的列线图模型;采用Hosmer-Lemeshow检验验证模型的预测效果。结果 313例AL患者中,总病死率为4.2%(13/313),血流感染全因病死率为3.5%(11/313)。313例中,革兰阴性菌感染254例(81.1%),主要为大肠埃希菌115例(45.3%)、肺炎克雷伯菌80例(31.5%)和铜绿假单胞菌29例(11.4%),死亡10例(3.9%);革兰阳性球菌感染51例(16.3%),主要为链球菌属22例(43.1%)、葡萄菌属20例(39.2%)、屎肠球菌7例(13.7%),死亡0例;真菌感染8例(2.6%),其中热带念珠菌4例(1.3%)、近平滑念珠菌2例(0.6%)、光滑念珠菌1例(0.3%)、新型隐球菌1例(0.3%),死亡3例(37.5%)。不同治疗阶段、危险分层、敏感抗生素使用时机、发热总时长,以及化疗方案中是否使用糖皮质激素、感染菌是否碳青霉烯类耐药、白血病是否缓解的患者不良事件发生率比较,差异均有统计学意义(均P<0.05)。logistic二元回归分析结果显示,化疗方案中使用糖皮质激素、发热总时长≥7 d、敏感抗生素使用时机≥24 h、感染菌碳青霉烯类耐药均为AL化疗后合并血流感染患者发生不良事件的独立危险因素(均P<0.05)。建立AL化疗后合并血流感染患者不良事件发生的列线图预测模型,经校准与验证列线图模型具有较好的校准度和区分度。结论 AL患者化疗后血流感染病原Objective To investigate the distribution of pathogenic bacteria of bloodstream infection after chemotherapy in patients with acute leukemia(AL),to analyze the risk factors for the occurrence of adverse events and to construct a nomogram model to predict the occurrence of adverse events.Methods The clinical data of 313 AL patients with bloodstream infection who were admitted to the First Hospital of Jilin University from January 2018 to December 2020 were retrospectively analyzed,and the incidence,fatality and distribution characteristics of pathogenic bacteria after chemotherapy in AL patients were analyzed;the occurrence of adverse events(death or infectious shock)in patients with different clinicopathological characteristics were compared.Unconditional logistic binary regression model multifactor analysis was used to screen independent risk factors for the occurrence of adverse events in AL patients with bloodstream infection after chemotherapy;the nomogram model for predicting the occurrence of adverse events was developed by using R software;the Hosmer-Lemeshow test was used to verify the predictive effect of the model.Results Of the 313 AL patients,the overall fatality rate was 4.2%(13/313),the all-cause fatality rate of bloodstream infection was 3.5%(11/313).Of the 313 cases,254 cases(81.1%)were Gram-negative bacteria infection,mainly including 115 cases(45.3%)of Escherichia coli,80 cases(31.5%)of Klebsiella pneumoniae,and 29 cases(11.4%)of Pseudomonas aeruginosa,and 10 cases(3.9%)died;51 cases(16.3%)were Gram-positive cocci infection,mainly including 22 cases(43.1%)of Streptococcus spp.,20 cases(39.2%)of Staphylococcus spp.,7 cases(13.7%)of Enterococcus faecalis,and 0 case died;8 cases(2.6%)were fungal infection,including 4 cases(1.3%)of Candida tropicalis,2 cases(0.6%)of Candida subsmoothis,1 case(0.3%)of Candida smooth,1 case(0.3%)of new Cryptococcus,and 3 cases(37.5%)died.The differences in the occurrence rates of adverse events were statistically significant when comparing different treatment stage,ri

关 键 词：急性白血病 血流感染 病原菌分布 危险因素 预测模型 

分 类 号：R733.71[医药卫生—肿瘤]