肝纤维化无创模型对丙氨酸氨基转移酶低于2倍正常值上限慢性乙型肝炎患者治疗时机的判定价值

Value of non-invasive models of liver fibrosis in judgment of treatment liming in chronic hepatitis B patientswith ALT 〈 2xupper limit of normal

目的评价无创肝纤维化模型FIB-4、S指数、APRI、GP模型，APGA、PAPAS对治疗时机难以确定的丙氨酸氨基转移酶（ALT）〈2×正常值上限（ULN）慢性乙型肝炎患者显著肝纤维化的诊断价值，帮助判定该人群的治疗时机。方法389例治疗时机难以确定的ALT〈2×ULN慢性乙型肝炎患者，经皮肝穿刺活组织检查取得其病理结果，分别对其进行血常规、血清生物化学指标、乙型肝炎病毒学标志物、HBVDNA等常规血清学检测。根据肝组织病理学分为非显著肝纤维化组324例（s〈2）和显著肝纤维化组65例（s≥2），参考原始文献构建预测肝纤维化的无创模型。统计学处理采用SPSSl9．0软件，采用受试者工作特征（ROC）曲线方法评价比较不同无创模型预测该人群显著肝纤维化的价值。结果各无创模型对本研究人群肝纤维化程度都具有一定诊断价值，各模型APRI，FIB-4，APGA，S指数，PAPAS和GP模型的曲线下面积分别为0．718，0．691，0．758，0．729，0．673和0．691。其中APGA的ROC曲线下面积最大（0．758，95％CI：0．673～0．844），各变量中γ-谷氨酰转肽酶与肝组织纤维化程度存在显著正相关。结论肝纤维化无创模型能较好判断治疗时机难以确定的ALT〈2×ULN慢性乙型肝炎患者的显著肝纤维化，帮助该人群治疗时机的判定，其中以APGA模型更佳，能在一定程度上减少肝活组织检查的需求。

Objective To investigate the value of non-invasive liver fibrosis models, FIB-4, S index, aspartate aminotransferase to platelet ratio index （APRI）, globulin-platelet （GP） model, aspartate aminotransferase/platelet/gamma-glutamyl transpeptidase/alpha-fetoprotein （APGA）, and platelet/age/ phosphatase/alpha-fetoprotein/aspartate aminotransferase （PAPAS）, in the diagnosis of marked liver fibrosis in chronic hepatitis B （CHB） patients with ALT 〈 2xupper limit of normal （ULN）, as well as treatment timing for this populatiorL Methods A total of 389 CHB patients with ALT 〈 2xULN who were admitted to Beijing Ditan Hospital and whose treatment timing was difficult to judge were enrolled. Transdermal liver biopsy was performed to obtain pathological results, and routine serological tests were performed, including routine blood test, serum biochemical parameters, hepatitis B virus （HBV） markers, and HBV DNA. According to liver pathology, the patients were divided into non-marked liver fibrosis group （S 〈 2） with 324 patients and marked liver fibrosis group （S ≥ 2） with 65 patients. The non-invasive models for predicting liver fibrosis was established with reference to original articles. SPSS 19.0 software was used for statistical analysis, and the receiver operating characteristic （ROC） curve was used to compare the value of different non-invasive models in predicting marked liver fibrosis in this population. Results All the non-invasive models had a certain diagnostic value for liver fibrosis degree in these patients, and the areas under the ROC curve for APRI, FIB-4, APGA, S index, PAPAS, and GP model were 0.718, 0.691, 0.758, 0.729, 0.673, and 0.691, respectively. APGA had the largest area under the ROC curve （0.758, 95% CI 0.673-0.844）, and gamma-glutamyl transpeptidase was significantly positively correlated with liver fibrosis degree. Conclusion The non-invasive models of liver fibrosis can identify marked liver fibrosis in CHB patients with ALT 〈 2×ULN in whom it is difficult to judge treatment timing and help to determine treatment timing for them. APGA model has the highest value and can reduce the need for liver biopsy to the certain degree.