Letter: clinical outcomes of HBsAg loss in chronic HBV infection.

Abstract

Hepatitis B surface antigen (HBsAg) loss is a milestone that marks the natural history of chronic hepatitis B virus (HBV) infection. Therefore, HBsAg loss is the most desired endpoint in chronic hepatitis B (CHB) but occurs uncommonly.1 Its estimated annual incidence is 0.45–2.38%.2 HBsAg loss is associated with excellent clinical outcomes. Baseline HBsAg, HBV DNA, older age and platelet counts are associated with higher chance of spontaneous HBsAg loss.3-6 And baseline low HBsAg levels is the strongest predictor of spontaneous HBsAg loss compared to other factors.3, 4, 6 Moreover, rates of spontaneous HBsAg loss will be higher along with follow-up.3 The most importance is that spontaneous loss of HBV DNA and HBsAg are important predictors of reduced HCC risk.1, 7 In a recent issue of Aliment Pharmacol Ther, Chen and coworkers8 conducted a case–control study to explore whether there is difference of clinical outcomes between patients with spontaneous and nucleos(t)ide analogue (NAs) treated HBsAg loss. This cohort study was based on 312 CHB patients with spontaneous HBsAg loss and 110 patients with nucleos(t)ide analogue treated HBsAg loss. After a mean follow-up period of 107 months after HBsAg loss, the clinical outcomes (including the incidence of anti-HBs seroconversion, HCC development, overall mortality and variceal bleeding) between patients in the two groups are comparable. Propensity score analysis revealed similar results. This is the first study with long-term follow-up to compare the difference in the clinical outcomes after HBsAg loss between CHB patients with spontaneous and nucleos(t)ide analogue-treated HBsAg loss.8 And this is the main strength of this cohort study. However, the main drawback is the less information about HBV DNA of the included patients. Persistent virus replication and cirrhosis are the two main risk factors of HCC development. HBV covalently closed circular DNA (cccDNA) is responsible for viral persistence. On the other hand, cccDNA is strongly associated with the level of HBV DNA. Therefore, serum HBV DNA, but not HBsAg, reflects the amount of cccDNA and the replication efficiency of HBV in patients with CHB.9 This may explain that HCC can occur in the absence of cirrhosis. Considering the indications of anti-viral therapy with nucleos(t)ide analogues, more patients in the NAs treated group might be with positive or higher HBV DNA than those in the spontaneous HBsAg loss group. Therefore, more patients in the NAs treated group were with cirrhosis (20%) and HCC development (3.6%) than others (7.1% and 0.3%, respectively). Also, patients in the nucleos(t)ide analogues treated group might have higher chance of HCC development in the long run because of the pre-existing fibrosis before and during nucleos(t)ide analogue therapy.8 HBsAg loss was defined as persistent negativity of HBsAg for at least 1 year.8 Patients developed HCC dozens of months after HBsAg loss. The level of HBsAg (or ±) at the time of HCC development was not described. The levels of HBsAg and HBV DNA are connected with necessity of adjuvant anti-viral therapy.10 We applaud Chen and coworkers’ interesting and important work. However, due to the small sample size (especially after propensity score analysis) and limitations of retrospective analysis, more prospective studies with large sample size are warranted to certify these findings. Declaration of personal and funding interests: None.