摘要目的：探讨 HBeAg 阴性 CHB 患者在干扰素抗病毒治疗过程中 HBsAg 水平的变化对治疗中 HBsAg 消失或血清学转换的预测作用。方法2008年5月至2011年8月于首都医科大学附属北京地坛医院就诊的 HBeAg 阴性 CHB 初治患者，HBsAg 阳性、血清 HBV DNA 阳性持续6个月以上， ALT 异常3个月以上，收集的患者包括聚乙二醇干扰素-α-2a （peg-IFN-α-2a）治疗3个月以上的在治疗患者和已治疗结束的患者。入组患者给予 peg-IFN-α-2a 180μg／周皮下注射治疗，治疗前和治疗期间每间隔3个月检测血清 HBV DNA、HBsAg 水平和抗-HBs，以 HBsAg 消失为主要疗效评价指标。组间差异采用卡方检验，应用受试者工作特征（ROC）曲线预测治疗12周和治疗24周的 HBsAg 水平绝对值和HBsAg 下降幅度对治疗96周和120周时 HBsAg 的消失。结果共有81例患者进入数据分析，根据患者治疗中 HBsAg 下降所达到的水平分为完全应答组12例（14．8％），部分应答组20例（24．7％）和应答不佳组49例（60．5％）。治疗12周完全应答组和部分应答组 HBsAg 水平值下降幅度分别为0．62（0．06，0．91）lg IU／mL 和0．19（－0．01，0．48）lg IU／mL，两组比较差异无统计学意义（Z ＝1．581，P ＝0．209），在调整基线的差异后（计算患者下降幅度的中位数），完全应答组和部分应答组下降幅度均高于应答不佳组的下降幅度0．00（－0．01，0．14）lg IU／mL，χ2＝9．00，P ＜0．01]。治疗24周，3组间 HBsAg的下降幅度差异均有统计学意义（χ2＝27．72，P ＜0．01）。无论预测治疗96周还是120周治疗的HBsAg 消失，24周HBsAg 临界值的预测能力均强于12周，ROC 曲线下面积均差异有统计学意义（χ2值分别为＝3．880和4．412，P 值分别为0．049和0．036）。结论治疗早期 HBsAg 水平变化可预测peg-IFN-α-2a 治疗 HBeAg 阴性 CHB 患者疗效。

abstractsObjective To investigate the predictive value of on-treatment hepatitis B surface antigens (HBsAg)levels for HBsAg loss/seroconversion in hepatitis B e antigen (HBeAg)-negative chronic hepatitis B (CHB)patients treated with pegylated interferon (peg-IFN)-α-2a.Methods HBeAg-negative CHB naive patients were collected from 2008 May to 2011 August in Beijing Ditan Hospital affiliated to Capital Medical University.Eligible patients had been HBsAg-positive,HBeAg-negative and with detectable hepatitis B virus (HBV)DNA for>6 months,had elevated alanine aminotransferase for>3 months,had no other liver diseases,and had been treated with peg-IFN-α-2a>3 months or had finished the treatment.All the enrolled patients were administered with peg-IFN-α-2a 180 μg/week.Serum HBV DNA,HBsAg,and anti-HBs levels were measured at baseline and every 3 months during peg-IFN-α-2a&nbsp;treatment.The primary efficacy of response to the treatment was HBsAg loss.Comparison between groups was conducted by chi-square test.The absolute levels of HBsAg and the decline rate of HBsAg at week 12 and week 24 were assessed using receiver operating characteristic (ROC)curve for predicting HBsAg loss at week 96 and week 120.Results A total of 81 patients were included into the data analysis. According to on-treatment HBsAg decline,patients were divided into completed response group (12 cases, 14.8%),partial response group (20 cases,24.7%),and poor response group (49 cases,60.5 %).On week 12 of treatment,mean HBsAg decline levels in completed response and partial response groups were 0.62(0.06,0.91)lg IU/mL and 0.19 (-0.01 ,0.48)lg IU/mL,respectively,which was not statistical different (Z =1 .581 ,P =0.209 ).Adjusting for differences of baseline factors (calculation of the mean decline of HBsAg),completed response group and partial response group had much more pronounced decline than poor response group 0.00 (-0.01 ,0.14)lg IU/mL,χ2 =9.00,P <0.01 ].On week 24, difference of HBsAg decline among the 3 groups was statistically significant (χ2 =27.72,P <0.01).The predictive ability of HBsAg cut-off level at week 24 was better than week 12 for HBsAg loss at either 96 or 120 weeks,and the differences of the area under the ROC curve were both significantly different (χ2 =3.880,P =0.049 and χ2 =4.412,P =0.036,respectively).Conclusion Dynamics of HBsAg level in early treatment of peg-IFN-α-2a can predict the therapeutic effect of patients with HBeAg-negative CHB.