QUANTIFICATION OF HEPATITIS B SURFACE ANTIGEN AS A PREDICTOR OF OFF-TREATMENT SUSTAINED VIROLOGICAL RESPONSE IN CHRONIC HEPATITIS B PATIENTS TREATED WITH ORAL NUCLEOS(T)IDE ANALOGUE

초록

Background and Aims: The absolute value or reduction of hepatitis B surface antigen (HBsAg) level can be a useful marker for on-treatment response or durability of off-treatment response. We examined the appropriate cutoff value of serum HBsAg level that predict off-treatment sustained virological response in patients with chronic hepatitis B (CHB) treated with oral nucleos(t)ide analogues (NAs). Methods: From June 1998 to Dec 2010, 81 patients with CHB who discontinued NAs were enrolled [HBeAg (+), n = 50(62%)]. Oral NAs were lamivudine (n = 53), adefovir (n = 15), lamivudine combined with adefovir (n = 4), entecavir (n = 9). The mean duration of antiviral treatment and the follow-up period after discontinuation of treatment were 31±16 and 32±24 months respectively. The sustained virological response after treatment discontinuation (SVR) was defined as follows; complete (cSVR, HBV DNA undetectable) or partial (pSVR, normal ALT and detectable HBV DNA <20,000 IU/ml in HBeAg positive and HBV DNA <2,000 IU/ml in HBeAg negative). Serum HBsAg levels were determined by the ARCHITECT i2000 HBsAg (Abbott, USA, 0.09–250 IU/mL). Results: Thirty-three of 81 (40%) patients had complete or partial SVR [cSVR, 11/81 (14%); pSVR 22/81 (27%)]. The relapse rate were 37/81 (46%), 5/81 (6%) at 6 months, 12 months after discontinuation of treatment. Baseline HBV DNA, ALT, presence of HBeAg were not different in patients with or without cSVR. The mean treatment duration, the reduction of HBsAg level at 6 month on-treatment, and HBsAg level at the end of treatment were significantly different in patients with or without cSVR (47 vs. 28 months, P = 0.011; −0.9 vs. +0.2 log IU/ml, P < 0.05; 0.2 vs. 3.5 log IU/ml, P < 0.05). In multivariate analysis, only HBsAg level at the end of treatment was significant (P = 0.007). Serum HBsAg level <2 log IU/ml at the end of treatment were predictive of SVR [(AUROC, 0.991; 95% confidence interval [CI], 0.000–1.000; P < 0.05); sensitivity, 100%; specificity, 93%; positive predictive value, 69%; negative predictive value, 100%]. Conclusions: In CHB patients with long term nucleos(t)ide analogue therapy, absolute value of HBsAg off-treatment can be a predictive of SVR, and treatment discontinuation can be considered.