酸反流指数在儿童病理性胃食管反流诊断中的价值附12年4109例食管pH值监测结果分析

The evaluation of acid reflux index in the diagnosis of pathological gastroesophageal reflux in children:the analysis of data of esophageal pH monitoring of 4109 cases in 12 years

目的了解酸反流指数（RI）在儿童病理性胃食管反流（GER）诊断中的价值,探讨儿童病理性GER的诊断标准。方法对在浙江大学医学院附属儿童医院进行食管pH值监测的数据进行归纳和整理,并除外复查和少数未能完成监测的数据。根据Boix-Ochoa标准（金标准）综合评分〉11.99诊断病理性GER,对RI进行受试者工作特征（ROC）曲线分析,获得最佳诊断界点值（cutoff value）。通过McNemar卡方检验和Kappa检验,评价各种以RI为基准的病理性GER诊断标准所得结果的差异及其吻合度,并进一步评价各种标准诊断GER的敏感度、特异度等指标的差异。结果1994年9月至2006年5月有5000例患儿进行食管pH值监测,有效数据者4109例进入分析,其中男性2692例,女性1417例。新生儿476例,-1岁1553例,-3岁658例,-7岁837例,-12岁513例,-16岁72例。病种构成有13类,依次为：支气管哮喘366例,支气管炎468例,支气管肺炎385例,迁延性肺炎143例,慢性咳嗽184例,慢性咽炎29例,新生儿呼吸暂停102例,新生儿呕吐283例,新生儿肺炎105例,呕吐或伴腹痛1817例,食管裂孔疝140例,非心源性胸痛55例,营养不良32例。ROC曲线分析显示RI最佳诊断界点为2.9,其敏感度为90.4%,特异度为95.6%（假阳性率〈5.0%）。ROC曲线下面积为0.981（95%CI：0.977-0.984）,sx-为0.002,用于诊断病理性GER有统计学意义（P=0.000）。分别以RI≥4.0%（RI 4.0标准）、RI≥2.9%（RI 2.9标准）为病理性GER诊断标准,以手术确诊的140例食管裂孔疝患儿进行分析,与Boix-Ochoa标准进行比较。RI 2.9标准的总体吻合度高于RI 4.0标准（κ系数分别为0.892、0.715,P均=0.000）。除了特异度,RI 2.9标准诊断病理性GER的敏感度、准确性和阴性预测值均明显高于RI 4.0标准,而阳性预测值两者相近。结论RI 2.9标准诊断病理性GER有显著意义,与RI 4.0标准相比,有助于提高儿童病理性GER诊断的敏感度和准确性。

Objective The aim of the present study was to evaluate the diagnostic value of acid reflux index（RI） in the pathological gastroesophageal reflux（GER） in children and to discuss the diagnostic criteria of GER.Methods The patients performed by esophageal pH monitoring in Children＇s Hospital of Zhejiang University School of Medicine were all enrolled in this study from 1994 to 2006,except for the cases repeated or unfinished the measurement.According to the Boix-Ochoa criteria,in which pathological GER was confirmed if Boix-Ochoa score larger than 11.99,receiver operating characteristic（ROC） of RI was analyzed and the best diagnostic cutoff value for pathological GER was obtained.The difference and goodness of fitness of different results from other diagnostic criteria of pathological GER based on all kinds of RI were compared by statistical method such as McNemar test and kappa test.The sensitivity and specificity among different diagnostic criteria were evaluated.Results A total of 4109 cases（male 2 692 and female 1 417） performed esophageal pH monitoring were enrolled in this study.Among them,476 cases were neonate,1 553 younger than 1 year,658 younger than 3 years,837 younger than 7 years,513 younger than 12 years,and 72 younger than 16 years.There were thirteen kinds of diseases,including asthma（366 cases）,bronchitis（468）,pneumonia（385）,persist pneumonia（143）,chronic cough（184）,pharyngitis（29）,neonate apnea（102）,neonate vomiting（283）,neonate pneumonia（105）,vomiting and/or upper abdominal pain（1817）,hiatal hernia（140）,non-cardiac chest pain（55）,and malnutrition（32）.ROC analysis indicated that the diagnostic cutoff value of RI was 2.9,and the sensitivity for pathological GER was 90.4%,while the specificity was 95.6%（false positive rate lower than 5.0%）.The area under curve was 0.981,standard error was 0.002,and 95% confidence interval was from 0.977 to 0.984.It showed statistical significance for diagnoses of pathological GER（P=0.000）.According to the new diagnostic criteria of pathological GER for both RI 4.0 criteria（RI more than 4.0%） and RI 2.9 criteria（RI more than 2.9%）,the difference and goodness of fitness of results from Boix-Ochoa criteria were compared in 140 cases with hiatal hernia confirmed by surgery.The goodness of fitness from RI 2.9 criteria was significantly higher than that from RI 4.0 criteria（κ coefficient was 0.892、0.715,respectively,P all equal to 0.000）.The sensitivity,accuracy,and negative predictive value for diagnostic pathological GER were significantly higher from RI 2.9 criteria than that from RI 4.0 criteria except for the specificity,while positive predictive value was almost equal between the two criteria.Conclusions RI 2.9 criteria had of significance value in the diagnosis of pathological GER.It was helpful to improve the sensitivity and accuracy of GER diagnoses in children compared with RI 4.0 criteria.