HBeAg-negative chronic hepatitis B: why do I treat my patients with pegylated interferon-alfa?

Chronic hepatitis B (CHB) virus infection is a global public-health problem associated with increased morbidity and mortality 1, 2. Approximately 15–20% of patients with CHB develop cirrhosis within 5 years, while 15% of patients with compensated cirrhosis and more than 60% of those with decompensated cirrhosis die within 5 years 2. Moreover, all patients with chronic hepatitis B virus (HBV) infection are at increased risk of hepatocellular carcinoma (HCC), although the risk increases substantially in patients with long-term high viraemia and cirrhosis 2.

HBeAg-negative CHB, which is currently the predominant type of CHB worldwide, represents a late phase in the natural history of chronic HBV infection 2. It may develop immediately after HBeAg seroclearance and seroconversion to anti-HBe or more frequently after a long inactive chronic HBV carrier phase 2. Treatment is often indicated for HBeAg-negative CHB patients because spontaneous remission rarely occurs and these patients have more advanced liver disease compared with HBeAg-positive patients 1, 2. Given the identical serological profile, HBeAg-negative CHB patients should be differentiated from inactive chronic HBV carriers. Differentiation is not always easy, and is based on close follow-up of HBeAg-negative patients with sequential determinations of alanine aminotransferase (ALT), serum HBV DNA levels and possible assessment of histological lesions in the liver by non-invasive methods or liver biopsy 3.

The aim of treatment in chronic HBV infection is HBV eradication expressed by clearance of HBsAg from serum and of covalently closed circular HBV DNA (ccc HBV DNA) from hepatocytes 1, 4. However, since HBV eradication is only rarely achieved with currently available anti-HBV agents 1, 4, a more realistic aim is the normalization of ALT activity, persistent inhibition or at least significant suppression of HBV replication and prevention of cirrhosis and HCC 1, 4-6. Currently, two classes of agents are licensed for the treatment of CHB: interferon-alpha (IFNa, standard or pegylated) with both antiviral and immunomodulatory activities, and nucleos(t)ide analogues (NAs) with only antiviral activity 1, 4-6. Despite the good tolerance and safety and the excellent on-treatment efficacy, NAs are not an optimal therapeutic approach because of the need for long-term, perhaps indefinite, treatment. Although a proportion of patients with HBeAg-positive CHB may stop NA therapy after a few years if they achieve stable anti-HBe seroconversion, the treatment duration is practically unknown for most patients with HBeAg-negative CHB who may stop NA therapy only if they achieve HBsAg clearance 1, 4-6. Thus, the indefinite duration is the major limitation of current NA therapy in HBeAg-negative CHB patients raising safety and family planning issues. Moreover, long-term NA therapy requires careful long-term monitoring for virological response with sequential serum HBV DNA determinations and continuous supervision and encouragement for patient compliance, while it results in a gradually increasing cost of treatment 1, 4-6.

Conversely, the main advantage of IFNa-based therapy is the predetermined finite, usually 48 weeks, treatment duration which offers a chance for sustained virological response (SVR) off-therapy in approximately one fourth of patients with HBeAg-negative CHB 7, 8. Another advantage of IFNa is that off-treatment SVR are usually durable and improve the patient's long-term outcome 9. Despite its advantages, IFNa therapy is not extensively used in CHB because of the parenteral administration, the need for frequent clinical and laboratory monitoring, its side effects profile and especially its limited efficacy. Standard IFNa has now been largely replaced by pegylated (PEG)-IFNa which has an improved pharmacokinetic profile (once instead of thrice weekly injection) with a longer half-life and without wide fluctuations in serum concentrations, ultimately maximizing adherence, reducing side effects and improving suppression of viral replication 1, 4. Therefore, the current scientific guidelines recommend a 48-week course of PEG-IFNa as a first-line option for the treatment of patients with HBeAg-negative CHB without contraindications to IFNa, as this is practically the only therapeutic option that can offer a chance for off-treatment SVR after a finite course of therapy in this setting 4-6.

Efficacy of current PEG-IFNa therapy in HBeAg-negative CHB

Cohort studies in HBeAg-negative CHB, which used insensitive virological assays, showed that 12 or 24 month courses of standard IFNa (3 or 5 million units thrice weekly) achieve sustained long-term off-therapy biochemical and virological responses in 22–30% of patients 8, 10 who often (>40%) clear HBsAg and have improved outcome and survival 8, 9. Subsequently, a 48-week course of PEG-IFNα-2a (180 μg/week) was found to be as effective as the combination of PEG-IFNα and lamivudine (LAM) (100 mg daily) and to induce higher SVR off-therapy than LAM alone (43% vs. 29%, P = 0.007) 7. In particular, PEG-IFNa was reported to induce combined biochemical and virological responses in approximately 36% of patients after 24 weeks of post-treatment follow-up 7. In addition, loss of serum HBsAg occurred in 12 patients in the PEG-IFNα-2a ± LAM groups and none of the patients in the LAM groups. Three years after the end of therapy, the percentage of patients with normal ALT or HBV DNA ≤10 000 copies/ml was significantly higher in those treated with PEG-IFNα than with LAM (31% vs. 18%, P = 0.032 and 28% vs. 15%, P = 0.039 respectively) 12. After a follow-up period of 5 years, 12% of patients treated with PEG-IFNα-2a with or without LAM achieved HBsAg clearance. Importantly, high rates of HBsAg clearance were achieved in patients with HBV DNA levels <10 000 copies/ml, 1 year post-treatment and 5 years post-treatment 12.

Baseline predictors of response to PEG-IFNα

Since PEG-IFNα achieves SVR in a minority of patients with HBeAg-negative CHB, it has been a major challenge to identify the patients who are likely to benefit from treatment based on parameters at baseline or early during treatment.

Several studies have been published investigating possible baseline predictors of response to PEG-IFNα. Multivariate analyses of HBeAg-negative CHB patients treated with PEG-IFNα alone, LAM alone, or both in combination found that high baseline ALT, low baseline HBV DNA, younger age and female gender were independent predictors of a combined ALT and HBV DNA response at 24 weeks post-treatment 13. However, viraemia and liver enzymes tend to fluctuate over time making the prediction of response based on these variables in a given patient rather unreliable. HBV genotype has become a new predictive factor in HBeAg negative CHB, with genotype D associated with a poorer response to treatment than other genotypes.

Recently, genomics have been investigated in CHB based on findings in hepatitis C virus (HCV) therapy showing that interleukin 28B (IL28B) polymorphisms predict the response to PEG-IFNα and ribavirin (RBV). Lampertico et al. 14 retrospectively analysed 101 HBeAg-negative CHB patients (92% genotype D) who were followed up for a median of 11 years (range: 1–17) after a median of 23 months (range: 10–48) of either standard or PEG-IFNa therapy. Patients with IL28B rs12979860 genotype CC were found to have higher end-of-treatment (69% vs. 45%, P = 0.01) and higher SVR rates (31% vs. 13%, P = 0.02) than non-CC patients. Interestingly, the rate of serum HBsAg clearance was 29% in CC compared with 13% in non-CC genotype carriers (P = 0.039). The authors suggested that IL28B polymorphisms might be used to define the pretreatment probability of a virological response in the difficult-to-cure patient population of HBeAg negative carriers with genotype D. However, Brouwer et al. 15 have recently reported contradictory results. They retrospectively analysed 133 HBeAg-negative CHB patients (81% genotype D) treated with PEG-IFNa with or without RBV or LAM for 48 weeks and showed that polymorphisms near the IL28B gene were not associated with on- and post-treatment kinetics of HBV DNA and HBsAg levels, or with 24 week post-treatment responses. Similarly, in a Greek cohort of HBeAg-negative CHB patients the IL2B polymorphisms were recently reported not to be associated with response to PEG-IFNα therapy 16. Therefore, further studies are required to clarify the potential significance of IL28B and perhaps other polymorphisms for the pretreatment prediction of virological responses to PEG-IFNα therapy.

On-treatment predictors of response – stopping rules for PEG-IFNα

Traditionally, HBsAg had been primarily used to diagnose HBV infection. Recently, two commercially available assays for HBsAg quantification, the Architect HBsAg assay (Abbott Diagnostics, Abbott Park, IL, USA) 17 and the Elecsys HBsAg II quant assay (Roche Diagnosticks, Indianapolis, IN, USA) 18 have been developed. It is important to note that the results of both assays are very highly correlated (r ≥ 0.96) and in close agreement across all major HBV genotypes 19. The availability of these easy to use assays has revived the interest in the utility of HBsAg levels promoting numerous clinical research studies 20. Recent data have shown that HBsAg levels reflect the transcriptional activity of the cccDNA and that they can be useful as a surrogate marker to predict the natural course of HBV-infection and treatment outcomes 20. Indeed, the potential use of HBsAg levels in the management of HBeAg-negative CHB patients treated with PEG-IFN is rather attractive given the relatively small proportion of patients who achieve SVR and the high probability of post-treatment relapses. It should be noted, that HBsAg is the only detectable serological marker in HBeAg-negative CHB patients who achieve virological responses during treatment.

Existing data suggest that lower HBsAg levels at baseline may be associated with a better treatment response. Takkenberg et al. 21 have analysed the predictive value of baseline HBsAg levels in 48 patients with HBeAg-negative CHB who received PEG-IFNa and adefovir (ADV) for 48 weeks and were followed for 2 years thereafter. Baseline HBsAg was the only independent predictor of HBsAg loss at week 144 (odds ratio: 0.22, P = 0.01). Specifically, 75% of patients with HBsAg loss had a baseline HBsAg level below 400 IU/ml corresponding to a positive predictive value of 100% and a negative predictive value (NPV) of 95%. However, since HBsAg levels show significant variations during the natural course of chronic HBV infection and perhaps among different HBV genotypes, this variability could have important implications for using quantitative HBsAg as a marker for predicting treatment responses 22, 23. In support of this, a recent study by Brunetto et al. 24 provides new insight into the influence of HBV genotypes on both on-treatment HBsAg kinetics and end of treatment HBsAg levels that are associated with a SVR. Their results show that baseline HBsAg levels were significantly higher for A than B, C or D genotypes. Moreover, on-treatment HBsAg kinetics varied between genotypes. Therefore, the authors have suggested genotype-specific time frames and specific end-of-treatment thresholds to improve response-guided treatment of HBeAg-negative CHB.

A decrease in HBsAg levels during PEG-IFNα treatment could be also useful as a predictor of off-treatment SVR and/or HBsAg clearance in patients with HBeAg negative CHB. In a relatively small study including 48 patients with HBeAg-negative CHB who were treated with PEG-IFNa for 48 weeks, there was a 89 and 92% chance of SVR if HBsAg declined >0.5 log₁₀ at week 12 and >1 log₁₀ IU/ml at week 24 respectively. In contrast, in patients who failed to achieve this decline, the chance of a SVR defined as undetectable (<70 copies/ml) serum HBV DNA at 24 weeks post-treatment was only 10% 25. However, this study, lacked sufficient power for a reliable algorithm and results were not confirmed in another study of PEG-IFNα in HBeAg-negative CHB 26. Brunetto et al. 27 showed that HBsAg level <10 IU/ml at week 48 and an on-treatment decline of more than 1 log₁₀ IU/ml were significantly associated with sustained HBsAg clearance 3 years after treatment (both P < 0.0001). Interestingly, the predictive value of HBsAg levels at week 48 was higher than that of HBV DNA levels.

Rijckborst et al. 26 investigated the role of early on-treatment changes in HBV DNA and HBsAg levels to identify HBeAg-negative CHB patients with a low probability of achieving an off-treatment SVR which can lead to early withdrawal of ineffective courses of PEG-IFNα. In 102 HBeAg-negative CHB patients treated with PEG-IFNα with or without RBV for 12 months in the PARC trial, none of the 20 patients without a decline of HBsAg levels and a <2 log₁₀ copies/ml decrease in HBV DNA achieved a SVR defined as serum HBV DNA <10 000 copies/ml and normal ALT at 6 months after the end of treatment (NPV = 100%). Most patients enrolled in the PARC trial (81/102) had genotype D and this stopping rule was also confirmed in a validation dataset (n = 91) that included HBeAg-negative CHB patients mostly with genotype D who were treated with PEG-IFNα for 48 or 96 weeks 28 The results of the stopping rule were good in two studies (P = 0.001) with a NPV of 100% in genotype D patients allowing treatment discontinuation in 19% and keeping all other patients on treatment with subsequent sustained responses. Based on these findings, this stopping rule (no decline in HBsAg and no reduction in HBV DNA ≥2 log₁₀) has now been included in the recommendations of recent scientific guidelines 4. This stopping rule was also confirmed in a Greek cohort of predominantly genotype D patients with HBeAg-negative CHB who were treated with Peg-IFNα-2a for 48 weeks 29.

The strategy of stopping early based on the combination of HBsAg and HBV DNA levels was further supported by the results of a recently published cost-effectiveness analysis. Using a decision-analytic Markov model, Iannazzo et al. 30 showed that the cost-effectiveness of antiviral therapy in HBeAg-negative CHB patients could be significantly improved using first-line PEG-IFNα followed by a switch to NAs in patients who fulfil the week-12 HBV DNA/HBsAg stopping rule. Therefore, the application of this early stopping rule optimizes the effectiveness of PEG-IFNα therapy in patients with HBeAg-negative CHB (genotype D) by avoiding unnecessary treatment to those who have no chance of achieving a sustained response.

Extended PEG-IFNa administration

Initial studies using conventional IFNα in patients with HBeAg-negative CHB have shown that these patients usually relapse after the end of treatment and might benefit from prolonged administration of IFNα for more than 48 weeks 10, 31. Therefore, a similar approach was used with PEG-IFNα. In a pilot USA study 32, 13 HBeAg negative CHB patients received PEG-IFNα with or without LAM for 60 weeks. Six months after stopping treatment, nine patients (62%) had HBV DNA <20 000 copies/ml with a >2 log₁₀ reduction from baseline and five patients (38%) had undetectable HBV DNA. Although it was a small, uncontrolled pilot study, it offered the basis for studying the administration of PEG-IFNα beyond 48 weeks. Recently, Lampertico et al. completed a multicentre randomized study comparing 96-weeks vs. 48-weeks of PEG-IFNα therapy in 128 HBeAg-negative, predominantly genotype D (94%), CHB patients 33. At the end of treatment, there was no difference in the proportion of patients with a virological response (HBV DNA <2000 IU/ml) between the 96-week (67.3%) and the 48-week (58.8%) groups (P = 0.37). However, at 48 weeks post-treatment, HBV DNA <2000 IU/ml was achieved in a significantly higher proportion of patients in the 96-week group (28.8%) than in the 48-week group (11.8%) (P = 0.03). HBsAg clearance was not achieved in any of the patients in the 48-week group and in three (5.8%) patients in the 96-week group, while two additional patients in the 96-week group had <10 IU/ml HBsAg levels compared with none in the 48-week group.

This was the first time that 96-week PEG-IFNa was found to improve the response in difficult-to-treat genotype D HBeAg-negative CHB patients in a controlled trial setting. Of course, these findings should be confirmed in other and ideally larger studies as well, before the duration of PEG-IFNa therapy can be extended in HBeAg-negative CHB in clinical practice.

Despite the potential for improved efficacy, the possible side effects are of concern for the prolonged use of PEG-IFNα. In the Italian study 33, the authors stated that treatment was well-tolerated without an increase in adverse events or discontinuation rates in patients treated for 96-weeks compared with those treated for 48 weeks. It should be noted, however, that PEG-IFNα was given at the standard 180 μg weekly dose for the first year and a reduced weekly dose of 135 μg in the second year, probably to prevent the side effects of extended PEG-IFNα use. PEG-IFNα seems to be better tolerated in CHB than in chronic hepatitis C patients with lower rates of adverse events and a lower incidence of depression, but an extended course of PEG-IFNα may not be appealing to all patients. Therefore patients should be carefully selected for extended PEG-IFNα therapy to find a balance between the benefits of treatment and patient tolerance if a prolonged course of PEG-IFNα is being considered.

PEG-IFNα and NA(s) combinations – possible future use of PEG-IFNα

There have been different approaches to combine the immunomodulatory properties of PEG-IFN with the direct antiviral activity of NAs to increase treatment response. However, results have not been confirmed in patients with HBeAg-negative CHB so far. In the large registration trial of PEG-IFNα-2a 7 the combination of PEG-IFNα with LAM showed a higher on-treatment virological response rate but no benefit in off-treatment SVR or serological response. In a multicentre randomized controlled study including 60 HBeAg-negative CHB patients, Piccolo et al. 34 reported similar SVR (HBV DNA <2000 IU/ml) 24 weeks after the end of treatment in those treated with a 48-week combination of PEG-IFNα and ADV or PEG-IFNα alone (23% vs. 20%, P = 0.75) with only one patient (3%) in the combination group achieving HBsAg loss. Moreover, in the PARC trial 35 including 138 HBeAg-negative CHB patients the addition of RBV did not improve the efficacy of a 48-week course with PEG-IFNα offering similar rates of combined responses (HBV DNA <10 000 copies/ml and normal ALT) 6 months after the end of treatment (20% vs. 16%, P = 0.49). Based on the above findings, current international guidelines do not recommend the use of PEG-IFNα in combination with NA(s) for the treatment of CHB. It should be noted, however, that there are no data on the efficacy and safety of the combination of PEG-IFNα with the first-line NAs, entecavir and tenofovir. Since large trials evaluating the effects of these combinations in both HBeAg-positive and HBeAg-negative CHB are ongoing and their results should appear in the near future, the role of PEG-IFNa and NA combinations remains open.

A new approach for the treatment of HBeAg-negative CHB patients might be the add-on or switch to PEG-IFNα therapy in NAs responders to accelerate the decline in HBsAg levels and induce HBsAg clearance which will result in safe discontinuation of antiviral treatment. This approach is supported by observations showing that most patients treated with NAs achieve complete maintained inhibition of HBV replication but no substantial decline in HBsAg levels 36. This is probably because of the mechanism of action of NAs which affect reverse transcription of pregenomic RNA, but not the cccDNA and subgenomic RNA which have a translational activity associated with HBsAg production 37.

Certain small studies have evaluated this new strategy and provide promising results. In one study with 20 consecutive HBeAg-negative CHB patients, ADV was given for 20 weeks followed by ADV and PEG-IFNα for 4 weeks and finally PEG-IFNα alone for 44 weeks. Twenty-four weeks after the end of PEG-IFNα, 10 patients (50%) had a partial (HBV DNA <10 000 copies/ml) or complete (HBV DNA <70 copies/ml) virological response 38. In another study 39, 30 patients with HBeAg-negative CHB were randomized to receive PEG-IFNα-2a for 24 weeks followed by telbivudine (TBV) for 24 weeks or vice versa. At the end of follow-up, the SVR (HBV DNA <2000 IU/ml) was significantly higher in patients treated with TBV first followed by PEG-IFNα than vice versa (46.7% vs. 13.3%, P = 0.046). Following a slightly different concept, Kittner et al. 40 added PEG-IFNα to a stable and effective NA therapy and closely monitored the kinetics of HBsAg levels in 12 CHB patients (nine HBeAg-negative). Two patients including one of the nine HBeAg-negative cases had a rapid decline in HBsAg resulting in anti-HBs development at weeks 32 or 40, while PEG-IFNα was discontinued in the remaining patients without a substantial decline in HBsAg levels after 24 weeks of PEG-IFNα. These preliminary results suggest that the combined use of NAs and PEG-IFNα may be effective, although the optimal concept for such an approach is still unclear. Further studies are under way to explore these strategies, to reveal their potential benefits and determine whether they are worth pursuing.

Conclusions

PEG-IFNα therapy, usually given for 48 weeks, is practically the only finite treatment to offer a chance of SVR in patients with HBeAg-negative CHB. Since only one-fourth of these patients achieve a SVR after PEG-IFNα therapy, the recent identification of strong early on-treatment predictors of no response is an important step. In particular in HBeAg-negative CHB patients mostly with genotype D, a lack of decline in HBsAg levels and in HBV DNA levels ≥2 log₁₀ copies/ml at week-12 has nearly 100% NPV for an off-treatment SVR and is now recommended as a stopping rule for early discontinuation of ineffective PEG-IFNα. Prolonging PEG-IFNα therapy to 96 weeks seems to provide higher SVR rates but the application and efficacy of this approach requires further study. The role of newer combined approaches with PEG-IFNα and NA(s) are currently under evaluation.