We have shown here that the incidence of post-treatment biochemical breakthrough as well as virologic relapse was low in CHB patients who showed undetectable serum HBV DNA (<50 copies/mL) at cessation of lamivudine after sufficient lamivudine treatment. Notably, initially HBeAg-negative or female patients who achieved low viral titer after sufficient lamivudine treatment showed low post-treatment biochemical breakthrough rate. In addition, we found that hepatitis flare was infrequent and antiviral retreatment for CHB patients with hepatitis flare was safe and effective although virologic relapse and biochemical breakthrough were common after cessation of lamivudine therapy.
The present study has distinctive features that we analyzed the comprehensive long-term post-treatment clinical course including hepatitis flare and retreatment results both in initially HBeAg-positive and HBeAg-negative patients. Recently, cessation of antiviral agent discussed in the previous studies
[12, 13, 19, 20], but biochemical breakthrough, hepatitis flare, and retreatment outcome were not closely investigated, especially in HBeAg-negative patients. Post-treatment virologic relapse and biochemical breakthrough incidence rates in our study were low in patients who achieved undetectable serum HBV DNA level and were lower than those of previous studies
[12, 13, 21–24]. Moreover, patients in our cohort were regularly and closely followed up after cessation of lamivudine. Considering good results of NA retreatment in relapsed patients, our data could be one evidence for the importance of early detection and early treatment of biochemical breakthrough before hepatitis flare with close follow-up after cessation of lamivudine, especially in patients with predisposing factors for post-treatment biochemical breakthrough as well as virologic relapse.
Despite that achievement of undetectable serum HBV DNA level has been recommended as the endpoint of CHB treatment
[3, 8, 14, 15], in fact, many clinicians have used lamivudine continuously due to potential virologic relapse, biochemical breakthrough, or hepatitis flare that can occur after cessation of lamivudine. However, life-long lamivudine treatment can induce drug resistance mutation
, and Asian patients tended to show less durable serologic response than Western patients despite long-term lamivudine use
. Considering these practical clinical settings, our results can provide useful evidence for determining who can stop and when to stop antiviral treatment. Furthermore, our data could strengthen the previous recommendations that lamivudine cessation might be reasonable option both in initially HBeAg-positive and HBeAg-negative patients who achieved low viral titer after sufficient lamivudine treatment.
In terms of hepatitis flare after cessation of antiviral agents, it has been one of the important concerns for clinicians when they stopped antiviral agents in CHB patients. It occurred in about 17-19% during the median follow-up period of 3-6 months after lamivudine cessation in the previous studies
[12, 16, 24–26], but they had limits due to the small cohorts or short-term follow-up period. In contrast, in our study, hepatitis flare was observed only in 8.7% of all enrolled off-treated CHB patients during the median follow-up period of 29 months. Thus, our results might be more acceptable because our outcome was based on the long-term follow-up period. In spite of our results, we have to say that the cessation of antiviral agents should be very cautious in patients with advanced liver disease because hepatitis flare could cause devastating outcome in these patients
. There should be sufficient discussion between doctor and patients about risk and benefit of continuation or cessation of antiviral agents, and frequent follow-up scheme should be applied to these patients with advanced liver disease if antiviral agent is cessated.
In multivariate analysis, serum HBV DNA level at the cessation of lamivudine was independent predisposing factor for post-treatment virologic relapse and biochemical breakthrough. In addition, younger age at lamivudine cessation was inversely associated with the post-treatment virologic relapse, which was similar to the previous studies
[12, 19], suggesting that younger patients have more powerful immune responses than older patients. Interestingly, we found that HBeAg-negative or female patients who achieved low viral titer were inversely associated with post-treatment biochemical breakthrough. The relationship of female with lower risk of post-treatment biochemical breakthrough was a unique finding in our study that has not been discussed in the previous literature. This result might be helpful for female CHB patients of childbearing age. Although we did not find the accurate mechanism for this relationship, hormonal difference might have influence on this association. In addition, given the low incidence rate of post-treatment biochemical breakthrough despite frequent virologic relapse in HBeAg-negative patients of our study, the association of HBeAg-negative patients with low post-treatment biochemical breakthrough might be explained by different virologic virulence or immune response between initially HBeAg-positive and HBeAg-negative patients. However, there need to be further studies to elucidate the mechanism for these relationships in the future.
With regard to antiviral responsiveness in CHB patients who received antiviral retreatment, it is an additional important consideration when clinicians determine the cessation of antiviral therapy. To date, some previous studies have reported about the clinical outcome after retreatment, but the number of patients was smaller without discussion of HBeAg-negative patients, or the follow-up period after retreatment was shorter compared to that of our study
[12, 19]. Interestingly, we found that antiviral retreatment was safe and effective even in patients with hepatitis flare. Moreover, most re-treated patients achieved virologic and biochemical response rapidly. Therefore, our results suggest that lamivudine might be cautiously discontinued and intermittent antiviral retreatment might be considered in appropriately selected patients.
In terms of HBV genotype, which has been known to be associated with the antiviral response to oral NA(s), several previous researches performed in South Korea have already demonstrated that approximately 100% of the CHB patients in Korea had genotype C
. A vast majority of patients in our cohort could be considered to be infected by HBV genotype C. Thus, in the present study, the relationship between HBV genotype and post-treatment outcomes after lamivudine cessation could not be analyzed.
This study had some limitations. First, our study is a retrospective one that selection bias could be possible For example, median serum HBV DNA level was not different between patients with initial HBeAg positive and HBeAg negative patients at the initiation of LAM treatment as shown in Table
1, which was different from the previous reports
[28, 29]. This finding might be explained by the retrospective design or relatively small number of initial HBeAg negative patients. However, we tried to avoid selection bias by enrolling all the consecutive patients including those who showed serum HBV DNA level >104 copies/mL at cessation of lamivudine in a given period. In addition, prospective trial for off-treatment of NA is difficult to be performed because of potential ethical problem. Second, we could not compare antiviral efficacy among the patients who were retreated by entecavir, lamivudine, clevudine, or adefovir because of the small number of the patients. There need to be further well-designed randomized prospective studies with large sample size to compare the antiviral efficacy among retreatment drugs.