Introduction

More than 350 million people worldwide are infected with hepatitis B virus (HBV) which is a leading cause of liver-related morbidity that accounts for 1 million deaths annually [1, 2]. Currently available antiviral treatment for chronic hepatitis B can be divided into two classes of therapeutic agents: nucleos(t)ide analogues (NAs) and interferon (IFN). Nucleosides include lamivudine [3, 4], telbivudine [5], and entecavir [6, 7]; nucleotides include adefovir [8, 9] and tenofovir [10]. The direct, potent antiviral effects of NAs induce an on-treatment response in most patients, but post-treatment relapse commonly occurs after treatment discontinuation. Long-term use of NAs triggers the emergence of drug-resistant variants possessing mutations in the hepatitis B virus (HBV) polymerase gene. Among the NAs currently available, entecavir or tenofovir is recommended as the first-line treatment because of the low rate of drug resistance. In contrast, IFN has both antiviral and immunomodulatory actions [11, 12]. The major advantages of IFN include a finite course of treatment, absence of drug resistance, and an opportunity to obtain a post-treatment durable response to therapy; however, a response to IFN is achieved in only a minority of patients with chronic hepatitis B.

In this article, we have reviewed previous reports on combination therapy with NA and IFN for chronic hepatitis B. Regimens of combination therapy can be classified into two main groups: (1) simultaneous therapy with the drugs in the combination and (2) sequential combination therapy in which treatment with one drug follows that of the previously administered one. To compare the results of previous trials, we noted that age/sex of the included subjects, HBV genotypes, and mode of vial transmission varies among the different studies. These differences may affect the results and their interpretations as older age, male sex, HBV genotypes C and D (vs. A and B), and vertical transmission are associated with a poor response to IFN therapy [13–15]. In particular, HBV genotypes have specific geographic distributions, with genotype A being prevalent in Northwest Europe, North America, and Central Africa, genotypes B and C being common in Southeast Asia, China, Japan, and Oceania, and genotype D being prevalent in Southern Europe, the Middle East, and India, although it has a nearly worldwide distribution.

Theoretical background

Nucleos(t)ide analogues directly inhibit viral replication by targeting at least one of the three replication steps: priming of HBV DNA polymerase, reverse transcription of negative-strand HBV DNA from pregenomic RNA, and synthesis of positive-strand HBV DNA. IFN also possesses antiviral activity but does not act directly on the virus or replication complex. Instead, it acts by inducing IFN-stimulated genes to establish a non-virus-specific antiviral state within the cell. In addition to their role as antivirals, IFNs are important immunomodulators that interact with the adaptive and innate immune responses. Combining NA and IFN, with their different mechanisms of action, in a therapeutic regimen is theoretically an attractive approach for treating chronic hepatitis B.

The action of NAs has little or no effect on the decrease in the intrahepatic HBV replicative intermediate, covalently closed circular DNA (cccDNA). In the experimental woodchuck hepatitis virus system, cccDNA persisted even when viral production was strongly reduced by NA treatment [16, 17]. To reduce the level of intrahepatic cccDNA, the immunomodulatory activity of IFN, which presumably induces cytotoxic T cell activity for immune clearance of infected cells, may be required. However, a high HBV DNA load is associated with an inefficient T cell response to HBV-related antigens, such as hepatitis B surface antigen (HBsAg) [18]. Several studies have shown that a decreased viral load induced by NA treatment can result in the subsequent restoration of CD4 followed by CD8 cellular immune response against HBV [19, 20]. The rationale for combination therapy is based on the concept that suppression of viral replication by NA can decrease viral protein synthesis on the surface of hepatocytes, which may restore the immune response and optimize the immunomodulatory effects of IFN for clearing infected cells.

Simultaneous combination with NA and IFN

Table 1 shows a summary of previous studies examining simultaneous combination therapy with NA and IFN for chronic hepatitis B. The first trial was reported by Mutimer et al. [21] in the UK. Since this study was designed to assess the safety and tolerability of combination treatment, the duration of treatment was only 16 weeks, and few patients showed sustained seroconversion from hepatitis B e antigen (HBeAg) to anti-HBe (antibody to HBe) by this short-term therapy with lamivudine and IFN-α. Barbaro et al. [22] reported the results of a randomized trial conducted in Italy where the 24-week combination with lamivudine and IFN-α increased the rate of sustained HBeAg seroconversion compared to the 52-week lamivudine monotherapy (33 vs. 15 %; P = 0.014). Tatulli et al. [23] in Italy found that the 52-week combination with lamivudine and IFN-α resulted in a sustained loss of serum HBV DNA, based on the results of a solution hybridization assay, and normalization of alanine aminotransferase (ALT) in only 14 % of HBeAg-negative patients, but drug-resistant mutation variants did not emerge in any patients. However, from these previous studies, it is still unclear whether combination therapy with lamivudine and IFN confers an additional benefit compared to IFN monotherapy.

Table 1 Simultaneous combination therapy with nucleos(t)ide analogues and interferon
Full size table

Three randomized controlled trials (2 in HBeAg-positive patients [24, 25] and 1 in HBeAg-negative patients [26]) did not show that 1-year combination therapy with lamivudine and pegylated IFN-α was superior to monotherapy with pegylated IFN-α in terms of the rate of sustained response. The results of these globally conducted trials, which included many patients with various HBV genotypes, appear to be reliable. All studies found that the combination therapy had greater on-treatment viral suppression and higher rates of sustained post-treatment response than therapy with lamivudine alone, but no difference was observed in the sustained post-treatment virologic response compared to that with pegylated IFN-α alone. Janssen et al. [24], for example, found that more patients in the patient group receiving the 52-week pegylated IFN-α + lamivudine combination than in the group receiving the 52-week pegylated IFN-α monotherapy showed a response, as assessed by serum HBeAg loss at the end of treatment (44 vs. 29 %; P = 0.01). However, this difference was not sustained; 35 % of the combination-therapy group and 36 % of the monotherapy group showed a sustained HBeAg loss at the end of follow-up (P = 0.91). The trial also showed that pegylated IFN-α therapy improves liver histology, particularly in responders to therapy, but that the addition of lamivudine to therapy with pegylated IFN-α did not further improve histological outcome [27] and that genotypes C and D are associated with a lower rate of response to IFN than genotypes A and B.

To date, few studies have examined the combination of IFN and other NAs that are more potent than lamivudine. The telbivudine + IFN combination is prohibited because of the high risk of severe polyneuropathy [28]. Interestingly, Wursthorn et al. [29] from Germany found that 48-week combination therapy with adefovir dipivoxil and pegylated IFN-α led to marked decreases in cccDNA in the liver, which has been correlated with reduced HBsAg in serum. However, the rate of the post-treatment sustained response was not reported in this study. Another group from the Netherlands [30] showed that intrahepatic cccDNA levels at the end of 48-week treatment with adefovir dipivoxil and pegylated IFN-α were predictive of a sustained response defined as HBV DNA <2,000 IU/mL and normal ALT. The efficacy of combining IFN and other NAs, such as entecavir or tenofovir, remains to be elucidated.

Sequential combination starting with IFN followed by NA

Table 2 shows a summary of previous reports concerning sequential combination therapy starting with IFN followed by NAs for chronic hepatitis B. Hasan et al. reported that the rate of sustained HBeAg seroconversion was only 6.2 % in patients in Kuwait receiving IFN-α alone for 4 weeks, followed by the IFN-α + lamivudine combination for 12 weeks, and lastly by lamivudine alone for 36 weeks; this rate was similar to that observed in patients receiving lamivudine alone for 48 weeks [31].

Table 2 Sequential combination therapy starting with IFN followed by nucleos(t)ide analogues
Full size table

In contrast, a randomized trial by Chan et al. in China [32] showed that the rate of sustained virologic response, defined as HBeAg seroconversion and a HBV DNA level of <500,000 copies/mL, was 36 % in patients receiving pegylated IFN-α alone for 8 weeks, followed by the pegylated IFN-α + lamivudine combination for 24 weeks, and lastly by lamivudine alone for 28 weeks; this rate was significantly higher than that observed in patients receiving lamivudine alone for 52 weeks (14 %; P = 0.011). At the end of the treatment period, 21 % of patients in the sequential combination treatment group developed a lamivudine-resistant mutant, compared to 40 % of patients in the lamivudine monotherapy group. Follow-up of this study demonstrated that sequential combination with pegylated IFN-α followed by lamivudine maintained a higher long-term virologic response than lamivudine monotherapy for up to 3 years [33]. However, this study did not include a study arm of pegylated IFN-α alone.

Sequential combination starting with NA followed by IFN

Table 3 shows a summary of previous studies which examined sequential combination therapy starting with NAs followed by IFN for chronic hepatitis B. In a pilot study [34] by Serfaty et al. in France, sustained responses, defined as serum HBV DNA clearance based on the results of a branched DNA assay and ALT normalization, were achieved in 57 % of patients who received lamivudine alone for 20 weeks followed by the lamivudine + IFN-α combination for 4 weeks, and lastly by IFN-α alone for 24 weeks.

Table 3 Sequential combination therapy starting with a nucleos(t)ide analogue followed by IFN
Full size table

Some groups have studied similar protocols for sequential therapy, but the results have been conflicting. Consistent with the results reported by Serfaty et al. [34], Sarin et al. [35] in India reported that the addition of 4-week lamivudine before starting 24-week pegylated IFN-α therapy resulted in a significantly higher rate of sustained HBeAg clearance (39 %) than that with 24-week pegylated IFN-α monotherapy (14 %; P = 0.05). In contrast, Manesis et al. [36] found that in HBeAg-negative patients in Greece, where genotype D is predominant, the rate of sustained response to sequential therapy, defined as HBV DNA of ≤30,000 copies/mL and normal ALT, was only 22 %, which did not differ from that obtained in age/sex-matched historical controls treated with IFN-α alone for 12 months (14 %; P = 0.36). In another report in Greece [37], sequential therapy did not raise the rate of sustained virologic response, defined as HBV DNA levels of <400 copies/mL, in HBeAg-negative patients compared to lamivudine monotherapy for a median duration of 25 months (33 vs. 17 %; P = 0.40), although no patients in the sequential therapy group showed emerging resistance to lamivudine. A group in China, where genotype B or C is predominant, reported very similar results [38].

In Japan and other East Asian countries, genotype C is the most prevalent HBV type [39, 40], and most patients with chronic hepatitis B acquire the virus perinatally [13]. Thus, response rates to IFN-based therapy in these countries are lower than those reported in Europe and the USA. In our previous study [41] using sequential therapy with lamivudine alone for 16–32 weeks, followed by the lamivudine + IFN combination for 4 weeks and lastly by IFN alone for 20 weeks, the rate of sustained loss of HBeAg was only 29 %. The rate of HBeAg loss during lamivudine treatment was higher among sustained responders than that among non-responders. In a multicenter trial, Minami et al. [42] found that patients who lost HBeAg during lamivudine treatment were more likely to show a sustained response to sequential therapy. Okuse et al. [43] reported that sequential therapy was effective for patients with acute exacerbations of chronic hepatitis B, particularly those in whom HBeAg had become negative during lamivudine treatment.

To date, a small study by Moucari et al. [44] from France has been the only one to evaluate the efficacy of sequential therapy with adefovir dipivoxil followed by IFN-α. Sustained virologic response, defined as serum HBV DNA of <10,000 copies/mL, was achieved in 50 % of patients, but only 20 HBeAg-negative patients were included in this study.

We recently reported the outcomes of sequential therapy with entecavir followed by IFN-α [45]. Among the 24 patients receiving entecavir alone for 36–52 weeks, followed by the entecavir + IFN-α combination for 4 weeks, and lastly by IFN-α alone for 20 weeks, the rate of sustained response, defined by HBeAg loss, HBV DNA of <10,000 copies/mL, and normal ALT, was 21 %; this was not higher than the rate found in our previous study using lamivudine [41]. In the study carried out in China, Chen et al. [46] included only patients with acute exacerbation (ALT >10-fold the upper limit of normal) who were treated with entecavir alone for 12–26 days before the ALT had declined to five- to ten-fold the upper limit of normal, followed by the entecavir + pegylated IFN-α combination for 2 weeks, and then by pegylated IFN-α alone for 22–46 weeks. Sustained virologic response, defined as HBV DNA of <10,000 copies/mL, was obtained in 69 % of HBeAg-positive and in 80 % of HBeAg-negative patients with acute exacerbation of chronic hepatitis B.

One objective of sequential therapy starting with NA is to lower the viral load before IFN therapy is initiated, thereby restoring treatment sensitivity as low HBV DNA levels are associated with a favorable response to IFN. Another objective of sequential therapy is to prevent the relapse of hepatitis following the discontinuation of NA therapy through the use of IFN. The high risk of viral relapse after treatment may be attributed to the persistence of cccDNA in the liver, which is correlated with HBV antigen levels in the serum. Using HBsAg and hepatitis B core-related antigen (HBcrAg) levels, Matsumoto et al. [47] proposed a model for predicting relapse after the discontinuation of NA therapy. In our study of sequential therapy using entecavir [45], few patients showed a decrease in HBsAg or HBcrAg to the level meeting the criteria of safe discontinuation of NA. Taken together with the fact that at least 21 % of our patients achieved a sustained response, we suggest that the switch to IFN-α contributes to the safe termination of NA therapy in some patients [48].

Combination with NA and IFN in the guidelines

Combination therapy with NA and IFN is not recommended in the guidelines proposed by the Asian–Pacific Association for the Study of the Liver (updated in 2008) [13] and the American Association for the Study of Liver Diseases (updated in 2009) [14] because there has been no large clinical trial to confirm the benefits of combination therapy over monotherapy in inducing a higher rate of sustained response. The most recently updated guidelines proposed by the European Association for the Study of the Liver (updated in July 2012) [15] also does not recommend combination therapy of IFN with lamivudine or telbivudine. However, the limited information currently available on the efficacy and safety of combining IFN with other NAs has raised an unresolved issue of assessing the safety and efficacy of combining IFN with a more potent NA, such as entecavir or tenofovir.

Guidelines proposed by the Japanese Study Group of the Standardization of Treatment of Viral Hepatitis [49] basically recommend IFN as the first-line treatment for chronic hepatitis B patients aged <35 years to attain a “drug-free state” and entecavir for patients aged ≥35 years to persistently suppress HBV DNA (as tenofovir disoproxil fumarate has not been licensed in Japan to date). In patients aged <35 years and harboring HBV DNA in titers of ≥7 log copies/mL, sequential treatment with entecavir followed by IFN is recommended as the first-line therapy if HBeAg is negative and as the second-line therapy (next to IFN monotherapy) if HBeAg is positive. In patients aged ≥35 years and harboring HBV DNA of ≥7 log copies/mL, sequential treatment is recommended as the second-line therapy (next to entecavir) if the HBeAg is positive.

Conclusions

It remains unclear whether combination therapy is superior to monotherapy for treating chronic hepatitis B. Consequently, controlled trials comparing combination and monotherapy are necessary. In previous studies comparing the lamivudine + IFN combination and lamivudine monotherapy in a finite course, combination therapy was associated with higher rates of sustained post-treatment response and lower rates of drug resistance than lamivudine monotherapy. However, NAs are generally administered indefinitely due to high rates of post-treatment relapse. Additionally, even when NAs are administered alone, concern for drug resistance has significantly decreased (≤1.2 % in 3–5 years [50, 51]) with the development of newer high-potency NAs, such as entecavir and tenofovir. In previous studies comparing the lamivudine + IFN combination and IFN monotherapy, combination therapy showed greater on-treatment viral suppression, but no difference in the post-treatment sustained response was observed when compared to therapy with IFN alone. The efficacy of combining IFN with a more potent NA remains to be evaluated. Further studies are needed to determine whether switching to IFN contributes to the safe discontinuation of therapy, particularly in patients with decreased HBsAg and/or HBcrAg levels during long-term NA treatment [52].