列线图预测模型在儿童尿路感染中的应用及诊断意义  

Application of nomogram prediction model in children with urinary tract infection and its diagnostic significance

作　　者：向燕君 冯兰英[1] 基鹏[2] 邓山鹰[3] 陈月梅[3] 宋亚莉 XIANG Yanjun;FENG Lanying;JI Peng;DENG Shanying;CHEN Yuemei;SONG Yali(Outpatient Department,West China Hospital,Sichuan University,Chengdu 610041,Sichuan,China;Department of Critical Medicine,West China Hospital,Sichuan University,Chengdu 610041,Sichuan,China;Department of Laboratory Medicine,West China Hospital,Sichuan University,Chengdu 610041,Sichuan,China)

机构地区：[1]四川大学华西医院门诊部,成都610041 [2]四川大学华西医院重症医学科,成都610041 [3]四川大学华西医院实验医学科,成都610041

出　　处：《临床检验杂志》2022年第8期595-599,共5页Chinese Journal of Clinical Laboratory Science

摘　　要：目的探讨列线图预测模型在儿童尿路感染中的应用及诊断意义。方法选取2019年1月至2021年8月四川大学华西医院就诊的1094例患者作为研究对象,分为>2岁组(共850例,诊断尿路感染334例)和≤2岁组(共244例,诊断尿路感染113例)。男童清洁中段尿培养菌落计数>10^(4)/mL,女童>10^(5)/mL判断为尿路感染。收集患者清洁中段尿,用激光法进行尿沉渣分析,并进行细菌培养及计数;收集EDTA-K_(2)抗凝全血,使用电阻抗法进行全血细胞分析;收集血清,用电化学发光法检测白介素-6和降钙素原,免疫比浊法检测C反应蛋白。使用受试者工作特征(ROC)曲线分析评价各指标的诊断效能,挑选P值<0.05的指标,使用logistic回归构建列线图预测模型。结果>2岁组入选指标为性别、尿白细胞计数、尿细菌计数、尿上皮细胞计数,ROC曲线下面积(AUC)分别为0.757、0.708、0.761、0.586(P均<0.001);≤2岁组入选指标为性别、血淋巴细胞计数、尿白细胞计数、尿细菌计数、尿上皮细胞计数,AUC分别为0.664、0.592、0.813、0.749、0.593(P均<0.05)。>2岁和≤2岁儿童的尿路感染诊断列线图模型C-index分别为0.872和0.857。对列线图进行内部重抽样,校正曲线提示模型符合度良好。结论尿液沉渣分析的白细胞计数和细菌计数,在儿童尿路感染诊断中有着很好的提示作用。利用尿液沉渣分析指标构建儿童尿路感染列线图预测模型,可快速进行儿童尿路感染诊断。Objective To explore the application and diagnostic significance of nomogram prediction model in children with urinary tract infection.Methods A total of 1094 patients who were treated in West China Hospital of Sichuan University from January 2019 to August 2021 were selected as the research objects and divided into 2 groups:>2 years old group(850 cases in total,including 334 cases diagnosed with urinary tract infection)and≤2 years old group(244 cases in total,including 113 cases diagnosed with urinary tract infection).The colony counts in midstream urine culture>10^(4)/mL in boys and>10^(5)/mL in girls were diagnosed as urinary tract infections.The middle urine was collected and urine sediment analysis was performed by laser method,and the colony numbers of bacterial culture in clean midstream urine were counted.The complete blood cell count was performed by electrical impedance method in EDTA-K_(2) anticoagulated whole blood.The levels of interleukin-6 and procalcitonin in serum were detected by electrochemiluminescence and C-reactive protein was detected by immunoturbidimetric method.Receiver operating characteristic(ROC)curve analysis was used to evaluate the diagnostic efficacy of each parameter.The parameter with P value<0.05 was selected to build a nomogram prediction model by logistic regression was used.Results The selected parameters in the group older than 2 years were gender,urinary white blood cell count,urinary bacteria count and urinary epithelial cell count.The area under the ROC curve(AUC)were 0.757,0.708,0.761,0.586,respectively(P<0.001).The selected parameters in the group less than or equal to 2 years old were gender,blood lymphocyte count,urine white blood cell count,urine bacteria count and urine epithelial cell count.The AUC were 0.664,0.592,0.813,0.749 and 0.593,respectively(all P<0.05).The C-index of the nomogram model for the diagnosis of urinary tract infection in groups of children older than and less than 2 years old were 0.872 and 0.857,respectively.By resampling the nomograms inte

关 键 词：儿童尿路感染 列线图 预测模型 

分 类 号：R446.5[医药卫生—诊断学]