【作者】 姚津剑；

【导师】 林菊生；

【作者基本信息】 华中科技大学 ， 消化内科， 2010， 博士

【摘要】 目的：慢性乙型肝炎的临床转归包括乙肝病毒携带者肝、硬化、肝癌、乙型重型肝炎,它是一组有明确病因的复杂多基因疾病,应用SNP-MaP(snp microarray and pooling)完成全基因组扫描,寻找慢性乙型肝炎重症化的易感位点,并扩大样本验证其易感性位点。方法：根据慢性乙型肝炎的临床转归的表型将样本分为乙肝病毒携带者,肝硬化,肝癌,乙型重型肝炎。应用affymatrix SNP 6.0扫描各组中100个样本随机挑选86个组成DNA pool,并完成3个生物学重复。分别运用两种不同的统计学方法进行分析,挑选差异性最强的SNP位点。在乙肝病毒携带者234例,肝硬化300例,肝癌300例,乙型重型肝炎171例进行个体验证。结果：通过严格的统计学挑选,一共挑选7个SNP进行个体的基因分型。发现两个与肝硬化易感性的多态性位点分别是转录因子TBX5 SNP位点rs3825214在于对照组相比(共显性模型P=0.024,显性模型P=0.008 OR1.8(1.1-2.9))和膜蛋白SCUBE3 SNP位点rs 3800385在于对照组相比(显性模型P=0.0076 OR 1.6(1.1-2.32)).且SNP位点rs3800385在肝癌在于对照组相比较的共显性模型中具有统计学意义结论：从分子流行病学的角度证明了TBX5和SCUBE3在肝硬化易感性中起重要作用。SCUBE3与肝癌的易感性相关。目的：前瞻性随访乙型重型肝炎病人长期预后,并对影响乙型重型肝炎长期预的危险因素进行分析,从中筛选出对预后有明显影响的因素,构建预后模型并与现有的预后模型进行对比其预后判断效能,在一个独立回顾性的样本进行验证,判断对预后预测的准确性。方法：从2007年6月到2008年12月前瞻性从4个医院进行样本收集254例乙型重型肝炎病人,同样的时间段内在回顾性中心收集乙型重型肝炎病例43例,记录所有的临床资料对上述病人情况进行随访。记录所有的住院期间临床资料。依据随访的结果,分析生化指标和并发症对死亡的影响,同时探讨肝硬化重型肝炎程与非肝硬化重型肝炎临床指标的差异。应用单因素分析和多元回归模型,建立预后模型,运用ROC曲线下面积比较本研究的模型与现有的重型肝炎预后模型预测能力。然后在回顾性的中心验证预后模型的预测效能。结果：前瞻性中心平均年龄为40.58(14-69)肝硬化比例为27.16%,男性比例为88.89%,而回顾性中心平均年龄为43.3(20-66)肝硬化比例为27.9%,男性比例为90.7%。随访截止时间为2009年12月,平均随访时间为19.5(0-30)。总的生存率为34.68%,其中前瞻性中心的生存率为35.03%而回顾性的中心生存率为32.58%。分析发现生存组和死亡组中年龄、凝血酶原时间、凝血酶原活动度、纤维蛋白原、总胆红素、直接胆红素、血清γ谷氨酰转肽酶、乳酸脱氢酶、氯、肌酐、MELD评分、肝硬化、脾大、自发性腹膜炎、腹水、肝性脑病、肝肾综合症均有统计学差异(P<0.05)。而在肝硬化组和非肝硬化组中年龄、感染周数、凝血酶原时间、国际化标准比值、凝血酶原活动度、谷丙转氨酶、谷草转氨酶、胆固醇、肌酐、尿酸、HBV-DNA.MELD评分、死亡率上均有统计学差异(P<0.05),多因素回归分析构建预后模型P=eX／1+eX,X=(4.542+0.671*国际化标准比值+0.003*总胆红素mmol／L+0.034*肌酐mmol／L-0.025*氯mmol／L+0.445*肝硬化),在ROC曲线下面积比较,本研究模型ROC面积较MELD评分系统,Dhiman模型,Sun模型及西安标准曲线下面积显著性大于(P<0.05)，依次为(0.80,CI95%0.75-0.86),MELD(0.73,CI95%0.67-0.8),Sun(0.71,CI 95%0.64-0.7791,Dhiman模型(0.66,CI 95%0.61-0.71),西安标准(0.58,CI95%0.528-0.65)。重复在回顾性的中心进行验证发现本研究的模型ROC曲线下面积0.93大于MELD 0.87.结论：本研究模型在乙型重型肝炎预后预测上优于MELD等预后模型,乙型重型肝炎并发症不能作为预后模型,由于并发症因其发生率较低且出现较晚。肝硬化乙型重型肝炎用于预后模型构建是可行的。更多还原

【Abstract】 Objectives:the adverse outcomes of chronic HBV infection are a polygenic trait. We performed a genome-wide screen to identify progression related variants in the respective adverse outcomes including HBV carrier, cirrhosis, liver cancer and acute liver failure.Methods:The initial genome-wide screen by affymatrix SNP 6.0 was done in pooled DNA in randomly selected 86 individual from 100 samples among each outcome. SNPs fulfilling the selection criteria of the pooling stage were genotyped individually in this same sample and additional sample, of which totally included heath HBV carrier 234、cirrhosis 300、liver cancer 300 and acute liver failure 171.Results:we combined two different and stringent statistical approaches (a single-point method and, as a filter for the most significant SNPs, a sliding-window method) to select SNPs of high statistical confidence, finally 7 SNPs is genotyped individually, two genomic loci encoding TBX5 rs382514(additive model P=0.024, dominant model P=0.008 OR 1.8(1.1-2.9)) and SCUBE3 rs 3800385 (additive model P=0.02, dominant model P=0.0076 OR 1.6(1.1-2.32)) was significantly associated with cirrhosis. And SNP rs3800385 of SCUBE3 is significant associated with liver cancer in additive model.Conclusions:Two-stage genetic analysis from above experiments suggests a role for TBX5 and SCUBE3 in HBV-related cirrhosis. Objectives:To describe and analyze the clinical features, complications and follow-up outcomes of HBV-related acute liver failure.To establish prognositic model according to results of follow-up study and compare with current acute liver failure prognositic models such from MELD, Sun, Dhiman prognositic model.To validate accuracy of our model in a indenpent retrospective corhort.Methods:Prospective cohort study consisted of 4 hospital and retrospective cohort included 2 hospital,254 and 43 consecutive patients with acute liver failure admitted over the same 18-month period from the prospective cohort and retrospective cohort respectively.Detailed clinical, Laboratory data and complications collected and analyzed the significant diffence between survival vs. nosurvival group and cirrhosis vs. nocirrhosis group. Logistic regression was also carried out to build prognositic model. Compared our model with current prognostic models by ROC curve, and to validate its predicting accuracy in retrospective independent cohort.Results:The median age of the prospective and retrospective was 40.58 (14-69) and 43.3(20-66) respectively; Cirrhostic rate in prospective group and retrospective group is 27.16% and 27.9% respectively; the overall group was predominantly male 88.89% and 90.7% in prospective group and retrospective group respectively.The follow-up study time is averagely 19.5 (0-30) months from time discharge from hospital and the censored time is Dec,2009. Overall survival rate was 34.68%(103 of total 297), and 35.03% in prospective group and 32.58% in retrospective group Independent risk analysis in survival and nosurvivor goup indicated age, PT, PTA,Fb, Total Bilirubin, DB, GGT, Lactate dehydrogenase, Chlorine, Creatinine,MELD; and Complication of Splenomegaly, Peritonitis, Ascites, Hepatic encephalopathy, Hepatorenal syndrome, Cirrhosis was significant between survival group and death group, the difference of clinical characteristics also existed in cirrhostic and nocirrhostic patients.multiple logistic regression applied to create prognositic model, the model comes to P= eX/1+eX, (X=4.542+0.671*INR+0.003*TB mmol/L-0.025*chlorinemmol/L +0.034*creatinine mmol/L+0.445*Cirrhosis). From the ROC curves, it was found that ROC of our model (0.8, confidence interval 95% 0.75-0.86) is significantly superior to MELD (0.73, confidence interval 95% 0.67-0.8) which is better than Sun Model (0.71, confidence interval 95%0.64-0.77), Dhiman model (0.66, confidence interval 95% 0.61-0.71) and China model (0.58, confidence interval 95% 0.52-0.65), To validite the accuracy of our model in predicting prognosis,43 patients from restrospective group was analyzed, and compared our model with MELD scoring system which was better than other prognostic models in our study, the result displayed AUC of our model was 0.9368 compared to 0.8776 of MELDConclusions:our model is superior to MELD in predicting survival rate of HBV-related acute liver failure.complications of the HBV-induced acute liver can not be proposed as prognostic factors for its low incidence and end-stage occurrence. Consideration of cirrhosis in prognositc model building is feasible and necessary in China.Key Words:prognosis; hepatitis B virus; prognostic model; follow-up study更多还原