【作者】 傅锐斌；

【导师】 吴平生； 吴书林；

【作者基本信息】 南方医科大学 ， 内科学， 2009， 博士

【摘要】 研究背景和目的心房纤颤（房颤）是一种最常见的心律失常,由于其可以造成心排量减少、心室率过快过慢,可加重心肌缺血及恶化心功能,并且可引起左房血栓形成和全身栓塞并发症（其中脑栓塞形成发病率高）,极大地威胁人民健康,造成了巨大的社会经济负担。近年来,有关于房颤的多种治疗方法取得了巨大进步,特别是房颤射频消融治疗取得了明显进展,但是对房颤的发病机制,对引起房颤电解剖重构和引起房颤高凝的具体原因和机理等问题,仍没有非常明确的认识,这也是造成目前房颤及其并发症预防治疗效果不尽如人意的主要原因。最先引起人们认识到房颤与炎症相关的例子是心外科手术后同时出现的高炎症水平状态和高房颤发病率,目前越来越多的研究证据表明房颤与炎症密切相关,而炎症状态不依赖于引起房颤的基础心脏病和其他可能引起炎症的非房颤状态。这些研究证据包括房颤个体心房组织炎症细胞浸润增多、炎症因子表达增多,包括炎症患者血液炎症标记物水平显著增高,包括使用有抗炎作用多效性药物（如他汀类药物、ACEI/ARB类药物和鱼油等）以及糖皮质激素等治疗显著减少房颤发病的证据。也有一些研究得到不同的结果,尤其是有关房颤和炎症关系的研究绝大多数是横断面研究,只能说明两者是一种共存关系,而不能说明炎症究竟是不是引起房颤的病因,因此房颤就是一种炎症性疾病这种观点目前还不像“动脉硬化实质上是一种炎症性疾病”那么广为接受。肿瘤坏死因子-α（TNF-α）由单核细胞和巨噬细胞产生,同时它也是单核细胞和巨噬细胞的最强的旁分泌激活剂,单核细胞受到激活后可产生一系列的细胞因子,其中就包括白介素-6（IL-6）,IL-6又可刺激肝脏产生急性期反应物C-反应蛋白（CR-P）。CRP是公认的一系列炎性疾病的标记物。大规模临床研究显示,动脉粥样硬化斑块可发现有相当数量的TNF-α和CRP,这与心血管危险事件的发生是显著相关的。迄今为止,不少研究已经对CRP与房颤的关系进行了研究,但是极少有关房颤的研究关注TNF-α。TNF-α基因启动子区的一些SNPs已经被发现与循环TNF-α水平显著相关,并且研究还发现一些炎症性疾病,如慢性阻塞性肺病和冠心病,这些SNPs位点基因型分布与健康人有显著性差异或不同SNPs位点的基因型预示不同患这些炎性疾病的风险。然而,据我们所知,目前尚没有研究对TNF-α基因多态性,尤其是其启动子区SNPs现象和房颤的关系进行探讨。血栓性静脉炎和脓毒血症时血液处于高凝状态是炎症引起高凝状态的典型例子,事实上,心脑血管硬化和心衰的研究有非常多的证据表明,炎症水平是决定血栓性事件发生的重要原因之一。虽然房颤时高凝状态和血小板功能激活是非常明确的,是造成房颤血栓形成的重要原因之一（其他两个原因包括心房血流淤滞和内皮功能损害）,但是房颤高凝状态和血小板功能激活与炎症水平的关系,尽管对指导采取何种预防血栓形成的治疗策略意义很大,目前已有的研究较少涉及这方面,国内外几乎没有相关的研究报告。还有房颤时凝血功能和血小板功能的激活是由房颤本身所引起,还是由房颤合并的基础心脏病所引起,目前尚存在有较大争议。而关于特发性房颤、特别是孤立性房颤的内皮功能、凝血功能和血小板功能变化,目前还未见系统的研究,同时关于它们的抗栓治疗问题,2006ACC/AHA/ESC房颤指南的推荐还有不尽明确的地方。心房纤颤（房颤）时可出现明显的心房形态学改变、如心房扩大和二尖瓣关闭不全有关房颤时左心房结构和功能变化的研究报告较多,但关注房颤时右心房改变的研究相对较少。目前关注房颤心房扩大发生机制的研究非常少。而炎症是促进组织坏死的因素之一。例如动实验结果表明,过度表达TNF-α可导致动物出现心肌炎、心肌细胞坏死、心肌纤维化、收缩功能障碍、心肌肥厚和心室扩张等变化,并使其死亡率增高。本研究的主要目的包括:中国人TNF-α启动子区SNPs位点基因型的分布特点,研究房颤发病是否伴发明显的炎症标记物的变化,研究决定炎症上游因子TNF-α表达效率的基因启动子区各SNP位点不同基因型在房颤患者与非房颤对照个体分布情况是否不同,研究房颤发病的预测因子、尤其是不同基因型是否预示房颤发病风险高低;研究房颤本身是否伴发高凝状态和血小板激活,研究高凝状态和血小板激活可能的影响因素,尤其是炎症因素在这个过程起何种作用;研究特发性房颤和孤立性房颤内皮功能、凝血功能和血小板功能的变化特点及对抗栓治疗方案的指导;研究房颤患者本身引起的心脏、特别使双心房形态学改变及其与炎症因素的相互关系。研究方法入选房颤患者102例（作为研究组）和性别、年龄和危险因素匹配的对照组个体117例。房颤通过心电图或入院诊断进行确认。排除标准包括:高龄,合并明显动脉硬化性疾病,先心病、风心病、心肌病、左室肥厚和心功能不全、等器质性心脏病,糖尿病,慢性炎症性疾病,结缔组织疾病,60天内手术史和急危重症,肿瘤和重要器官功能衰竭。详细记录患者病史（尤其是房颤发病时间和特点）、体检情况(重点进行血压测定、心脏体检（包括心率、节律和脉率等各项和身高、体重、腹围等）、临床检验（常规化验、血生化指标、血凝血指标和炎症指标等）和临床检查结果（常规胸片、心电图和心脏超声等检查结果）。血脂、血糖、血肝肾功能等指标采用酶法测定,血纤维蛋白原（Fib）、部分凝血活酶时间、凝血酶原时间和凝血时间等采用酶法测定,血电解质测定采用电解质分析仪直接电极法、血CRP测定采用免疫浊度法、甲状腺功能测定采用放射免疫法。心脏超声检查常规测量包含以下指标:左房前后径（LAd）、右房上下径（RAd）、室间隔收缩期厚度（IVSd）、二尖瓣返流程度（MR）和左室射血分数（LVEF）。患者组进行射频消融治疗者在术前72小时进行经食道心脏超声检查。留取研究对象的全血和血清标本,部分标本用于临床送检,全血标本用于基因组DNA提取,血清冷冻于-70℃低温冰箱整批用于TNF-α、P-Selectin和vWF等细胞因子的检测。全血基因组DNA提取采用德国QIAGEN全血DNA小量提取试剂盒(QIAamp^?DNA Blood Mini Kits)提取,主要步骤包括细胞裂解、过柱吸附、杂质洗涤和DNA洗脱。采用琼脂糖凝胶电泳方法确定DNA质量,并采用光密度比对法确定其浓度。使用Prime Premier5软件设计TNF-α启动子区目的片段引物,并交公司进行合成。采用PCR方法(美国Promega PCR扩增试剂盒GoTaq^?PCR CoreSystemsⅠ)扩增基因组DNA的TNF-α启动子区目的片段,主要步骤包括变性、退火和延伸,并反复进行多个循环。PCR产物进行琼脂糖凝胶电泳并与DNAmarker进行比对,确认片段长度约为808bp。产物送公司进行专业纯化和核酸测序。测序结果使用Chromas软件进行读图,使用Pubmed网站Blast功能进行比对,确定每个个体基因型,尤其是各SNP位点的基因型。血清细胞因子检测采用免疫酶联吸附分析方法（ELISA）检测,其中血清TNF-α检测采用美国R&D公司TNF-αELISA检测试剂盒（Human TNF-α/TNFSF1AImmunoassay Kit）,可溶性P-选择素（sP-Selectin）检测采用武汉博士德公司P-Selectin检测试剂盒（Human P-Selectin ELISA Kit）,血管病性假血友病因子（vWF）检测采用上海太阳生物技术有限公司血管性血友病因子（vWF）含量测定试剂盒。主要步骤包括:结合抗原（样本）、结合酶标抗体、洗涤、加底物显色、终止反应并在酶标仪上进行光密度读数。使用软件Curve Expert 1.3对各指标进行标准曲线制作和标本光密度读数-浓度转换。统计学分析采用SPSS13.0软件进行,计量资料行正态性检验,正态性检验采用单样本K-S拟和优度检验（1-Sample K-S Test）。如符合正态分布,数据用x^-±S表示;如不符合正态分布,数据用中位数（median）（四分位数间距,IQR）表示。对于显著偏态分布的资料,进行对数转换后再进入协方差和回归分析。两组正态分布的计量资料均数比较采用Student t检验,两组中有一组以上是非正态分布的计量资料均数比较采用Mann-Whitney U检验。两组计量资料均数比较同时要控制其他对应变量有影响的因素时采用协方差分析,将可能有影响的因素作为协变量。计数资料表示为率,两个率的比较采用卡方检验（Fisher’s ExactTest）。两组正态分布资料相关分析采用Pearson相关;对于非双变量正态分布资料的相关分析,采用Spearman相关。两变量相关分析同时控制其他因素对这种相关关系的影响时采用偏相关分析,将其他因素作为控制变量。房颤独立预测因子估计采用logistic回归分析,房颤分组作为应变量,其他房颤组与对照组比较差异显著性P＜0.20的变量以及性别、年龄等一般情况纳入作为自变量。-308G/A不同基因型房颤风险推测采用卡方检验的“risk estimate”。重要的研究指标（包括炎症促凝指标Fibrinogen和sP-Selectin）的影响因素估计采用多元回归分析,将这些指标分别作为应变量,将性别、年龄、分组及单因素相关分析中分别与这些应变量相关关系显著性P＜0.20的计量变量和不同分组应变量指标差异显著性P＜0.20的分类变量作为自变量。研究结果房颤组女性40例,男性62例（60.8%）;对照组79例患者有性别记录,其中女性31例,男性48例（60.8%）,两组性别比例相比无显著性差异（P＞0.05）。房颤组年龄为56.7±10.9岁,对照组年龄为55.2±9.2岁,两组年龄相比无显著性差异（P＞0.05）。房颤组包括阵发性房颤87例,持续性房颤6例,长期持续性房颤9例。对照组包括房室结折返性室上性心动过速患者30例、窦性心律个体49例、健康人DNA样本38例。房颤组特发性房颤为61例,合并高血压患者为41例（40.1%）;对照组健康人+房室结折返性室上性心动过速患者72例,单纯性高血压者为45例（38.5%）,两组高血压合并率相比无显著性差异（P＞0.05）。TNF-α启动子区检测片段内-851、-575、-419和-244四个SNP位点未检测到少见基因型（或少见碱基）。房颤组和对照组之间-863、-857、-806、-782、-646、-572、-570、-376G/A和-238等各位点基因型分布均无显著性差异（P＞0.05）,这些位点少见基因型的总频率分别为24.9%、24.4%、7.2%、1.0%、1.9%、2.4%、1.0%、0.5%和2.4%。房颤组-308non-GG和-308GG分布频率分别为21%和79%,其中有1例是-308AA携带者,对照组-308GA和-308GG分布频率分别为10.1%和89.9%,未发现-308AA携带者,两个构成比相比有显著性差异（P=0.035）。-308non-GG与-308GG患房颤优势比（OR）为2.368（95%CI为1.078-5.205）。提示房颤患者-308位点基因型分布与对照组相比有显著性差异,房颤患者少见基因型-308non-GG携带率更高,并且-308non-GG提示更高的患房颤风险。房颤组和对照组的各项血液指标相比显示房颤组和对照组血TNF-α（pg/ml）的浓度分别为3.84（3.07,4.49）和2.85（2.28,3.60）,相比有显著性差异（P＜0.001）;血CRP水平分别为（mg/L）1.70（1,3.12）和1.00（0.95,2.10）,相比有显著性差异（P=0.015）;血Fib（g/L）水平分别为3.03（2.62,3.65）和2.58（2.28,3.02）,相比有显著性差异（P=0.002）;sP-Selectin（ng/ml）水平分别为26.5±5.8和21.3±3.8,相比有显著性差异（P＜0.001）;血GPT（U/L）水平分别为22.50（17.00,29.50）和18.00（13.50,23.50）,相比有显著性差异（P=0.002）。房颤组和对照组vWF（%）水平分别为107.31（90.30,128.58）和81.57（71.12,101.25）,相比无显著性差异（P=0.372）。其余血常规、血糖、血脂、肝肾功能、血电解质、甲状腺功能等血液指标房颤组和对照组相比均未见显著性差异（P＞0.05）。提示房颤患者比对照组有明显的炎症水平增高、血小板功能激活和凝血激活。阵发性房颤患者和对照组血Fib水平分别为3.20±0.84和2.77±0.60 g/L（P=0.005）,阵发性房颤患者和对照组血血sP-selectin水平分别为26.2±5.9和21.5±3.9 ng/mL（P＜0.001）。协方差仍然提示阵发性房颤Fib和sP-selectin水平分别较对照组显著增高。房颤组和对照组心动过缓发生率分别为29.7%和6.8%,相比有显著性差异（P＜0.001）。房颤组和对照组患者二尖瓣返流（MR）的发生率分别为52.3%和14.9%,相比有显著性差异（P＜0.001）。房颤组和对照组LAD（mm）分别为33.0（35.0,38.0）和28.0（30.0,33.0）,相比有显著性差异（P＜0.001）;RAD（mm）分别为47.0（43.0,51.0）和41.0（37.0,43.0）,相比有显著性差异（P＜0.001）;IVSD（mm）分别为10.6（10.3,10.8）和9.6（9.2,9.9）,相比有显著性差异（P＜0.001）。提示房颤心动过缓发生率增高和心脏形态学改变更明显。-308GG与-308non-GG个体血TNF-α浓度分别为3.37（2.72,4.44）和3.36（2.83,4.16）,两者相比无显著性差异（P=0.600）。Logistic回归分析显示,TNF-α、sP-selectin和LAd均是房颤的独立预测因子,其OR值分别为1.375（95%CI:1.024-1.846,P=0.035）、1.248（95%CI:1.121-1.389,P＜0.001）和1.321（95%CI:1.125-1.552,P=0.001）。炎症因素（TNF-α或CRP）与促凝血因素（sP-Selectin或Fib）之间均有显著的正相关关系,尤其是CRP与Fib之间的相关系数为0.560（P＜0.001）。以年龄、性别和分组（是否房颤）作为控制变量的偏相关分析显示TNF-α与Fib、TNF-α与sP-Selectin以及CRP与sP-Selectin三对相关没有统计学意义,CRP与Fib之间相关仍较密切且具显著性（r=0.421,P＜0.001）,控制变量分组（是否房颤）的作用在各对关系中均有统计学意义。多元回归分析表明,分组（是否房颤）、sex、age和lgCRP的标准回归系数分别为0.302（P=0.004）、-0.396（P＜0.001）、0.221（P=0.032）和0.265（P=0.013）,显示对血浆Fib水平均有显著性影响,房颤、女性、高龄和高炎症水平提示高血Fib水平。多元回归分析表明,分组因素（是否房颤）标准回归系数为0.446（P＜0.001）,显示对血清sP-Selectin水平有显著性影响。亚组分析结果表明,特发性房颤患者和未合并高血压的对照组个体分别为61例和34例（其中房室结折返性阵发性室上速18例,健康个体16例）,两组之间性别比例、年龄、血细胞计数、血生化指标水平、血电解质水平和血小板计数等相比均无显著性差异,两组间血vWF水平无显著性差异,而两组血清sP-selectin水平分别为25.97±6.17 ng/mL和22.08±4.72 ng/mL（P=0.005）,血浆Fib水平分别为3.19±0.97 g/L和2.74±0.57 g/L（P=0.031）。说明与对照相比,特发性房颤患者有明显的高凝状态和血小板激活。特发性房颤中49例阵发性房颤患者血sP-selectin和Fib水平分别为25.50±6.35ng/mL和3.15±0.70g/L,显著高于对照组患者（P值分别为0.018和0.015）。亚组分析结果表明,孤立性房颤患者和＜60岁未合并高血压的对照组个体分别为38例和21例（其中房室结折返性阵发性室上速14例,健康个体7例）,两组之间性别比例、年龄、血细胞计数、血生化指标水平、血电解质水平和血小板计数等相比均无显著性差异,两组间血清vWF和血浆Fib水平无显著性差异,而两组血清sP-selectin水平分别为26.80±6.28ng/mL和22.51±4.97 ng/mL（P=0.009）。提示孤立性房颤患者有显著的血小板激活。方差分析显示,对照组、阵发性房颤和持续性房颤LAd值分别为30.6±3.8,35.1±4.0和43.4±7.1mm,方差分析显示三组间LAd值有显著性差异,两两比较显示阵发性房颤和持续性患者LAd值分别与对照组相比有显著性差异（P值均＜0.001）,持续性房颤与阵发性房颤有显著性差异（P=0.005）。方差分析显示,对照组、阵发性房颤和持续性房颤RAD值分别为42.8±4.8,46.2±6.2和55.8±7.0mm,方差分析显示三组间LAD值有显著性差异,两两比较显示阵发性房颤和持续性患者RAD值分别与对照组相比有显著性差异（P=0.001和P＜0.001）,持续性房颤与阵发性房颤有显著性差异（P=0.001）。提示与窦性心律相比,房颤可出现显著的心房形态学改变,持续性房颤心房形态学改变显著重于阵发性房颤。本研究未发现持续性房颤持续时间对心脏形态学改变有显著影响,考虑可能原因是入选的这两类患者例数太少。单因素分析显示,LAD与RAD显著密切相关（r=0.567,P＜0.001）。提示房颤时左右心房解剖重构是同步的。房颤患者CRP水平与LAd偏相关分析（控制性别、年龄、有无高血压和有/无二尖瓣关闭不全）结果显示,偏相关系数r=0.254,P=0.078;房颤患者CRP水平与RAd偏相关分析（控制性别、年龄、有无高血压和有/无二尖瓣关闭不全）结果显示,偏相关系数r=0.255,P=0.078。提示CRP水平与LAd和RAd的相关性均未达到统计学意义。研究结论1.本研究未发现研究人群在-851、-575、-419和-244四个SNPs位点出现少见碱基。2.首次发现房颤患者在TNF-α启动子区-308位点少见基因型携带率显著增高,并首次证实-308A携带者预示患房颤风险更高,提示TNF-α基因多态性影响房颤的易患性,支持炎症因素作为房颤发病的病因之一。3.发现炎症标记物TNF-α、血小板功能标记物sP-selectin和左房大小标志LAd为房颤发生的独立预测因子,支持炎症因素在房颤的发病过程起作用。4.证实房颤时存在凝血功能激活。血Fib水平的显著影响因素包括房颤发生率、性别、年龄和炎症因素。房颤、女性、高龄和高炎症水平提示高血Fib水平。首次发现炎症因素对房颤患者凝血指标有显著性影响。5.证实房颤时存在血小板功能激活,未发现炎症因素对其有显著性影响。6.证实阵发性房颤发作间歇期有明显的高凝状态和血小板激活。7.特发性房颤患者凝血功能和血小板功能激活,孤立性房颤患者凝血功能无明显激活,但血小板功能显著激活。提示房颤可独立引起凝血和血小板功能激活。8.房颤时左右房直径均显著增大,首次专门对房颤右房的形态学变化做了研究,未发现炎症因素对房颤心房重构过程有显著性影响。9.持续性房颤炎症标记物水平增高和心房直径扩大的程度超过阵发性房颤,提示房颤负荷越大、炎症激活和心房扩大越严重。更多还原

【Abstract】 BackgroudAtrial fibrillation（AF）,the most commonly encountered arrhythmia in clinical practice,is associated with a 2-fold increase in total and cardiovascular mortality,as well as the potential for substantial morbidity,including stroke,congestive heart failure,and cardiomyopathy.Its incidence and prevalence are increasing,and it represents a growing clinical and economic burden.Owing to relative inefficacy and side effects of current pharmacological and Non-pharmacological therapy for AF,it remains a great challenge to improve primary and secondary AF prevention strategies to reduce this potentially enormous health burden.However the pathogeneses,especially risk factors underlying this arrhythmia should be fully understood at first step.Recent findings have demonstrated a mechanistic link between inflammatory processes and the development of AF,a link that may be a target for more effective treatment and prevention strategies.Evidence for an inflammatory contribution to at least some forms of AF was initially suggested by high incidences（25%to 40%） of AF after cardiac surgery.Till now,more and more study have showed that atrial fibrillation associated closely with inflammation,which didn’t depend on the underlying cardiovascular disease.Some of these study found that significantly more inflammatory cells infiltrated or more inflammatory factors expressed within the atrial tissue of atrial fibrillation individuals, some others discovered the levels of circulatory inflammatory markers elevated, when compared to the respective indexes of non-af individuals.Moreover,some pleiotropic agents such as ACEIs,ARB,fish oil,etc.,and corticosteroids had been found to reduce the morbidity of AF in many clinical trials.However,the studies that showed correlation between AF and fibrillation are mostly cross-sectional,and could only imply the“coexistence”but not causal relation of AF and inflammation.Futhermore,there are still some studies showed different results,which didn’t support the relationship between AF and inflammation at all.As a result,unlike atheroclerosis,AF was not accepted as a kind of inflammatory disease by the majority.Tumor necrosis factor（TNF）-αis essentially produced by monocytes and macrophages,and,in turn,it is the strongest known paracrine activator of monocytes and macrophages.Upon stimulation,these cells secrete a variety of products including interleukin（IL）-6,stimulating the liver to produce the acute phase reactant C-reactive protein（CRP）.CRP is an acknowledged indicator of increased systemic inflammation across a wide range of diseases.TNF-αand CRP both are found in considerable quantities in atherosclerotic lesions,and they have also been associated with increased cardiovascular risk in numerous large population-based studies. Hitherto,a lot of studies has investigated the role of GRP in AF,but seldom ones have concentrated on the role of TNF-αin either the pathogenesis or prothrombotic state in AF.Some SNPs in TNF-ahave been found to associated with the level of circulatory TNF-αand some kinds of inflammatory disease,such as COPD and CHD,but to our knowledge,no one has investigate the relationship between SNPs in the TNF-αpromoter and AF.Both thrombophlebitis and pyemia are typical examples of hypercoagulatory states caused by inflammation.In fact,inflammation had been showed as an important reason that lead to thrombotic events in a lot of studies of atherosclerosis and heart failure.Further investigations have showed a prothrombotic state with blood hypercoagulation,platelet activation,damaged atrial endothelium and atrial blood stasis in AF,while these changes are due to the AF itself,or coexistent cardiovascular risk factors remains a matter of debate.Moreover,the relationship between inflammation and this prothrombotic state,especially blood hypercoagulation and platelet activation,has been seldom investigated.Nevertheless,this could be very important for our decision to chose an anti-thrombotic therapy,when the association of AF with inflammation get much clearer.Untill now,no study has specially focused on the changes of coagulation,platelet and endothelial fuction in idiopathic AF and lone AF,which should help to answer whether AF per se or the possibly coexisting cardiovascular disease cause the prothrombotic state in AF patient.In addition,it is somewhat vague the recommendation of antithrombotic therapies for these two specified AF populations in 2006 ACC/AHA/ESC AF guideline.The objectives of this study include:To investigate whether the levels of several inflammatory markers elevate in AF.To examine the proportion of genotype at the SNPs sites of TNF-αpromoter in AF and sex,age and risk factors matched controls, and make clear whether some of them imply great risk to the pathogeneses of AF.To explore the independent risk factors for AF.To investigate whether AF patients have hypercoagulatory state and platelet activation and if so,the influencing factors promoting prothrombotic statte and the role of inflammation in prothrombotic state. To study the association of AF with the incidence of bradicardia and the changes of cardiac morphology.To evaluate the differences of inflammatory markers and ultorsonic indexes of atrium between different types of AF.To observe the effect of duration of AF history on inflammation markers and ultrasonic indexes of atrium.Study population and methods102 AF patents were recruited into the study group,117 sex and age matched individual were recruited into the control group.Cases of AF were identified by ECG or charge diagnosis.Exclusion criteria were as follows:＞75 years old,overt atherosclerosis,congenital heart disease,rheumatic heart disease,myocardiopathy, LV thickening,heart dysfunction,diabete,chronic inflammatory disease,surgery and severe disease,tumor and function failure of important organ.Detailed medical history,physical examination,routine biochemical testing （including Fib and CRP testing）were performed in addition to a 12-lead electrocardiograms（ECG）.By transthorasic echocardiography,valvular functions, left ventricular functions（ejection fraction）,interventricular septum（IVSd）,left atrial diameter（LAd）and right atrial diameter（RAd）were evaluated.LAd was measured in parasternal long axis view.Patients with AF who would be submitted to radiofrequecy ablation still underwent tranesophageal echocardiography in 72 hours before procedure.Each blood sample was obtained from a peripheral vein in drying tube and potassium citrate containing tube in the morning situation of fasting.Serum was immediately extracted,aliquoted,and stored at-80℃until analysis by a batch.Whole blood in the former used to extract genomic DNA with QIAamp^?DNA Blood Mini Kits within 3 days.The procedure included mainly cells lysis,column adsorption, washing and elution.DNA samples were then comfirmed by electrophoresis in agarose gel and their concentration were determined by the method of optical density before being stored at-20℃.The primers for PCR amplification of target fragment of TNF-αpromoter were designed using software Prime Premier 5 and synthesized by Shanghai Shenggong Co.Ltd.PCR amplification for each DNA sample was carried out using Promega PCR kit(GoTaq^?PCR Core SystemsⅠ).The main procedure include denaturation,annealing and extension,for about 30-35 circles.After comfirmed by electrophoresis in agarose gel,PCR products were send to Guangzhou Jingrui company for purifying and sequenceing.The sequencing results were read using Chroma software and blasted at Pubmed to determine genotype at the SNPs sites for every subjectSerum TNF-α,sP-Selectin and vWF were determined using commercially available enzyme-linked immunosorbentassay（ELISA）kits according to manufacturer’s instructions.The interassay coefficient of variation is within 5%and intraassay coefficient of variation is within 10%according to manufacturer’s instructions.The Curve Expert 1.3 was used to make standard curve and transfer optical density into concentration for each indices.Continuous data are expressed as mean±SD if normally distributed and as median （interquartile range）if nonparametrically distributed.Differences in clinical features, blood indices between two groups were evaluated with 2-sample t tests for normally distributed continuous variables,Mann-Whitney U tests for nonparametrically distributed continuous variables,and chi-square tests for categorical variables.Because the distribution of TNF-α,CRP,Fib,vWF,GPT,LAd,RAd,neutrophil and monocyte was skewed,logarithms transformation were applied to remove skewness of these variables before they were included into multivariate linear regression analysis and analysis of covariance.To account for covariate imbalance,analysis of covariance was introduced with the potential confounding factors as covariate.Correlation was evaluated with the Pearson coefficient when both variables are normally numberic ones,and with the Spearman coefficient when one or above of them aren’t normally numberic variables.Partial correlation was introduced to control for the factors that potentially influence the correlation between two variable.A stepwise multiple logistic regression analysis was performed for identifying independent predictors for the presence of AF with sex,age and those varialbles showing difference with significance less than 0.20 between subjects with AF and those without AF in univariate analysis as explanatory variables.The odd ratio of the presence of-308A to the absence of-308A for AF was estimated by the function of“risk estimate”in the crosstab model of SPSS 13.0.Multiple variates linear regression analyses were performed respectively for identifying factors that determine important indices such as TNF-α,CRP,Fib,and sP-Selectin,including age,sex,grouping（AF or not）,and variables associated with each of these indices with significance of less than 0.20 in univariate analysis.A p value of＜0.05 was accepted as significant in all analyses.All statistical analyses were made by using SPSS 13.0 program（SPSS Inc.;Chicago,Illinois）Results219 individuals were enrolled totally,with 102 AF patients in the AF group and 109 non-AF individuals in the control group.The proportion of male/female are 62/40 in AF group and 48/31 in control group（only 71 individuals has a sex record）, and the proportions showed no significant differentce between two groups（P＞0.05）. The AF group and the control group had a mean age of 56.7±10.9（29.0 to 75.0）years old and 56.7±10.9（38.5 to 75.0）years old respectively,with no significant difference found between two groups.The AF group include a sample of 87 paroxysmal AF,6 persistent AF and 9 long-persistent AF,most of whom were submitted to radiofrequency therapy.the control group include a sample of 30 individuals with AVNRT,49 healthy individuals and 38 DNA specimens of healthy persons.41 patients in the AF group and 45 individuals in the control group complicated with hypertension.Two groups had a similar proportion of hypertension（P＞0.05）.Less frequent alleles weren’t found at the SNPs site of-851,-575,-419 and-244 in TNF-αpromoter.The overall proportion of less frequent genotypes at the SNPs site of-863,-857,-806,-782,-646,-572,-570 and-238 of TNF-αpromoter is 24.8%, 24.3%,7.1%,1.0%,1.9%,2.4%,1.0%and 1.9%respectively.The ratioes of less frequent genotypes to ancestral genotype at these sites between the AF group and the control group all showed no significant difference（P＞0.05）.The proportions of less frequent genotypes-308（GA or AA）and ancestral genotype-308GG were 21%and 79%,respectively,in the AF group,with 1 case being homozygote of-308A,and those in the control group were 10.1%and 89.9% respectively,with no-308AA.The distribution of the-308 genotypes was significantly different between the AF group and the control group（P=0.035）with the former having a higher frequency of less frequent genotypes.The presence of the -308A allele（including both homozygous and heterozygous subjects）was associated with an increased risk for AF（odds ratio 2.368,95%CI:1.078-5.205）,compared with -308G homozygous subjects,which implied a higher incidence of AF in less frequent -308 genotypes subjects.In univariate analysis,when compared with the control group,the AF group has a significantly higher levels of circulatory TNF-α（3.84pg/ml（3.07,4.49）versus 2.85 pg/ml（2.28,3.60）,P＜0.001）,CRP（1.70mg/L（1,3.12）versus l.00mg/L（0.95,2.10）, P=0.015）,Fib（3.03g/L（2.62,3.65）versus 2.58g/L（2.28,3.02）,P=0.002）, sP-Selectin（26.5±5.8 ng/ml versus 21.3±3.8ng/ml,P＜0.001）and GPT（22.50 U/L（17.00,29.50）versus 18.00 U/L（13.50,23.50）,P=0.002）.The level of blood vWF show no significant difference between the two groups.These implied subjiects with AF had more severe inflammation and prothrombotic state than those without AF.No statistically significant differences in indices of blood cells,blood glucose, blood lipid,blood electrolytes,hepatic function,renal function and thyroid function, were found when comparing the AF group and the control group.These implied the above indices didn’t influence the differences of the other indices between the two groups.Both fibrinogen and sP-selectin levels were significantly higher in individual with paroxysmal AF than those in the control group（Fib:3.20±0.84 and 2.77±0.60 g/L, respectively,P=0.005;sP-selectin:26.2±5.9 and 21.5±3.9 ng/mL,P＜0.001）.Both differences were still significant after adjusted for age,sex and risk factors in analysis of covariance.The incidences of bradicardia in the AF group and the control group were 29.7% and 6.8%,respectively,with a significant difference between the two groups（P＜0.001）.The subjects with AF had a higher incidence of mitral regurgitation then those without AF（52.3%and 14.9%,P＜0.001）.In univariate analysis,some of the echocardiographic indices were significantly higher in the patients of the AF group than those in the control group.The indices include LAD（33.0mm（35.0,38.0）and 28.0mm（30.0,33.0）,P＜0.001）,RAD（47.0mm（43.0,51.0）and 41.0mm（37.0,43.0）, P＜0.001）and IVSD（10.6（10.3,10.8）and 9.6（9.2,9.9）,P＜0.001）.These imply subjuects with AF have higher incidence of bradycardia and more severe myocardium remodeling,especially atrium remodelling.No statistically significant difference of TNF-αwere found when comparing the groups defined by the presence and absence of the-308A allele（3.37（2.72,4.44）and 3.36（2.83,4.16）,P=0.600）.In multivariate linear regression model,the presence of AF and monocyte number were showed to be independent factor affecting serum TNF-α,with regression coefficient being 0.331（P=0.002）and 0.199（P=0.046）,respectively.The presence of-308A wasn’t a independent factor influencing TNF-α.A multivariate linear analysis using serum CRP level as dependent variable showed 1gFib,1gvWF,1gNeu and lgIVSD were independent predictors,with regression coefficient being 0.343（P =0.002）,0.288（P=0.005）,0.239（P=0.023）,0.347（P=0.001）respectively.TNF-α,sP-Selectin and LAd were showed to be independent predictors for the presence of AF in a multivariate logistic regression model with OR being（TNF-α: 1.375,95%CI（1.024-1.846）,P=0.035;sP-Selectin:1.248,95%CI（1.121-1.389）;P＜0.001 and LAd:1.321,95%CI（1.125-1.552）,P=0.001）.There were statistically positive correlations between inflammatory indices （TNF-αor CRP）and prothrombotic indices（Fib or sP-Selectin）,among which the efficient for the correlation between CRP and Fib was 0.560（P＜0.001）.After controlling for sex,age and grouping（AF or not）,the correlations between TNF-αand Fib,TNF-αand sP-Selectin,and between CRP and sP-Selectin became statistically insignificant,while the statistical significance for the correlation between CRP and Fib was retained（r=0.421,P＜0.001）.Grouping（that is,AF or not）,sex,age and lgCRP were independent determinors of Fib.Their standard coefficients were 0.302（P=0.004）,-0.396（P＜0.001）,0.221 （P=0.032）and 0.265（P=0.013）,which implied the presence of AF,woman,elder and inflammation all associated with higher level of Fib.The presence of AF was the only independent factor influencing the level of sP-Selectin in a multivariate linear regression.Its standard coefficients was 0.446（P＜0.001）.Subgroup analysis showed raised Fib and sP-selectin levels in the subjects with idiopathic AF,compared to those in the subjects without risk factors,which implied a hypercoagulable state and platelet activation in idiopathic AF.Compared with age- and risk factors-matched control,the subjects with lone AF had a higher levels of sP-selectin but not Fib,which implied lone AF associated with platelet activation rather than hypercoagulation.The association of LAD with RAD was statistically significant in univariate analysis.This should imply the simultaneous morphous changes in left and right atrium.The levels of inflammatory markers and prothrombotic factors were not independent predictors for both atrial remodelling.LAd were 30.6±3.8,35.1±4.0 and 43.4±7.1mm for subjuects without AF,with paroxysmal AF and with persistent AF respectively.Analysis of variance showed significant difference in LAd between these three groups,with persistent AF patients having a higher LAd than paroxysmal AF patients and both groups having a higher LAd than subjects without AF.RAd were 42.8±4.8,46.2±6.2 and 55.8±7.0mm for subjuects without AF,with paroxysmal AF and with persistent AF respectively.Analysis of variance showed significant difference in RAd between these three groups,with persistent AF patients having a higher RAd than paroxysmal AF patients and both groups having a higher RAd than subjects without AF. Conclusion1.We didn’t found less frequent alleles at the site of-851,-575,-419 and-244 in TNF-αpromoter.2.The subjects with AF had a higher rate of less frequent genotypes at-308 than those without AF.The presence of less frequent allele-308 A represent an underlying diathesis predisposing to this common arrhythmia.3.TNF-α,sP-selectin and LAd were independent predictors for AF.This should imply that inflammation plays a role in the pathogenesis of AF.4.The subject with AF had statistically higher level of Fib.Sex,age and inflammation were the independent factors affecting plasma fibrinogen level.Female,old age and high level of CRP implied high level of fibrinogen.5.The level of sP-selectin was higher in the subjects with AF.Inflammation wasn’t a significant influential factor.6.Patients with paroxysmal AF had higher levels of fibrinogen and sP-selectin during sinus rhythm compared to those in control groups.7.The levels of plasma fibrinogen and serum sP-selectin were both higher in the subjects with idiopathic AF got higer,which implied that idiopathic AF associated with hypercoagulatory state and platelet activation.The level of serum sP-selectin but not that of plasma fibrinogen was higher in the subjects with lone AF,which implied lone AF associated with platelet activation.8.Both LAD and RAD were significantly larger in subjects with AF,which didn’t rely on inflammation.9.Subjects with persistent and long-persistent AF had higher level of inflammatory markers and larger atrium than subjects with paroxysmal AF.更多还原