【作者】 尹泽钢；

【导师】 周华东；

【作者基本信息】 第三军医大学 ， 神经病学， 2014， 博士

【摘要】 研究背景：脑白质病变(WML)，又称脑白质疏松，多见于老年人群。WML是脑白质区弥散性、融合性病变，边缘模糊，在头颅磁共振(MRI)的T1加权像呈低信号，在T2加权像和液体衰减反转恢复(FLAIR)序列呈高信号。因为在FLAIR中WML可以很好地区别于腔隙，所以FLAIR序列成为评估WML存在与否和严重程度最常用的影像学手段。WML是一种脑小血管病，存在一系列病理学改变，包括髓鞘脱失、血管壁通透性增高、血脑屏障破坏、脑血管自调节障碍、白质低灌注和胶质增生等。WML可分为脑室旁白质病变(PVWML)和深部白质病变(DWML)，虽然它们常常共存，但是有证据表明这两者在病因和临床意义上不完全相同。人脑中的神经元之间存在多重联系网络，不仅掌控着运动和感觉系统，也与注意力、记忆、语言、视空间能力和情绪等神经行为功能密切相关，脑白质的病变会损害这些联系网络，从而影响到这些神经行为功能，特别对于处在生命末期的老年人群，这些影响尤为明显。经过30多年的研究，越来越多的证据表明WML与卒中、认知功能下降、痴呆、抑郁、残疾和全因死亡率有关，对老年人群的健康构成了巨大的威胁。年龄和高血压是公认的WML的危险因素，而其他危险因素尚未被充分研究。考虑到WML对老年人群身体健康的严重危害及对社会医疗体系造成的沉重负担，研究其他危险因素的作用，尤其是可干预的危险因素，具有重要意义。另外，环境危险因素并不能完全解释WML的发病情况，对双胞胎人群的研究和家系研究已经证实，WML具有很高的遗传倾向。研究WML的遗传学基础，对于WML和相关疾病(如卒中和痴呆)的早期诊断和个体化治疗有重要价值。因此，本研究的第一个目标是在中老年人群中探索代谢综合征和血脂水平与WML的关系，这一问题在既往研究中没有得到充分阐述，且充满争议；另一个目标是探索脂联素基因多态性和WML的关系，这在既往研究中从未涉及。第一部分代谢综合征与脑白质病变关系的研究背景：代谢综合征(MetS)是一系列可干预的血管危险因素的集合，包括腹型肥胖、血脂紊乱、高血压和高血糖等。既往研究报道其与脑静息性腔梗(SLI)有关，后者和WML同属脑小血管病，具有相似的发病机制，提示代谢综合征可能也与WML相关，既往研究对此少有报道。本部分的研究目的是在中老年人群中探索代谢综合征及其组分与WML患病风险间的关系。方法：前瞻性连续收集我科50岁或以上入院病人，用头颅MRI评估WML患病与否和严重程度，以改良的Fazekas四分法分级。以经过修订的美国国家胆固醇教育计划成人治疗专题小组III (NCEP-ATPIII)标准诊断MetS，腹围参考亚洲人群标准。采用二分类logistic回归分析MetS及其组分与WML的关系，并对男女性别进行分层分析，最后采用多分类logistic回归分析MetS与WML各个严重程度的关系。在PVWML和DWML中分别考察。结果：1.从2012年6月至2013年1月，共852名患者纳入本次研究。入组患者平均年龄为66.0±9.4岁，52.0%为女性，其中轻度PVWML有311名(36.5%)，中度有77名(9.0%)，重度有71名(8.3%)；轻度DWML有298名(35.0%)，中度有116名(13.6%)，重度有66名(7.7%)。根据NCEP–ATPIII标准，MetS在总人群中患病率为38.4%，在女性中为43.8%，在男性中为32.5%。WML的严重程度随年龄增加而增加(Ptrend<0.001)，亦随MetS组分的个数增加而增加(Ptrend<0.001)。2.调整了年龄和性别后，logistic回归表明MetS与PVWML和DWML风险升高均有关。进一步调整了吸烟、饮酒、冠心病和卒中史后，这种关系依然具有统计学意义(对PVWML，OR：3.21，95%CI：2.26-4.55；对DWML，OR：2.93，95%CI：2.09-4.09)。对性别进行分层分析，MetS在男性和女性中都是WML的独立危险因素。3. MetS组分个数越多，PVWML和DWML的患病风险也越高(Ptrend<0.001)。较之组分个数为0者，组分个数最多(4或5个)的患者，其患PVWML的风险增加6.5倍，患DWML的风险增加3.7倍。4.将MetS的5个组分同时进入回归方程，并调整年龄、性别、吸烟、饮酒、冠心病和卒中史，PVWML的独立危险因素是血压升高(OR：2.69，95%CI：1.91-3.78)、血糖升高(OR：1.55，95%CI：1.08-2.23)和高密度脂蛋白胆固醇(HDL-C)降低(OR：1.42，95%CI：1.01-2.00)；而DWML的独立危险因素是血压升高(OR：2.59，95%CI：1.86-3.60)和HDL-C降低(OR：1.51，95%CI：1.09-2.09)。进一步对性别分层，血压升高在男性和女性中均为PVWML和DWML的危险因素，而血糖升高在男性中仍是PVWML的危险因素，HDL-C降低在女性中仍是DWML的危险因素。5.多分类logistic回归显示MetS与PVWML和DWML的严重程度有关。调整年龄、性别、吸烟、饮酒、冠心病和卒中史后，对轻、中、重度PVWML，OR值分别为2.97(95%CI：2.07-4.26)、4.14(95%CI：2.35-7.28)、5.44(95%CI：2.86-10.35)；而对轻、中、重度DWML，OR值分别为2.44(95%CI：1.72-3.47)、5.09(95%CI：3.05-8.48)和6.31(95%CI：3.30-12.06)。可见，OR值随着WML严重程度增加而增加。结论：1. MetS是中老年人群罹患PVWML和DWML的独立危险因素，MetS患者患PVWML和DWML的风险是无MetS者的3倍左右，且这种效应在男性和女性中均存在。另外，年龄和既往卒中史也是PVWML和DWML的独立危险因素。2. MetS组分个数越多，PVWML和DWML的患病风险也越高。较之组分个数为0者，组分个数最多(4或5个)的患者，其患PVWML的风险增加6.5倍，患DWML的风险增加3.7倍。3.在MetS的组分中，血压升高、血糖升高和HDL-C降低与PVWML患病率上升有关，而血压升高和HDL-C降低与DWML患病率上升有关。4.多分类logistic回归显示，MetS与各个严重程度的WML均有关，但与重度WML的关系最为密切，MetS患者罹患重度PVWML和DWML的风险较无MetS者分别增加了5.4倍和6.3倍。第二部分血脂水平与脑白质病变关系的研究背景：血脂水平与WML的关系在既往研究中充满争议，且缺乏全面系统的研究。因此本部分的研究目的是在中老年人群中探索血脂水平与WML患病风险间的关系。方法：前瞻性连续收集我科50岁或以上入院病人，用头颅MRI评估WML患病与否和严重程度，以改良的Fazekas四分法分级。最后采用多因素logistic回归分析血浆各血脂指标与PVWML和DWML的关系。各血脂指标包括总胆固醇(TC)、甘油三酯(TG)、高密度脂蛋白胆固醇(HDL-C)、低密度脂蛋白胆固醇(LDL-C)、载脂蛋白A-I(ApoA-I)、载脂蛋白B(ApoB)以及ApoB与ApoA-I的比值(ApoB/ApoA-I)。各血脂指标以四分位数法分为4个浓度等级，以最高浓度等级为参照行回归分析。结果：1.从2012年6月至2013年6月，共1282名患者(606名男性，676名女性，平均年龄65.9±9.4岁)纳入研究。轻度PVWML有469名(36.6%)，中度有110名(8.6%)，重度有99名(7.7%)；轻度DWML有486名(37.9%)，中度有147名(11.5%)，重度有91名(7.1%)。PVWML和DWML的患病率和严重程度随年龄增加而增加(Ptrend<0.001)。2.患PVWML者较未患者年龄更大(70.0±8.9VS61.3±7.5，P <0.001)，男性比例更高(50.9%VS43.2%，P=0.006)，冠心病、卒中史、高血压和糖尿病的患病率更高(P <0.05)，收缩压、舒张压和血浆FBG浓度更高(P <0.05)。两组间的吸烟、饮酒情况和体重指数(BMI)无统计学差异(P>0.05)。DWML组和无DWML组的比较与PVWML类似，不同的是，两组间血浆FBG浓度无差异(P=0.386)。3. TC、HDL-C、LDL-C、ApoA-I和ApoB的血浆浓度在PVWML各级间的分布存在统计学差异(P <0.05)。TC、HDL-C和ApoA-I的血浆浓度在DWML各级间的分布存在统计学差异(P <0.05)。4.当调整了年龄、性别、既往卒中史、冠心病史、高血压和糖尿病后，二分类logistic回归显示，HDL-C(Ptrend=0.044)和ApoA-I(Ptrend=0.019)血浆浓度降低可增加PVWML的患病风险，与最高浓度等级比较，最低浓度等级的患病风险均增加约1.5倍。同样，HDL-C和ApoA-I浓度降低也与DWML有关，最低浓度等级的患病风险分别增加约1.5倍和1.8倍。未发现其他血脂指标与WML有关。5.调整年龄、性别、既往卒中史、冠心病史、高血压和糖尿病后，有序多分类logistic回归显示HDL-C(OR：1.39，95%CI：1.01-1.93)和ApoA-I(OR：1.50，95%CI：1.09-2.07)的最低浓度等级与PVWML的严重程度有关，而ApoA-I(OR：1.49，95%CI：1.09-2.05)的最低浓度等级与DWML的严重程度有关。6.对不同性别进行分层分析，多因素logistic回归显示最低浓度等级HDL-C(Ptrend=0.006)和ApoA-I(Ptrend=0.003)增加PVWML患病风险的效应只在女性中存在。同样，只在女性中，HDL-C(Ptrend<0.001)和ApoA-I(Ptrend<0.001)浓度降低可增加DWML的患病风险。结论：1.调整可能的混杂因素后，HDL-C和ApoA-I血浆浓度降低是PVWML和DWML的独立危险因素，未发现其他血脂指标与WML有关。2.有序多分类logistic回归显示，HDL-C和ApoA-I的浓度与PVWML的严重程度有关，最低浓度等级者罹患更严重PVWML的风险增加约1.4倍和1.5倍；ApoA-I的浓度与DWML的严重程度有关，最低浓度等级者罹患更严重DWML的风险增加约1.5倍。3.对不同性别进行分层分析，发现最低浓度等级HDL-C和ApoA-I增加PVWML和DWML患病风险的效应仅在女性中存在。第三部分脂联素基因多态性与脑白质病变关系的研究背景：脂联素有抗动脉粥样硬化、抗炎性反应等生物学效应，脂联素单核苷酸多态性(SNP)与代谢综合征、高血压、糖尿病、卒中和HDL-C浓度有关，但尚无研究关注其与WML的关系。本研究通过横断面研究的方法在中老年人群中探索脂联素SNP位点与WML的关系。方法：前瞻性连续收集我科50岁或以上同意抽血和MRI检查的入院病人，以头颅MRI评估WML患病与否和严重程度，以改良的Fazekas四分法分级，将PVWML和DWML的分数相加，0-1分为对照组，2-6分为WML病例组。用连接酶链反应进行基因分型。最后采用多因素logistic回归分析脂联素各SNP位点与WML的关系。结果：1.2012年6月至2013年1月，811名连续研究对象入组，平均年龄66.4±9.1岁，47.6%为男性，病例组419名，对照组392名。与对照相比，病例组平均年龄更大(P <0.05)，冠心病、既往卒中、高血压和糖尿病的比例较高(P <0.05)，血浆TC、HDL-C和ApoA-I浓度较低(P <0.05)，收缩压、舒张压和FBG浓度较高(P <0.05)。2.所有SNP位点的基因分型成功率均高于97%。各SNP位点病例组和对照组分别进行HWE检验，所有研究组均符合HWE，具有群体代表性(P>0.05)。3.脂联素基因rs7649121位点的AA、AT、TT基因型频率在病例组分别为50.6%、41.7%，7.7%，在对照组分别为61.5%、35.2%、3.3%，在两组间的分布有统计学差异(χ2=13.407，P=0.001)，等位基因A、T的频率在病例组为71.5%和28.5%，在对照组为79.1%和20.9%，在两组间的分布有统计学差异(χ2=12.562，P <0.001)。rs1063539、rs2241767、rs1501299、rs16861194、rs12495941、rs182052、rs266729的基因型和等位基因频率在病例组和对照组间均无差异(P>0.05)。4.调整了年龄、性别、既往卒中史、冠心病、高血压、糖尿病和HDL-C后，AT基因型(OR：1.59，95%CI：1.01-2.50)和TT基因型(OR：3.44，95%CI：1.18-10.03)仍是WML的独立危险因素，在显性模型中，AT/TT基因型患WML的风险是AA基因型的1.7倍。结论：1.在中国中老年人群中首次发现，脂联素基因的rs7649121位点与WML有关，T等位基因是WML发生的易感基因。2.未发现脂联素基因的rs1063539、rs2241767、rs1501299、rs16861194、rs12495941、rs182052和rs266729位点与WML有关。更多还原

【Abstract】 Background:Cerebral white matter lesions (WMLs), also known as leukoaraiosis, are frequentlyobserved on neuroimaging in the brains of elderly patients. They are defined as diffuse,confluent white matter abnormalities, often with irregular margins. On magnetic resonanceimaging (MRI), they are seen as hypointensities on T1-weighted imaging andhyperintensities on T2-weighted imaging and fluid-attenuated inversion recovery sequences(FLAIR). FLAIR is now the preferred imaging modality to assess the presence and severityof WMLs because of clear distinction between WMLs and lacunes. WMLs aremanifestation of cerebral small vessel disease and reflect multiple pathologic changes,including loss and deformation of myelin sheath, changes in vessel wall permeability,disruption of the blood-brain barrier, hypoperfusion attributable to altered cerebrovascularautoregulation and gliosis. WMLs are often divided into two categories: periventricularWMLs (PVWMLs) and deep WMLs (DWMLs), with a possible dissimilarity in pathogenicmechanisms and clinical relevance. The human brain contains multiple networks of neuronsthat serve not only motor and sensation but also neurobehavioral functions such as attention,memory, language, visuospatial ability, and emotion. Damage to white matter might disrupthigher cortical functions and be related to various neurobehavioral syndromes, especially inelderly people. The clinical and pathologic significance of WMLs has been investigated byseveral studies over the last30years. Although considered as benign changes at first,WMLs have been proven by accumulated evidence to predict an increased risk of stroke,cognitive decline, dementia, depression, disability, and all-cause mortality. Whereasadvanced age and hypertension are the most widely accepted risk factors for WMLs, thecurrent understanding of other risk factors for WMLs remains less clear. Given thedetrimental impact of WMLs on individuals and healthcare systems, knowledge of the risk factors for WMLs, especially modifiable ones, is becoming more important. Furthermore,environmental risk factors account for only a part of the variance in the formation of WMLs,and twins and family studies have shown that genetic factors are contributory. Studying thegenetics of WMLs may be very helpful in early diagnosis and individual treatment ofWMLs and related diseases, such as stroke or dementia.Therefore, the aim of the present study was to investigate the associations of WMLswith metabolic syndrome and serum lipids in middle aged and elderly patients, which havenot fully elucidated in former studies. And the other aim was to explore the relationshipbetween WMLs and single nucleotide polymorphisms (SNPs) of adiponectin gene, whichhas never been investigated before.Section One Assotiation between WMLs and metabolic syndromeBackgroud:Metabolic syndrome (MetS) is a constellation of modifiable vascular risk factorsincluding abdominal obesity, dyslipidemia, hypertension, and hyperglycemia. It has beenrelated to silent lacunar infarction, which is another manifestation of cerebral small vesseldisease and has similar pathogenesis to WMLs. However, little is known about therelationship between MetS and the prevalence of WMLs. The aim of the study was toinvestigate the association between MetS, its components and WMLs in middle-aged andelderly patients.Methods:Consecutive inpatients aged50years and older were prospectively enrolled in thisstudy. All participants underwent MRI scans to assess the presence and severity ofPVWMLs and DWMLs using the4-point modified Fazekas’method. The MetS was definedaccording to the updated National Cholesterol Education Program’s Adult Treatment PanelIII criteria. Multivariate logistic regression analyses were performed to examine therelationship between MetS, its components, and WMLs. Further sex-specific analyses wereconducted. Finally, a multinomial logistic regression model was performed to evaluate theassociation between MetS and the severity of PVWMLs and DWMLs.Results:1. From June2012to January2013, a total of852consecutive patients were enrolledin the study. The mean age was66.0±9.4years;52.0%were female. Mild periventricular WMLs (PVWMLs) was found in311(36.5%), moderate in77(9.0%), and severe in71(8.3%), while mild deep WMLs (DWMLs) was found in298(35.0%), moderate in116(13.6%) and severe in66(7.7%). According to the updated NCEP-ATP III criteria, MetSwas present in38.4%of the total population,43.8%of females and32.5%of males. Theprevalence of PVWMLs and DWMLs increased with increasing age and number of MetScomponents ((Ptrend<0.001).2. In logistic regression model adjusted for age and sex, MetS was associated with anincreased risk of PVWMLs and DWMLs. Further adjustment for current smoking, dailydrinking, coronary heart disease (CHD), and prior stroke attenuated the ORs slightly, butthe association remained significant (OR:3.21,95%CI:2.26-4.55for PVWMLs; OR:2.93,95%CI:2.09-4.09for DWMLs). Moreover, in sex-stratified analyses, the results remainedsignificant in both men and women.3. There was a clear trend towards an increased prevalence of PVWMLs and DWMLswith increasing number of MetS components ((Ptrend<0.001). Compared to patients withoutany MetS components, those having4or5components were almost6.5times more likelyto have PVWMLs and3.7times to have DWMLs.4. When all the five components were entered simultaneously into the logistic model,with multivariable adjustment for age, sex, current smoking, daily drinking, CHD and priorstroke, the independent risk factor of PVWMLs were elevated blood pressure (BP)(OR:2.69,95%CI:1.91-3.78), elevated fasting blood glucose (FBG)(OR:1.55,95%CI:1.08-2.23) and low high-density lipoprotein cholesterol (HDL-C)(OR:1.42,95%CI:1.01-2.00), while elevated BP (OR:2.59,95%CI:1.86-3.60) and low HDL-C (OR:1.51,95%CI:1.09-2.09) showed an increased risk of DWMLs. In the sex-stratified analyses,elevated BP was an independent risk factor of PVWMLs and DWMLs in both sexes.However, the association between elevated FBG and PVWMLs was observed only in menand the impact of low HDL-C on DWMLs was observed only in women.5. A multinomial logistic regression was performed with adjustment for age, sex,current smoking, daily drinking, CHD and prior stroke to evaluate the association of MetSwith the severity of PVWMLs and DWMLs. The adjusted ORs were2.97(95%CI:2.07-4.26) for mild,4.14(95%CI:2.35-7.28) for moderate and5.44(95%CI:2.86-10.35)for severe PVWMLs, while the adjusted ORs for increasing grades of DWMLs were2.44 (95%CI:1.72-3.47) for mild,5.09(95%CI:3.05-8.48) for moderate, and6.31(95%CI:3.30-12.06) for severe, respectively. The ORs increased with the advancing grades ofPVWMLs and DWMLs, showing a dose effect.Conclusions:1. The presence of MetS was associated with an approximate3times increased risk ofPVWMLs and DWMLs in middle-aged and elderly patients, after adjustment forconfounding factors. This effect was observed in both sexes. Apart from MetS, age andprior stroke were independent risk factors of PVWMLs and DWMLs as well.2. The number of MetS components was associated with the presence of PVWMLsand DWMLs. Those having4or5components were almost6.5times more likely to havePVWMLs and3.7times to have DWMLs.3. the independent risk factor of PVWMLs were elevated BP, elevated FBG and lowHDL-C, while elevated BP and low HDL-C showed an increased risk of DWMLs.4. The presence of MetS was associated with the severity of PVWMLs and DWMLs inmultinomial logistic regression models. Patients with MetS were almost5.4times morelikely to have severe PVWMLs and6.3times more likely to have severe DWMLs,compared with those without MetS..Section Two Assotiation between WMLs and levels of serum lipidsBackground:The relationship between WMLs and levels of serum lipids is controversial in formerstudies. Thus, the aim of this study was to investigate the association of serum lipids withthe presence of PVWMLs and DWMLs in middle-aged and elderly subjects.Methods:Consecutive inpatients aged50years and older of our department were prospectivelyenrolled in this study. All participants underwent MRI scans to assess the presence andseverity of PVWMLs and DWMLs using the4-point modified Fazekas’ method.Multivariate logistic regression analyses were performed to examine the association ofWMLs with serum lipids including total cholesterol (TC), triglyceride (TG), HDL-C,low-density lipoprotein cholesterol (LDL-C), apolipoprotein A-I (ApoA-I), apolipoproteinB (ApoB) and ApoB/ApoA-I. The levels of the lipid profile were divided into four quartilesand the highest quartile was used as the reference. Results:1. From June2012to June2013, a total of1282consecutive patients (606men and676women,65.9±9.4years) were enrolled in the study. Mild PVWMLs was found in469(36.6%), moderate in110(8.6%), and severe in99(7.7%), while mild DWMLs was foundin486(37.9%), moderate in147(11.5%) and severe in91(7.1%). The prevalence ofPVWMLs and DWMLs increased with advancing age ((Ptrend<0.001).2. Patients with PVWMLs were older (70.0±8.9VS61.3±7.5，P <0.001) andshowed significantly higher proportion of male sex, hypertension, diabetes mellitus, priorstroke, and CHD, and higher levels of systolic/diastolic BP and FBG (P <0.05). Thesmoking and drinking status and body mass index (BMI) were not different between groups(P>0.05). The differences between patients with DWMLs and without DWMLs weresimilar to those with and without PVWMLs, except for levels of FBG (P=0.386).3. The distribution of serum TC, HDL-C, LDL-C, ApoA-I and ApoB levels weresignificantly different between four grades of PVWMLs (P <0.05). The distribution ofserum TC, HDL-C and ApoA-I levels were significantly different between four grades ofDWMLs (P <0.05)4. After adjustment for age, sex, prior stroke, CHD, hypertension, and diabetes,patients with the lowest HDL-C (Ptrend=0.044) or apoA-I (Ptrend=0.019) quartile wereapproximately1.5times more likely to have PVWMLs, compared with those with thehighest quartile. Likewise, patients with the lowest HDL-C (Ptrend=0.013) or apoA-I (Ptrend=0.002) quartile were approximately1.5or1.8times more likely to have DWMLs. Noneof other lipids were related to WMLs.5. After adjustment for age, sex, prior stroke, CHD, hypertension, and diabetes, thelowest HDL-C (OR:1.39,95%CI:1.01-1.93) or apoA-I (OR:1.50,95%CI:1.09-2.07)quartile were related to the severity of PVWMLs in ordinal regression models, while thelowest apoA-I (OR:1.49,95%CI:1.09-2.05) quartile was related to the severity ofDWMLs.6. In sex-specific analyses, the lower levels of HDL-C and ApoA-I predicted higherrisk of PVWMLs and DWMLs only in women.Conclusions:1. Decreased serum levels of HDL-C and ApoA-I were associated with an increased risk of PVWMLs and DWMLs independent of other risk factors.2. In odinal logistic regression models, serum levels of HDL-C and ApoA-I wererelated to the severity of PVWML, while levels of ApoA-I were related to the severity ofDWML.3. In sex-specific analyses, the lower levels of HDL-C and ApoA-I were associatedwith higher risk of PVWMLs and DWMLs only in women.Section Three Assotiation between WMLs and adiponectin gene polymorphismsBackground:Adiponectin manifests anti-atherosclerosis and anti-inflammatory effects. Singlenucleotide polymorphisms (SNPs) in adiponectin gene have been related to MetS,hypertension, diabetes, stroke and levels of HDL-C. However, the association of SNPs inadiponectin gene with WMLs has never been studied before. Therefore, the aim of thisstudy was to investigate the association between adiponectin gene SNPs and WML inmiddle aged and elderly subjects, using the cross-sectional design method.Methods:Consecutive inpatients aged50years and older of our department were prospectivelyenrolled in this study. All participants underwent MRI scans to assess the presence andseverity of WMLs using the4-point modified Fazekas’ method. For statistical analyses,WMLs severity, defined as the sum of the scores of PVWMLs and DWMLs, wasdichotomized into scores0-1(controls) versus2-6(cases). Multivariate logistic regressionanalyses were performed to examine the association of WMLs with SNPs.Results:1. From June2012to January2013, a total of811consecutive subjects were enrolledin the study. The mean age was66.4±9.1,47.6%were males. Among the participants,419were cases and392were controls. The cases were older (P <0.05), and showedsignificantly higher proportion of hypertension, diabetes mellitus, prior stroke, and CHD (P<0.05), and higher levels of systolic/diastolic BP and FBG (P <0.05) whereas lower levelsof TC, HDL-C and ApoA-I (P <0.05).2. More than97%samples were genotyped successfully for each SNP. Every groupconformed to the Hardy-Weinberg equilibrium (P>0.05).3. As for rs7649121, the AA, AT and TT genotype frequency were50.6%,41.7%and 7.7%in the case group respectively, whereas61.5%,35.2%and3.3%in the control group.The distribution of genotypes was different between the two groups significantly (χ2=13.407，P=0.001). The A and T allele frequency were71.5%and28.5%in cases whereas79.1%and20.9%in controls, showing significant difference (χ2=12.562，P <0.001). Nodifferences in genotype or allele frequency were observed in rs1063539, rs2241767,rs1501299, rs16861194, rs12495941, rs182052and rs266729between cases and controls.4. After adjustment for age, sex, prior stroke, CHD, hypertension, diabetes and HDL-Clevels, AT (OR:1.59,95%CI:1.01-2.50) and TT (OR:3.44,95%CI:1.18-10.03) wereassociated with higher risk of WMLs independently. In dominant model, AT/TT genotypewas1.7times more likely to have WMLs than AA genotype.Conclusions:1. Among the middle aged and elderly subjects, the present study firstly found thatrs7649121in adiponectin gene might be associated with WMLs, and T allele was probablya susceptible allele of WMLs.2. We failed to found any association between WMLs and other SNPs in adiponectingene including rs1063539, rs2241767, rs1501299, rs16861194, rs12495941, rs182052andrs266729.更多还原