质子磁共振波谱定量检测肝脏脂肪含量方法的建立及肝脏脂肪含量与胰岛素抵抗的关系

【作者】 刘蒙；

【导师】 高鑫； 曾蒙苏；

【作者基本信息】 复旦大学 ， 内科学， 2008， 博士

【摘要】 目的:在大鼠脂肪肝动物模型上应用质子磁共振波谱(~1H MRS)定量检测肝脏脂肪含量,评价其稳定性与精确性,初步建立~1H MRS定量测定肝脏脂肪含量方法,并在人体水平进一步验证该方法的稳定性与实用性。方法:3周龄近交系Wistar大鼠40只,随机分为两组,建立大鼠脂肪肝动物模型。实验组大鼠给予高脂饮食(脂肪供能占52%),对照组大鼠给予标准饮食(脂肪供能占12%),共喂养16周。第4、8、12、16周各从两组随机抽取5只大鼠,应用~1H MRS测定肝脏脂肪含量,以肝脏内水峰作为内参,测定并计算肝脏脂肪峰与水峰的峰下面积比值。所有大鼠处死后,肝组织行冰冻切片苏丹Ⅲ脂肪染色,应用Motic专业医学图像软件分析病理切片图像,计算肝脏脂肪含量,结果以肝组织内脂肪变性的面积占分析区域总面积的百分比来表示肝脏脂肪含量百分比。将~1H MRS测定的脂肪峰面积/水峰面积的比值与病理学测定的肝脏脂肪含量百分比进行比较,分析二者之间的相关性,初步建立~1H MRS测定肝脏脂肪含量的方法。然后在临床选择14例行肝穿刺活检的脂肪肝患者应用~1H MRS测定肝脏脂肪含量,评价其临床应用的可操作性和稳定性。所有患者均行经皮细针肝穿刺活检,肝组织行常规HE染色,应用Motic医学图像软件测定肝脏脂肪含量百分比。分析~1H MR测定结果与病理学测定的肝脏脂肪含量百分比之间的相关性。结果:随着饲养时间延长,高脂饮食组大鼠肝脏脂肪含量逐渐增加,应用~1H MRS在所有高脂饮食组大鼠肝脏内检测到了稳定的脂肪峰和水峰,重复性佳。当肝脏脂肪含量＜1%时,~1H MRS即可检测到波谱脂肪峰的明显改变。脂肪峰及脂肪峰与水峰的峰下面积比值随着肝脏脂肪含量的增加而升高,与病理学测定的脂肪含量变化一致。脂肪峰面积/水峰面积的比值与病理学测定肝脏脂肪含量百分比之间高度正相关(r=0.90,P＜0.001,斜率为0.145,Y轴截距为-0.005),初步建立了敏感、稳定的~1H MRS定量检测肝脏脂肪含量的方法。选择14例脂肪肝患者应用~1H MRS定量检测肝脏脂肪含量,验证其稳定性与实用性。在所有患者肝脏内检测到稳定的水峰与脂肪峰,~1H MRS谱线较动物水平检测到的谱线更加稳定。随着肝脏脂肪含量的增加,~1H MRS测定的脂肪峰及脂肪峰与水峰的峰下面积比值亦逐渐升高,当肝脏脂肪含量超过75%时,脂肪峰相对抑制了水峰。分析~1H MRS测定的脂肪峰面积/水峰面积的比值与肝穿刺病理学测定的肝脏脂肪含量百分比之间的相关性,发现二者之间亦呈高度正相关(r=0.91,P＜0.001,斜率为0.487,Y轴截距为0.065)。将动物水平的数据与人体水平的数据整合后分析~1H MRS检测的脂肪峰面积/水峰面积的比值与病理学测定的肝脏脂肪含量百分比之间的相关性,二者仍然呈高度正相关,并建立了稳定的线性回归方程(r=0.91,P＜0.001,斜率为0.565,Y轴截距为-0.037)。结论:在国内首次在动物水平建立了应用~1H MRS无创性相对精确测定肝脏脂肪含量的方法,并且人体水平进一步验证。该方法稳定、敏感、精确、受外因素干扰少、重复性佳、具备临床可操作性,与病理学测定结果相关性好。为NAFLD的临床与基础研究提供了定量测定脂肪含量的新技术。目的:在高脂饮食诱导的非酒精性脂肪肝病大鼠动物模型中观察肝脏脂肪含量与胰岛素抵抗的关系;同时分析既往无糖尿病史的非酒精性脂肪肝病(NAFLD)患者中,应用不同方法测定的肝脏脂肪含量与胰岛素抵抗之间的关系。方法:动物水平,与第一部分实验应用的是同一批大鼠模型,每周随访大鼠体重、空腹及随机血糖和胰岛素;每四周从两组各随机抽取5只大鼠行腹腔糖耐量试验(IPGTT),计算相关参数评价胰岛素抵抗及胰岛B细胞分泌功能。处死动物,肝组织病理切片行苏丹Ⅲ脂肪染色测定肝脏脂肪含量。临床收集2005年11月～2007年7月门诊既往无糖尿病史的NAFLD患者374例,其中133例行校正CT管肝脏脂肪半定量检查,15例行经皮肝穿刺活检病理学检查。除外病毒性肝炎、自身免疫性肝病、药物性肝炎、其他疾病导致的慢性肝损伤,大量饮酒史(男性≥40g/天,女性≥20g/天)。询问现病史及既往史,测量身高、体重、腰臀围、血压。实验室检查包括:肝功能,血脂谱,HBsAg,HCVAb。所有研究对象均行口服75g葡萄糖耐量试验(OGTT),测定空腹、口服75g葡萄糖粉后2小时血糖及胰岛素。15例行肝活检的患者同时测定OGTT试验5点血糖和胰岛素。以ATPⅢ定义为标准诊断MS。应用SPSS for windows13.0统计分析,所有结果以P＜0.05为差异具有统计学意义。结果:在动物水平,高脂饮食组大鼠的空腹及随机血糖均高于对照组;IPGTT试验中,第12周高脂饮食组大鼠的葡萄糖曲线下面积(AUCg)及胰岛素曲线下面积(AUCi)均高于对照组,HOMA-IR也较对照组升高,并且高于第4、8周水平,提示随着肝脏脂肪含量的增加,胰岛素抵抗逐渐加重;而AUCi/AUCg的比值却低于对照组,提示可能存在胰岛β细胞分泌功能障碍。应用~1H MRS检查和肝穿刺活检相对精确测定肝脏脂肪含量,在15例NAFLD患者,根据肝脏脂肪含量中位数进行二分位分割,肝脏脂肪含量≥27%组患者的BMI、腰围、收缩压、肝酶、TG、血糖、HOMA-IR均高于肝脏脂肪含量＜27%的组;OGTT试验中,肝脏脂肪含量≥27%组各点血糖均较肝脏脂肪含量＜27%组升高,且葡萄糖曲线下面积(AUCg)亦升高,而胰岛素曲线下面积(AUCi)及AUCi/AUCg的比值则低于肝脏脂肪含量＜27%组,但是因为样本例数过少,差异没有统计学意义。应用CT扫描半定量测定肝脏脂肪含量的方法,在133例NAFLD患者中,根据肝脏脂肪含量四分位区间,将患者按照肝脏脂肪含量由低到高分为4组。随着肝脏脂肪含量的增加,BMI、腰围、WHR、舒张压、肝酶水平逐渐升高,而HDL-c水平逐渐降低。校正年龄、性别后,肝脏脂肪含量与AST(r=0.534)、ALT(r=0.525)、GGT(r=0.348)、腰围(r=0.304)、BMI(r=0.232)、舒张压(r=0.292)、收缩压(r=0.219)、OGTT后2h血糖(r=0.228)之间呈正相关(P值均＜0.05),与HDL-c(r=-0.212)之间呈负相关(P＜0.05)。随着肝脏脂肪的增加,MS的伴有率亦逐渐增加(P=0.001)。校正年龄性别后,肝脏脂肪含量取代腰围成为MS独立的危险因素(OR值1.659,95%可信区间1.132-2.430)。同样,在代表肥胖的形体学参数中,只有肝脏脂肪含量为HOMA-IR独立的危险因素(P＜0.05)。将肝酶作为代表肝脏脂肪沉积的简易临床指标,在374例NAFLD患者中,肝酶异常者214人(57.2%)。肝酶异常组的体重、BMI、腰围、血压、TC、TG、空腹胰岛素、HOMA-IR均高于肝酶正常组(P＜0.05),而HDL-c低于肝酶正常组。校正年龄、性别后,腰围、TG为肝酶异常独立的危险因素[OR值分别为1.11(95%CI,1.05-1.17),1.36(95%CI,1.09-1.70)]。根据ATPⅢ定义诊断MS,374例NAFLD患者中有147例为MS(39.3%),肝酶异常组MS患病百分比为明显高于肝酶正常组(47.7%vs 28.1%,P＜0.001)。校正年龄和性别后,肝酶降低了舒张压、TG、HDL-c、空腹血糖对MS患病风险的贡献,GGT为MS独立的危险因素(OR值为1.52,95%CI为1.063-2.184)。结论:动物水平研究发现肝脏脂肪含量增加与胰岛素抵抗有关,并可能影响胰岛β细胞的分泌功能。在NAFLD患者中,发现肝脏脂肪含量与胰岛素抵抗及代谢综合征的主要组分密切相关,肝脏脂肪含量比BMI、腰围、WHR等传统代表肥胖与脂肪分布的形体学参数对胰岛素抵抗的影响更敏感,因为肝脏脂肪含量更确切地反映了内脏脂肪沉积;临床常用的肝酶指标与MS密切相关。更多还原

【Abstract】 PartⅠEstablishment of the Method of Quantitively Measuring hepatic fat Content by ~1H magnetic Resonance Spectroscopy In VivoObjectiveTo establish the quantitive method of measuring liver fat content by ~1H magnetic resonance spectroscopy in high-fat diet induced fatty liver model of Wistar rats,and to evaluate the stability and accuracy in patients with NAFLD.Methods40 inbred line male Wistar rats were randomly divided into high-fat diet group and control group.The model group was fed high-fat diet(52%of energy from fat) for 16 weeks,and control group standard diet(12%of energy from fat).Five rats randomly chosen from control group and model group respectively at the fourth, eighth,twelfth,sixteenth weekend,were measured intrahepatic lipid peak and water peak by ~1H MRS in vivo.Peak areas for all resonances were obtained and lipid resonances were quantified with reference to water resonance.The ratio of the lipid peak area relative to water peak area was computed.Then all rats were sacrificed. Liver tissues were stained with SudanⅢfor neutral fat,and the percentage of liver fat content was measured by the software for analysis of medical images.14 patients with fatty liver disease were collected for this study.All the patients measured intrahepatic lipid peak and water peak by ~1H MRS(as above) in vivo and performed liver biopsy. The liver tissues were stained by hematoxylin eosin,and the percentage of liver fat content was measured by the same method as that in rats.ResultThe liver fat content of rats in high-fat diet group gradually increased during the follow-up period.Intrahepatic lipid peak and water peak were detected in all the rats of model group.The spectral line by ~1H MRS was stable and sensitive.Even the percentage of hepatic fat content was less than 1%,the change of lipid peak was observed.The lipid peak area and the ratio of the lipid peak area relative to water peak area in liver rised with increased liver fat content.There was a close positive correlation between the ratio of the lipid peak area relative to water peak area by ~1H MRS in vivo and liver fat content.by histologic measurement.[r=0.90;P＜0001;slope 0.145;y-intercept -0.005]in the rats.The method of quantitive measurement of hepatic fat content by ~1H MRS was established at the animal level.Fourteen patients with NAFLD were selected to measure liver fat content by ~1H MRS and performed liver biopsy.Intrahepatic lipid peak and water peak were detected in all the patients.The spectral line by ~1H MRS was more stable than that in the rats.The lipid peak area and the ratio of lipid peak area relative to water peak area increased with the increment of liver fat content.When the liver fat content was more than 75%,the water peak was inhibited by the lipid peak.There was a close positive correlation between the ratio of the lipid peak area relative to water peak area by ~1H MRS in vivo and liver fat content.by histologic measurement.[r=0.91;P＜0001;slope 0.487;y-intercept 0.065].The data of rats and patients was analyzed together,there was a close positive correlation[r=0.91;P＜0001;slope 0.565;y-intercept -0.037]between the ratio of the lipid peak area relative to water peak area and liver fat content.by histologic measurement.ConclusionIt was the first study that established successfully the method of measurement of liver fat content by ~1H MRS in China.It was a safe,stable,reproducible,sensitive and non-invasive method for relatively accurate quantification of intrahepatic lipid content, and had a good correlation with histologic results,which provided a new method for clinical and basic study of NAFLD. PartⅡThe Relationship of Liver Fat Accumulation with Insulin Resistance by different methods of evaluating liver fat contentObjectTo observe the association of liver fat content and insulin resistance in high-fat diet induced fatty liver model of rats and the patients with nonalcoholic fatty liver disease(NAFLD) by different methods of evaluating liver fat content.Materials and MethodsThe fatty liver model of rats was the same as that in the firs part of the study. Weight,fasting and random blood glucose and insulin were determined respectively every week.Five rats randomly chosen from control group and model group respectively at the fourth,eighth,twelfth,sixteenth weekend,was performed IPGTT. Then rats were sacrificed.Liver tissues were stained with SudanⅢfor neutral fat, and the percentage of liver fat content was measured as described in the first part.Clinical data of 374 outpatients with NAFLD and without known diabetes were collected in from Nov.2005 to Jul.2007,in which 133 subjects were performed calibrated computerized tomography(CT) scan and 15 subjects were measured liver fat content by percutaneous liver biopsy.The subjects were excluded in those with the history of chronic virus hepatitis,autoimmune hepatitis,hepatitis induced by drugs and genetics,male subjects with an alcohol consumption≥40g/d and female≥20g/d. Each subject received the standard interview.Anthropometrics(height,weight,waist circumference,hip circumference and blood pressure ) were measured.Laboratory tests included liver enzymes,lipid profile,HBsAg and HCVAb,Fasting and 2 hour plasma glucose and insulin level were tested in OGTT test in all subjects.Fasting, 30min,60min,120min,180min glucose and insulin in OGTT test in the 15 patients who performed liver biopsy were tested.MS was defined according to ATPⅢ. Differences were considered statistically significant if P＜0.05.ResultsFasting and random glucose of rats in the high fat diet group were significantly higher than those in the control group during the follow-up period.The areas under the curves of serum blood glucose levels(AUCg) and insulin levels(AUCi) and HOMA-IR were higher in the rats of high fat diet group which fed for 12 weeks than those of control group.The degree of insulin resistance worsen with the increment of liver fat content.The ratio of AUCi/AUCg was lower in the rats of high fat diet, which suggested there was dysfunction of insulin secretion in the beta cell.According to median of liver fat content(27%) determined by liver biopsy,15 subjects who measured liver fat content by liver biopsy and ~1H MRS were divided into two groups.Levels of BMI,waist circumference,WHR,blood pressure,liver enzymes,TG,serum glucose,insulin,HOMA-IR and AUCg in the group whose liver fat content≥27%were higher than that in the group whose liver fat content＜27%, but HDL-c,AUCi and the ratio of AUCi/AUCg were lower,possibly because of small case size,the results were no statistically significant differences.Liver fat content in all subjects tested by calibrated CT scan were divided into 4 groups according the quartile of liver fat content.BMI,waist circumference,diastolic blood pressure,serum ALT,AST,GGT level elevated respectively with increaseing of liver fat content,while serum HDL-c level was decreased.After adjusted age and sex, liver fat content showed statistically significant positive correlations with AST(r=0.534),ALT(r=0.525),GGT(r=0.348),waist circumference(r=0.304), BMI(r=0.232),diastolic blood pressure(r=0.292),systolic blood pressure(r=0.219), 2h glucose after OGTT(r=0.228)(all P＜0.05),negative correlation with HDL-c (r=-0.212)(P＜0.05).When liver fat content increased by quartile order,the percentage of MS was increased accordingly(P＜0.05).After adjusted for age and sex, Logistic multivariable stepwise regression analysis demonstrated that liver fat contend instead of waist circumference was the only significant risk factor of MS among parameters representing fat distribution(include BMI waist circumference and WHR) (OR=1.659,95%CI 1.132-2.430).Stepwise multiple linear regression analysis demonstrated that only liver fat content was the significant risk factor of HOMA-IR, while BMI and waist circumference were not the risk factor of it.Taking liver enzymes as the markers of liver fat accumulation,374 subjects with NAFLD without known diabetes were analyzed in this study.Liver enzymes’ levels were found abnormal in 214 subjects(57.2%).Weight,BMI,blood pressure,TC,TG, fasting insulin,HOMA-IR of with abnormal liver enzymes were higher than that of subjects with normal liver enzymes,while level of HDL-c was lower(P＜0.05).Waist circumference and TG were the significant risk factors of abnormal liver enzymes after adjusted for age and sex[OR(95%CI):waist circumference 1.11(1.05-1.17), TG 1.36(1.09-1.70)].147 subjects(39.3%) were diagnosed MS using the ATPⅢcriteria.More subjects were diagnosed MS in the group with abnormal liver enzymes compared with those with normal liver enzymes(47.7%vs 28.1%,P＜0.001).After adjusted for age and sex,Logistic multivariable stepwise regression analysis demonstrated that level of liver enzymes decreased the contribution to MS of diastolic blood pressure,TG,HDL-c and fasting glucose,GGT was independently associated with the presence of MS[OR 1.52,95%CI(1.063-2.18)].ConclusionsWe found that liver fat content was associated with insulin resistance,and might affect the function of beta cell secretion in the fatty liver model of Wistar rats. Liver fat content was associated with most components of MS.Compared with BMI and waist circumference,liver fat content was a more sensitive affective factor of MS and insulin resistance,which suggested that visceral fat accumulation,particularly liver fat accumulation,was a more significant risk factor of MS.Liver enzymes as clinical markers of reflecting liver fat accumulation was associated closely to MS.更多还原