With successful nucleos(t)ide analog therapy, serum hepatitis B virus (HBV) DNA titer decreases to below detection limit, alanine aminotransferase (ALT) normalizes, and fibrosis of the liver may become improved [1–4]. The incidence of hepatocellular carcinoma can be decreased as well [5–7].

However, withdrawal hepatitis frequently occurs after treatment cessation, even in patients with undetectable HBV DNA levels. As this complication is sometimes lethal [8–10], it is crucial to determine the optimal conditions for stopping nucleos(t)ide analog therapy, especially for young patients wishing to have children, the avoidance of hyper-resistant viral strains, and regulating medical expenses.

Knowledge of intrahepatic HBV replication status is the key to defining the safe conditions for nucleos(t)ide analog therapy cessation. cccDNA level is a suitable indicator for this aim, although frequent liver biopsies are needed. Serum HBV DNA was believed to be a good estimator of cccDNA. However, the correlation of HBV DNA and intrahepatic cccDNA levels is lost under the nucleos(t)ide analog treatment and cannot therefore be adopted [11]. Alternative, non-invasive markers are desired.

The possibility of discontinuation of nucleos(t)ide analogs is also affected by the clinical goal, such as no withdrawal hepatitis, no viral relapse, or long-term clinical response. The majority of patients experience viral recurrence during withdrawal, but some will not develop hepatitis or will achieve clinical remission after a hepatitis flare.

In this issue, Jung KS et al. [12] report on virological remission (i.e., HBV DNA ≤2000 IU/ml) over 1 year following cessation of nucleos(t)ide therapy in patients HBeAg-positive or HBeAg-negative at the start of treatment. The virological relapse rate of each group was 57.8 and 54.4 %, respectively. Significantly and independently related factors for virological relapse were age >40 years at the start of nucleos(t)ide therapy and basal HBV DNA level >20,000,000 IU/ml in the HBeAg-positive group, and age >40 years at the start of nucleos(t)ide analog therapy and HB core-related antigen (HBcrAg) level >3.7 log U/ml in the HBeAg-negative group.

HBcrAg is an established viral activity marker [13]. HBcrAg levels correlated closely with those of intrahepatic cccDNA, even when HBV DNA became undetectable, during nucleos(t)ide analog therapy [11]. Thus, HBcrAg may represent a non-invasive marker of cccDNA with or without nucleos(t)ide analog therapy. We have also reported that a combination of HBsAg and HBcrAg levels at the end of nucleos(t)ide analog therapy was useful to predict long-term clinical remission [14, 15].

According to APASL guidelines, the appropriate duration of nucleoside analog administration is unknown, although HBV DNA and HBsAg are suggested as estimators of treatment efficacy [16]. There are no precise criteria for the cessation of nucleos(t)ide analog therapy in the EASL guidelines as well [17]. In the AASLD guidelines, HBsAg loss is regarded as the best predictor of sustained remission off-treatment, but is also reported to be infrequent with current therapies [18].

In this issue, the authors present clear criteria for successful discontinuation of nucleos(t)ide analog therapy using HBcrAg, which is expected to contribute to the enhancement of chronic hepatitis B patient management. Their study also underscores the importance of measuring multiple HBV antigens in addition to HBV DNA to better estimate the behavior of HBV in the liver.