非酒精性脂肪性肝病无创性诊断方程的构建

Establishment of a noninvasive diagnosis equation for nonalcoholic fatty liver disease

目的利用与非酒精性脂肪性肝病(NAFLD)相关的常用临床及实验室指标,构建无创性诊断方程,并评估该方程对NAFLD的预测及诊断价值。方法选取2016年11月-2017年11月西安医学院第一附属医院诊断为NAFLD患者127例和健康对照者30例,记录研究对象的性别、年龄、BMI、病史、ALT、AST、GGT、尿素氮(BUN)、尿酸(UA)、血肌酐(Cr)、TC、TG、HDL、LDL、糖化血红蛋白(HBA1c)、游离脂肪酸(FFA)、空腹血糖(FPG)、空腹胰岛素(FINS)、PLT、超声检查结果和Fibro Scan的检查结果等数据。计量资料组间比较采用t检验,相关性使用Pearson相关性分析,回归方程的构建使用多元线性回归方程模型,构建受试者工作特征曲线(ROC曲线),计算回归方程的敏感度和特异度。结果与脂肪肝相关的指标包括BMI(r=0.308,P=0.005)、ALT(r=0.379,P<0.001)、AST(r=0.318,P=0.004)、GGT(r=0.293,P=0.009)、UA(r=0.244,P=0.033)、FFA(r=0.249,P=0.030);对Fibroscan的受控衰减参数(CAP)相关指标进行多重回归分析,CAP的回归模型具有统计学意义(F=11.113,P<0.001),其校正决定系数R2=0.274,说明回归引起的变异在总变异中所占的比率为27.4%,对CAP影响最大的指标是ALT(β=0.358,P=0.001),其次是BMI(β=0.258,P=0.012<0.05),构建的回归方程为CAP=113.163+0.252×ALT+6.316×BMI,诊断方程的ROC曲线下面积为0.927,灵敏度为87.68%,特异度为90.00%,cut-off值为277.67,P<0.001,具有较高的诊断效能。结论与目前已有的诊断方程相比,此方程的ROC曲线下面积、特异度、灵敏度均较佳,计算方法最为简单,实用性及可操作性强,有利于便捷筛查出早期NAFLD,提高自我干预意识,进一步降低NAFLD的全球危害和疾病进程。可以作为一个诊断NAFLD的方程,供临床参考及推广。

Objective To establish a noninvasive diagnosis equation for nonalcoholic fatty liver disease（NAFLD） using related clinical and laboratory markers,and to investigate the value of this equation in the prediction and diagnosis of NAFLD. Methods A total of 127 patients who were diagnosed with NAFLD in The First Affiliated Hospital of Xi＇ an Medical University from November 2016 to November 2017 were enrolled,and 30 healthy individuals were enrolled as healthy controls. Related data were recorded,including sex,age,body mass index（BMI）,medical history,alanine aminotransferase（ALT）,aspartate aminotransferase（AST）,gamma-glutamyl transpeptidase（GGT）,blood urea nitrogen（BUN）,uric acid（UA）,serum creatinine（SCr）,total cholesterol（TC）,triglyceride（TG）,high-density lipoprotein（HDL）,low-density lipoprotein（LDL）,Hb Alc,free fatty acid（FFA）,fasting blood glucose（FPG）,fasting insulin（FINS）,platelet count（PLT）,and results of ultrasound examination and Fibro Scan examination. The t-test was used for comparison of continuous data between groups; the Pearson correlation analysis was performed to investigate correlation; the multiple linear regression equation model was used to establish the regression equation; the receiver operating characteristic（ROC） curve was plotted to calculate the sensitivity and specificity of this regression equation. Results The indices related to fatty liver included BMI（r = 0. 308,P = 0. 005）,ALT（r = 0. 379,P 0. 001）,AST（r = 0. 318,P = 0. 004）,GGT（r = 0. 293,P = 0. 009）,UA（r = 0. 244,P = 0. 033）,and FFA（r = 0. 249,P =0. 030）. A multiple regression analysis was performed for controlled attenuation parameter（CAP） on Fibro Scan; the regression model of CAP had statistical significance（F = 11. 113,P 0. 001）,and its adjusted determination coefficient R2 was 0. 274,suggesting that the variation caused by regression accounted for 27. 4% of all variations; ALT had the greatest influence on CAP（β = 0. 358,P = 0. 001）,followed by BMI（β = 0. 258,P = 0. 012）. The regression equation established was CAP = 113. 163 ＋ 0. 252 × ALT ＋ 6. 316 × BMI. This diagnostic equation had an area under the ROC curve of 0. 927,a sensitivity of 87. 68%,and a specificity of 90. 00%,at the cut-off value of277. 67（P 0. 001）,suggesting that it had high diagnostic efficiency. Conclusion Compared with current diagnosis equations,the equation established in this study has a larger area under the ROC curve and higher specificity and sensitivity. It also has a simple calculation method and strong practicability and operability and helps to screen out early NAFLD and improve the awareness of self-intervention,which can further reduce the harm and delay the progression of NAFLD in the world.