In patients with chronic hepatitis B (CHB), hepatitis B surface antigen (HBsAg) seroconversion is one of the ultimate goals of antiviral therapy. However, this is only achievable in a small proportion of patients receiving treatment. Other end-points that are commonly used include the normalization of alanine aminotransferase (ALT), viral suppression, and hepatitis B e antigen (HBeAg) seroconversion. However, HBeAg seroconversion is an inadequate end-point because it does not guarantee long-term remission and inactivity of the hepatitis B virus (HBV). Thus, although HBeAg seroconversion remains an important milestone in the natural history of CHB infection, a significant proportion (30–50%) of patients will either have ongoing active disease immediately after HBeAg seroclearance or undergo reactivation following a variable period of quiescence. The exact immunological mechanism for the continuation of disease activity after HBeAg seroconversion is not known. However, the continuing viral replication might be partly explained by the spontaneous mutations in the precore or core promoter regions that reduce the production of HBeAg. It has been shown that the precore and core promoter mutations start to develop even before HBeAg seroconversion. In Asian HBeAg-negative patients with detectable HBV DNA, 50–60% have the precore mutations, and up to 70% have the core promoter mutations. However, approximately 10% still have both precore and core promoter wild-type sequences.
There are considerable differences between the current major regional treatment guidelines as to the criteria for stopping therapy in both HBeAg-positive and -negative patients with CHB. This highlights the fact that there is no consensus regarding the treatment end-points. The guidelines will continue to evolve with increasing understanding of the natural history of CHB infection.
For HBeAg-positive patients, the current European Association for the Study of the Liver (EASL) guidelines1 state that the ideal end-point of therapy is sustained HBsAg loss with or without seroconversion. In contrast, the American Association for the Study of Liver Diseases (AASLD)2 and Asian Pacific Association for the Study of the Liver (APASL)3 guidelines view “durable” HBeAg seroconversion as an adequate end-point. The APASL guidelines also specify that the HBeAg seroconversion should be accompanied by undetectable HBV DNA.
However, the question remains whether treatment-induced HBeAg seroconversion together with virological response is truly durable. Even with pegylated interferon therapy, in a 5-year long-term follow-up study, despite the cumulative incidence of HBeAg seroconversion of up to 60%, only 29% of patients had HBV DNA <20 000 IU/mL, and 13% had HBV DNA <20 IU/mL.4 With lamivudine (LAM), an early study of 34 Korean patients who stopped LAM after HBeAg seroconversion showed a cumulative relapse rate of 49% after 2 years.5 Another study of 82 patients from Taiwan showed that 48% had relapsed by 12 months, with genotype C infections being associated with higher rates of relapse.6 A study of 132 Korean patients reported a relapse rate of 66% at 12 months after stopping LAM.7 A more recent Korean study of 178 patients who stopped LAM after achieving HBeAg seroconversion showed a lower relapse rate of 30% at 5 years.8 In 125 Chinese patients who stopped LAM after HBeAg loss or seroconversion, the 4-year cumulative relapse (defined as serum HBV DNA ≥ 2000 IU/mL) rates were 41% and 29%, respectively.9
Common to all these studies is the fact that older age and shorter duration of consolidation therapy after HBeAg seroconversion are associated with higher rates of relapse. However, it is important to note carefully the definition of relapse in the different studies. The low 5-year relapse rate of 30% in the previously-described study is likely due to the high cut-off level of >28 000 IU/mL used to define the reappearance of HBV DNA. When virological rebound was defined strictly as a 1 log increase in HBV DNA, the virological rebound rate was 82% at 4 years in a study of 22 patients who stopped LAM after HBeAg seroconversion. Seventy-eight percent of patients had undetectable HBV DNA at the time of last follow up in those who continued with LAM, compared to 0% in those who stopped (P < 0.001).10 Altogether, these studies show that off-treatment response after HBeAg seroconversion is not durable.
An argument can be made for those in the younger age group in whom antiviral therapy might be stopped with a lower risk of relapse after an extended consolidation treatment period. However, there is no agreement as to the acceptable age cut-off and the length of consolidation. Soo et al. reported a 31% cumulative relapse rate after 2 years following LAM cessation in 85 CHB patients who had received at least 24 months of consolidation.11
For HBeAg-negative patients, both the EASL and AASLD guidelines recommend that treatment should be continued until HBsAg clearance is achieved.1,2 The APASL guidelines, however, are less definite, and recommend that consideration should be given to stopping treatment if HBV DNA remains undetectable on three separate occasions, 6 months apart.3 In this issue of the Journal of Gastroenterology and Hepatology, Liu and colleagues reported the relapse rates of 61 HBeAg-negative CHB patients in whom LAM was stopped.12 These patients were treated for at least 24 months, and had undetectable HBV DNA (<200 IU/mL) and normal ALT levels for at least 18 months. Relapse was defined as HBV DNA ≥2000 IU/mL. In their cohort, 31 (51%) patients suffered relapse, with a cumulative relapse rate of 52% after 3 years. Similar to previous studies, younger age was associated with a lower relapse rate. The authors conclude that despite the cessation criteria recommended by the APASL guidelines, the maintenance of viral suppression was not durable.
Another recent study using less stringent cessation criterion was associated with a similarly high relapse rate.13 In this study, of those who achieved a protocol-defined response (HBV DNA <1.4 × 105 IU/mL and ALT <1.25 × upper limit of normal) with entecavir or LAM at 48 weeks, seven of 257 (3%) entecavir-treated and 10 of 201 (5%) LAM-treated patients sustained HBV DNA <60 IU/mL at 24 weeks off treatment. In contrast, those who continued treatment into year 2 had maintenance of virological suppression.13
Currently, HBsAg seroclearance remains an elusive goal for the majority of patients treated with oral antiviral agents or with pegylated interferon. Despite this, durable suppression of HBV DNA is now achievable with long-term antiviral therapy. The importance of viral load on long-term outcome cannot be over-emphasized, with evidence showing that the lower the HBV DNA, the lower the risk of hepatocellular carcinoma and cirrhosis development.14–16 Early concerns regarding the development of drug-resistant mutations with long-term treatment have largely been mitigated by the availability of highly-potent antiviral agents with a high genetic barrier to resistance, such as entecavir and tenofovir. Relapse, despite continual therapy after HBeAg seroconversion, is usually preceded by the development of resistance in the majority, and is often seen in patients treated with LAM monotherapy.17 However, the resistance rate is lower than that observed in HBeAg-positive patients. After treatment-induced HBeAg seroconversion, the resistance rate was reported to be 10% after a median treatment length of 79 months.10
Virological rebound following cessation of antiviral therapy can be associated with negative consequences. First, inadequate monitoring can result in severe flares of hepatitis. Second, re-challenging the HBV with the same drug after rebound of viral load can theoretically increase the chance of drug resistance. In regions where expensive antiviral drugs might not be readily available as first-line treatment, long-term treatment with LAM might be the only option. Long-term therapy is still advisable, since the risk of relapse from stopping therapy is greater than that of resistance. The adoption of the roadmap concept, with testing of HBV DNA at week 24, might further minimize resistance.
In summary, the study by Liu and colleagues has provided further evidence that off-treatment virological response is not durable, even with adherence to strict cessation criteria. For both HBeAg-positive and -negative patients, the ideal treatment end-points in the era of potent antiviral therapy with low resistance should be the seroclearance of HBsAg.