【作者】 杜娟；

【导师】 厉有名；

【作者基本信息】 浙江大学 ， 内科学（专业学位）， 2013， 博士

【摘要】 目的：在临床实践中,溃疡性结肠炎(UC)和克罗恩病(CD)的鉴别十分重要,然而常常由于临床表现不典型、模棱两可的内镜检查结果和影像表现以及肠镜下活检的深度不够,使得鉴别UC和CD成为一个难题。IBD的诊断指标包括一些生物学标记,本研究将相互独立的血清标记物作为参数进行整合,通过统计学工具和方法构建了一个用于鉴别仅结肠损伤的UC和CD的诊断模型,并进一步检验该诊断模型的效能。并通过meta分析探讨克罗恩病(CD)与肠结核(ITB)内镜表现和组织病理学特征,为两者的鉴别诊断提供依据。研究对象及方法：2006年2月至2011年2月,采用回顾性分析的方法,收集了来自浙江大学医学院附属第一医院的140名UC住院患者和174名CD住院患者的资料。首次住院治疗的这段时间收集周围静脉血液样本,根据所测的血清标记物的结果,我们构建了两个逻辑回归模型。为了评估最终拟合模型的有效性,我们还用了受试者工作特征(ROC)来评估该诊断模型的预测效果,ROC曲线下面积(AUC)用来评估其准确度。检索Pubmed、EBSCO、Web of science、中国生物医学文献数据库(the Cochrane Library and Chinese Biomedicine Database)、维普、万方数据库等数据库,时间1995年1月到2013年6月发表的关于克罗恩病和肠结核内镜表现和组织病理学特征的文献,由2名评价员独立采用QUADAS(Quality Assessment of Diagnostic AccuracyStudies)工具进行质量评价,应用Meta-disc1.4和stata12.0做异质性检验,根据异质性检验结果选择相应的效应模型合并,评价其敏感性、特异性、似然比和诊断比值比,描绘SROC曲线并计算曲线下面积,对于研究间存在较高异质性,用Meta回归分析找异质性来源,并做敏感性分析。结果：我们利用BIC来挑选出与疾病状态相关的预测变量。在无效模型中,利用BIC选出了预测变量Alb,TC,Plt以及Alb:Plt.在备择模型中,同样的方法选出了新的预测变量GPDA以及另加的传统预测变量TCa,两两相互作用项Alb:Plt,Alb:GPDA, TCa:TC和Plt:GPDA.CD/UC指数(CUI)结果为CUl=1.901+0.425Alb-3.324TC一7.444TCa+0.018Plt+0.087GPDA-0.0007Alb:Plt-0.004Alb:GPDA+1.839TC:TCa+0.003Plt:GPDA。UC患者的CUI大于CD患者的,CUI>0则递增性倾向于UC的诊断,而CUI<0则对应CD诊断的可能性更高。无效模型和备择模型的AUCs的平均值分别为0.66(95%置信区间：0.59-0.72)和0.73(95%置信区间：0.67-0.80)。截断点对应的灵敏度和特异度,备择模型中分别为0.55和0.80,而无效模型中分别为0.46和0.79。meta分析共纳入15篇文献,包括1271个研究对象,其中克罗恩病671个,肠结核600个。统计结果显示：以克罗恩病为阳性对照,其敏感性、特异性、阳性似然比、阴性似然比、诊断比值比和SROC曲线下面积分别为：阿弗他溃疡0.39,0.80,2.20,0.75,3.34,0.7252；肠腔狭窄0.35,0.72,1.37,0.89,1.54,0.4626；鹅卵石征0.28,0.96,5.25,0.79,7.05,0.6212；跳跃征0.61,0.57,1.52,0.71,2.52,为0.6420；纵形溃疡0.42,0.94,6.18,0.65,11.02,0.7898；微肉芽肿0.42,0.69,1.42,0.82,2.08,0.5768。而以肠结核为阳性对照,环形溃疡0.43,0.88,3.66,0.64,7.07,0.7515；回盲部扩张0.38,0.91,3.98,0.74,5.98,0.8404；干酪样坏死0.42,1.00,17.10,0.69,38.25,0.9976；肉芽肿0.73,0.63,1.78,0.50,4.83,0.7268；融合肉芽肿0.41,0.99,17.74,0.60,29.86,0.9705；每个切片肉芽肿大于5个0.26,0.94,4.45,0.80,5.52,0.5702；粘膜下肉芽肿0.30,0.90,2.92,0.76,4.00,0.6559；不成比例的粘膜下炎症0.52,0.75,2.84,0.59,4.52,0.6679；肉芽组织0.31,0.92,3.68,0.72,5.23,0.8723；ulcers lined by histiocyte0.42,0.95,6.33,0.55,12.52,0.9248。结论：根据血清标记物的检测结果构建的CUI可成为克罗恩病和溃疡性结肠炎的鉴别诊断的辅助工具,特别是在临床病史不明,内镜和影像学特征异常,活组织检查模棱两可的情况下。诊断性meta分析结果提示阿弗他溃疡、肠腔狭窄、鹅卵石征、跳跃征、纵形溃疡、微肉芽肿有助于诊断克罗恩病,而同时环形溃疡、回盲部扩张、干酪样坏死、肉芽肿、融合肉芽肿、每个切片肉芽肿大于5个、粘膜下肉芽肿、不成比例的粘膜下炎症和肉芽组织有助于诊断肠结核。因此,内镜结合病例组织活检的特异性表现对于鉴别克罗恩病和肠结核意义重大。更多还原

【Abstract】 Aim:To derive a model capable of distinguishing UC from CD, based on the measurement of independent serum markers.To find out endoscopic and histological features for differentiation between Crohn’ disease(CD) and intestinal tuberculosis(ITB) by meta-analysis.Methods:140UC and174CD hospitalized patients whose lesions were only in the colon at The First Affiliated Hospital, College of Medicine, Zhejiang University were recuited into this study. A panel of serum markers was measured for each patient. Then, a discrimination model using the Bayesian information criterion (BIC) was developed. The receiver operating characteristic (ROC) is used to evaluate the performance of the model, and the area under the ROC curve (AUC) is used as a measure of evaluating the accuracy of the model.A search of Pubmed, web of science, EBSCO, the Chinese Biomedicine Database, Weipu and Wanfang database was undertaken from1995January to2013June. All the English and Chinese literatures for differential diagnosis of CD and ITB by endoscopy and histological features were collected. Study quality was assessed and data extraction were performed. The software of Meta-disc1.4and Stata12.0were used for calculated the sensitivity, specificity, likelihood ratio, diagnostic odds ratio, summary receiver operating characteristic curve and area under the curve(AUC) to evaluate the differentiation value of CD and ITB. Meta-regression and sensitive analyses were used when there was heterogeneity.Results:Serum albumin (Alb), total cholesterol (TC), total calcium (TCa), platelet (Plt), glycyl proline dipeptidyl aminopeptidase (GPDA) and the product terms Alb:Plt, Alb:GPDA, TCa:TC, and Plt:GPDA were selected into the diagnosis model using BIC. The resulting CD/UC Index (CUI) is CUI=1.901+0.425Alb-3.324TC-7.444TCa+0.018Pit+0.087GPDA-0.0007Alb:Plt-0.004Alb:GPDA+1.839TC:TCa+0.003Plt:GPDA, with CUI>0incrementally favored a diagnosis of UC, while CUI<0corresponded to a higher likelihood of a diagnosis of CD. An average value of the AUC for the CUI model is0.73(95%confidence interval:0.67-0.80).Fifteen studies were included in the meta-analysis,with a total of1271cases with CD671cases and ITB600cases.The sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, diagnostic odds ratio and AUC are: CD as positive controll, Aphthous ulcers0.39,0.80,2.20,0.75,3.34,0.7252; Stricture0.35,0.72,1.37,0.89,1.54,0.4626; Cobblestone appearance0.28,0.96,5.25,0.79,7.05,0.6212; Skip lesions0.61,0.57,1.52,0.71,2.52,0.6420; Longitudinal ulcers0.42,0.94,6.18,0.65,11.02,0.7898; Microgranulomas0.42,0.69,1.42,0.82,2.08,0.5768. ITB as positive control, Transverse ulcers0.43,0.88,3.66,0.64,7.07,0.7515; Patulous ileocecal valve0.38,0.91,3.98,0.74,5.98,0.8404; Caseous necrosis0.42,1.00,17.10,0.69,38.25,0.9976; Granuloma0.73,0.63,1.78,0.50,4.83,0.7268; Confluent granulomas0.41,0.99,17.74,0.60,29.86,0.9705; Granulomas>5per section0.26,0.94,4.45,0.80,5.52,0.5702; Granuloma in submucosa0.30,0.90,2.92,0.76,4.00,0.6559; Disproportionate submucosal inflammation0.52,0.75,2.84,0.59,4.52,0.6679; Granulation tissue0.31,0.92,3.68,0.72,5.23,0.8723; ulcers lined by histiocyte0.42,0.95,6.33,0.55,12.52,0.9248.Conclusions:The CUI, derived from commonly available serum biomarkers, could try to differentiate ulcerative colitis from Crohn’s disease in patients with lesions only in the colon and unclear clinical features as a new menthod.Meta-analysis shows that aphthous ulcers, stricture, cobblestone appearance, skip lesions, longitudinal ulcers and microgranulomas are helpful to diagnosis of CD, while transverse ulcers, patulous ileocecal valve, caseous necrosis, granuloma, confluent granulomas, granulomas>5per section, granuloma in submucosa, disproportionate submucosal inflammation, granulation tissue and ulcers lined by histiocyte can help to diagnosis of ITB. So the endoscopy and histological features is contribute to differential diagnosis of CD and ITB.更多还原