四种诊断模型诊断乙型肝炎病毒携带者肝纤维化的效能分析

Diagnostic efficacy of four non-invasive diagnostic models in evaluating significant liver fibrosis in chronic hepatitis B virus carriers

目的探讨应用基于4因子模型(FIB-4)、肝脏瞬时弹性探测仪(FS)、天冬氨酸氨基转移酶/丙氨酸氨基转移酶比率(AAR)和天冬氨酸氨基转移酶/血小板比值指数(APRI)诊断乙型肝炎病毒(HBV)携带者肝纤维化的效能。方法2015年7月~2017年11月我院收治的203例慢性HBV携带者,均符合《慢性乙型肝炎防治指南》中相关诊断标准。使用FS行肝硬度测定(LSM),并根据临床检测结果,获得FIB-4、APRI和AAR计算结果。常规行肝活检,行肝组织病理学检查,将≥S2定义为显著肝纤维化。应用ROC曲线分析4种指标诊断显著肝纤维化的效能。结果在203例慢性HBV携带者中,经病理学检查,诊断为S0期24例(11.8%),S1期143例(70.4%),S2期32例(15.8%),S3期3例(1.48%),S4期l例(0.5%),其中≥S2期者36例(17.7%);36例≥S2期患者FIB-4为(2.2±1.6),显著大于167例<S2期的(1.3±0.9),显著肝纤维化人群LSM、AAR和APRI均显著大于无显著肝纤维化人群【分别为(15.3±5.1)对(6.3±4.2)、(0.8±0.4)对(0.5±0.1)、(1.0±0.6)对(0.5±0.2),P<0.01】;ROC曲线分析显示,LSM诊断显著肝纤维化的AUC为0.967(95%CI 0.945~0.990),其灵敏度为94.4%,特异度为86.0%,AAR诊断的AUC为0.8(95%CI 0.821~0.966),其灵敏度为72.2%,特异度为99.4%,APRI诊断的AUC为0.884(95%CI 0.810~0.958),其灵敏度为72.2%,特异度为99.8%,FIB4诊断的AUC为0.792(95%CI 0.689~0.895),其灵敏度为69.4%,特异度为90.1%。结论应用无创诊断模型诊断HBV携带者显著性肝纤维化有一定的意义,对指导临床处理有帮助,但还需大样本多中心研究证实。

Objective To investigate the diagnostic efficacy of four non-invasive diagnostic models in evaluating significant liver fibrosis in chronic asymptomatic hepatitis B virus carriers(ASC).Methods 203 chronic ASCs were recruited in ourhospital between July 2015 and November 2017,and FIB-4 scoring systems,liver stiffness measure(LSM)by Fibroscan,aspartate aminotransferase/alanine aminotransferase ratio(AAR)and aspartate aminotransferase/platelet ratio index(APRI)were obtained.The liver biopsies were performed,and≥S2 was defined as significant liver fibrosis.The diagnostic efficacy was evaluated by area under ROC.Results Out of the 203 chronic ASCs,the liver histopathological examination showed that no liver fibrosis in 24(11.8%),S1 in 143(70.4%),S2 in 32(15.8%),S3 in 3(1.48%)and S4 in l(0.5%),with 36(17.7%)having≥S2 liver fibrosis;the FIB-4 in 36 ASCs with≥S2 was(2.2±1.6),significantly higher than(1.3±0.9)in 167 with less than S2,and the LSM,AAR and APRI scores in ASCs with significant liver fibrosis were all significant higher than in those with less than S2 liver fibrosis[(15.3±5.1)vs.(6.3±4.2),(0.8±0.4)vs.(0.5±0.1),and(1.0±0.6)vs.(0.5±0.2),P<0.01];the AUC in predicting significant liver fibrosis by LSM was 0.967(95%CI 0.945 to 0.990),with sensitivity(Se)of 94.4%and specificity(Sp)of 86.0%,and that by AAR was 0.8(95%CI 0.821 to 0.966),with Se of 72.2%and Sp of 99.4%,by APRI was 0.884(95%CI 0.810 to 0.958),with Se of 72.2%and Sp of 99.8%,and by FIB4 was 0.792(95%CI 0.689 to 0.895),with Se of 69.4%and Sp of 90.1%.Conclusions The application of the four non-invasive diagnostic model in predicting significant liver fibrosis in ASCs might help clinical managements,which warrants further multi-centre investigation.