Why do I treat HBeAg-negative chronic hepatitis B patients with pegylated interferon?


  • Pietro Lampertico,

    Corresponding author
    • “A.M. e A. Migliavacca” Center for the Study of Liver Disease, 1st Division of Gastroenterology, Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, Università degli Studi di Milano, Milano, Italy
    Search for more papers by this author
  • Mauro Viganò,

    1. Hepatology Division, Ospedale San Giuseppe, Università degli Studi di Milano, Milano, Italy
    Search for more papers by this author
  • Massimo Colombo

    1. “A.M. e A. Migliavacca” Center for the Study of Liver Disease, 1st Division of Gastroenterology, Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, Università degli Studi di Milano, Milano, Italy
    Search for more papers by this author


Pietro Lampertico MD, PhD, 1st Division of Gastroenterology, Fondazione IRCCS Ca’ Granda , Ospedale Maggiore Policlinico, Università degli Studi di Milano, Via F. Sforza 35 – 20122 Milano, Italy

Tel: +39 0255035432

Fax: +39 0250320700

e-mail: pietro.lampertico@unimi.it


Chronic hepatitis B (CHB) in serum HBeAg negative patients is a difficult to cure, progressive disease leading to end-stage liver disease and hepatocellular carcinoma. Currently, there are two different treatment strategies for such patients: a finite course of Pegylated interferon (PEG-IFN) or long-term administration of the more potent and less resistance-prone nucleot(s)ide analogues (NUC), i.e. entecavir and tenofovir. Although NUC may ensure persistent viral suppression by preventing disease progression in most patients, they require lifelong administration with the hypothetical disadvantages of cost, lack of long-term safety data and, most important, the null rates of HBsAg seroclearance. On the other hand, 1 year of PEG-IFN has the advantage of providing an immune-mediated control of hepatitis B virus (HBV) infection, with the possibility of achieving a sustained off-treatment response in 20% of the patients, ultimately leading to HBsAg loss in approximately 50% of these. However, these sustained response rates can be significantly increased by carefully selecting candidates for PEG-IFN therapy based upon baseline ALT and HBV DNA levels, viral genotype and IL28B polymorphisms, by extending PEG-IFN therapy beyond 48 weeks and, most importantly, by applying early on-treatment stopping rules based upon HBsAg kinetics. Overall, PEG-IFN is an ideal treatment strategy in selected patients with HBeAg-negative CHB, because of its well-recognized and predictable safety profile and a unique mechanism of antiviral activity leading to long-lasting immune control. Because of these features, new therapeutic trials based upon a combination of PEG-IFN and third generation NUC such as entecavir and tenofovir, in both naïve and NUC-exposed patients, are ongoing to further increase the rates of HBsAg seroclearance, which remains the ‘ideal end-point’ in all HBeAg-negative CHB subjects.




adverse effects


chronic hepatitis C


end of treatment




negative predictive value


nucleot(s)ide analogues




sustained viral response

The goal of antiviral therapy in chronic hepatitis B (CHB) patients is to improve quality of life and survival by preventing the progression of liver damage to cirrhosis, end-stage liver disease, hepatocellular carcinoma (HCC) and death [1-3]. This goal can be achieved if HBV replication can be suppressed in a sustained or maintained manner either by short-term ‘curative’ treatment with interferon (IFN)-based treatment or long-term ‘suppressive’ therapy with potent and high barrier to resistance Nucleot(s)ide analogues (NUC), i.e. entecavir and tenofovir.

Unlike NUC, IFN has antiviral and immunomodulatory properties that result in direct inhibition of viral replication and the enhancement of the host's antiviral immune response. This synergic effect may result in sustained disease remission in one-fourth of patients and subsequent HBsAg seroclearance, i.e. the ideal end-point of anti-HBV treatment, in a significant proportion of them. Standard IFN has been largely replaced by Pegylated interferon (PEG-IFN) which has a better administration schedule (once weekly vs 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 [4, 5].

If maintained over time, a sustained viral response (SVR) to PEG-IFN treatment, i.e. HBV DNA levels less than 2000 IU/ml and normalization of serum alanine aminotransferase (ALT) levels, reduces the progression of fibrosis, thus preventing the development of the typical complications of end-stage liver disease and HCC and, ultimately, increasing survival [6-13]. However, the major limitations to its extensive use are the need for parenteral therapy and for clinical and laboratory monitoring, its side effects profile and the limited effectiveness in a large proportion of patients. According to the recently updated EASL 2012 HBV guidelines, a 48-week course of PEG-IFN is recommended in HBeAg-negative patients without contraindications, because it is practically the only ‘curative treatment strategy’ to offer a chance for sustained off-treatment response after a finite course of therapy compared with NUC which require long-term administration [1].

Registration trials

A large multinational registration trial compared the virological and biochemical responses of a 48-week course of treatment with PEG-IFNα-2a (180 μg/week) ± lamivudine (LMV) with LMV monotherapy in 537 HBeAg-negative patients [14]. After 24 weeks of follow-up, the percentage of patients with ALT normalization or HBV DNA levels <20 000 copies/ml was significantly higher in those treated with PEG-IFNα-2a ± LMV (60 and 44%; 59 and 43%, respectively) than with LMV alone (44 and 29%, respectively). The combination of PEG-IFNα-2a and LMV led to more profound end of treatment (EOT) virological response than PEG-IFNα-2a alone (92% vs 81%), yet the therapeutic effect was lost 6 months after treatment cessation when the rates fell to 43 and 44%, respectively. Loss of serum HBsAg occurred in 12 patients in the PEG-IFNα-2a ± LMV groups compared with none of the patients in the LMV group. Adverse effects (AE) were less frequent in patients treated with LMV alone than in those treated with PEG-IFNα-2a ± LMV therapy. In another, long-term follow-up study of 315 patients [15], the 3-year percentage of patients with normal ALT or with HBV DNA ≤ 10 000 copies/ml was higher in patients treated with PEG-IFN than with LMV (31% vs 18%; = 0.032 and 28% vs 15% P = 0.039, respectively). At year 5 of follow-up, 12% of 230 patients treated with PEG-IFNα-2a ± LMV achieved HBsAg seroclearance, compared with 28% (20/72) in patients with HBV DNA ≤ 10 000 copies/ml 1 year post-treatment (Fig. 1) [16].

Figure 1.

Rate of sustained viral response (SVR) (HBV DNA < 10 000 IU/ml) at 1 year post-treatment and HBsAg clearance at 5 years post-treatment in patients with HBV DNA ≤ 10 000 copies/ml 1 year after end of treatment.

How to improve the efficacy of PEG-IFN

Baseline predictors of response

Well-known predictors of a SVR include low baseline HBV DNA, high ALT levels, younger age, female gender and HBV genotype. Unfortunately, the fact that viremia and liver enzymes tend to fluctuate over time makes the prediction of response by these variables in a given patient rather unreliable [17]. In the subanalysis of the Marcellin et al. study in the PEG-IFN monotherapy arm, patients with genotypes B or C had a better chance of response than genotype D patients (P < 0.001), while the latter responded better to the combination with LMV than to PEG-IFN monotherapy (P = 0.015) [17].

More recently, genomics have been investigated in CHB motivated by research in the field of HCV showing that IL28B polymorphisms predict the response to PEG-IFN and Ribavirin (RBV). HBeAg-negative CHB carriers of the favourable IL28B genotype were shown to have higher on-treatment and off-treatment virological and serological responses, thus opening the door to personalized medicine. The only existing data in HBeAg-negative patients are in 101 subjects treated with either conventional IFN or PEG-IFNα-2a for 24 months and followed up for 11 years after treatment [18]. Patients with IL28B rs12979860 genotype CC were shown to have higher EOT (69% vs 45%, P = 0.01) and higher SVR (31% vs 13%, P = 0.02) than non-CC patients. Interestingly, CC patients had a higher cumulative probability of clearing HBsAg during an observation period of 16 years (38% vs 12%, P = 0.039), a finding which was confirmed by multivariate analysis. The IL28B genotype was shown to be an independent predictor of both virological and serological responses together with low baseline HBV DNA levels, high ALT levels and the duration of IFN therapy. To further identify any possible association between IL28B polymorphisms and HBV genotype, data in 93 genotype D patients were reanalysed with IL28B stratification and showed: 48% with CC, 41% with CT and 11% with TT genotypes. The rates of EOT response, SVR and HBsAg clearance were still significantly higher in CC than in non-CC carriers, i.e. 69% vs 44% (P = 0.014), 31% vs 12% (P = 0.028) and 29% vs 12% (P = 0.048). When the assessment of the IL28B polymorphism was extended to include rs8099917, which was recently shown to improve the prediction of a response to PEG-IFN/RBV therapy in chronic hepatitis C patients with the CT genotype of rs12979860, the favourable rs8099917 TT genotype was found in 100% of the rs12979860 CC patients compared with 31% of CT and 10% of TT patients, only. The 42 rs12979860 CT patients with the rs8099917 TT genotype had a significantly higher rate of SVR and HBsAg seroclearance (23% vs 3%, P = 0.045 and 23% vs 0%, P = 0.007 respectively) [19], suggesting that multiple IL28B polymorphisms may be required to define the pretreatment probability of a virological response at an individual level.

PEG-IFN+NUC combination

The assumption that the direct antiviral activity of NUC may add-on to the immunomodulatory properties of PEG-IFN has not been confirmed in HBeAg-negative CHB patients. An Italian multicentre study in 60 HBeAg-negative patients has shown a similar SVR, i.e. HBV DNA <2000 IU/ml 24 weeks after the EOT in those treated with a 48-week combination of PEG-IFNα-2a + adefovir (ADV) or PEG-IFNα-2a alone (23% vs 20%, P = 0.75) with only one patient (3%) in the combination group achieving HBsAg loss [20]. A pilot study with 20 consecutive HBeAg-negative patients treated with ADV for 20 weeks, then ADV+PEG-IFNα-2a for 4 weeks and finally PEG-IFNα-2a for 44 weeks showed that 24 weeks after EOT 10 patients (50%) had serum HBV DNA <10 000 copies/ml [21]. The only study comparing 48 weeks of PEG-IFNα-2a+RBV vs PEG-IFNα-2a alone showed that the addition of RBV did not improve the response (PARC trial), because the combined responses (HBV DNA <10 000 copies/ml and normal ALT) 6 months after the EOT were similar between the groups (20% vs 16%, P = 0.49) [22]. Although international guidelines do not recommend the use of PEG-IFN in combination with NUC, several studies are ongoing to determine whether this combination is worth pursuing.

Extended duration of PEG-IFN administration

Based on the experience showing that standard IFN must be administered for more than 48 weeks to increase SVR rates by reducing the risk of relapse [23], a similar approach was tested with PEG-IFN. In a pilot study in the US, seven HBeAg-negative patients received PEG-IFNα-2a (180 μg/week) for 60 weeks and six patients received PEG-IFNα-2a (180 μg/week) for 12 weeks followed by 48 weeks of PEG-IFNα-2a+LMV. Overall, the SVR defined a reduction in serum HBV DNA of ≥2 log10 copies/ml and HBV DNA <20 000 copies/ml at 24-week of follow-up, was achieved in nine patients. This study concluded that 60 weeks of PEG-IFNα-2a±LMV resulted in a higher response rate than historical controls treated with 48 weeks of PEG-IFN [24].

Recently, in an Italian multicentre study (PegBeLiver), 128 HBeAg-negative patients (mean age 45 years, 94% genotype D, 13% with cirrhosis) were randomized to weekly PEG-IFNα-2a 180 μg for 48 weeks (group A, n = 51), or PEG-IFNα-2a 180 μg for 48 weeks followed by PEG-IFNα-2a 135 μg weekly for additional 48 weeks (group B, n = 52) [25]. At the EOT, virological response (HBV DNA <2000 IU/ml) was similar among the groups (59, 67%) whereas the rates of SVR observed 1 year post-treatment were significantly higher in patients treated for 96 weeks (29% in group B, 12% in A, P = 0.03) (Fig. 2). Three patients in group B, compared with none in group A became HBsAg negative (6% vs 0%) and two additional patients in group B had <10 IU/ml HBsAg levels compared with none in group A. Extended treatment was well tolerated without an increase AE or discontinuation rates compared with patients treated for 48 weeks (Table 1). Another advantage of IFN therapy for CHB is that PEG-IFN is better tolerated in CHB than in hepatitis C patients, with a lower incidence of IFN-related adverse events and a significantly lower incidence of depression [26].

Figure 2.

Virological response rates (HBV DNA <2000 IU/ml) at end of treatment (EOT) and 12 months after EOT with PEG-IFN for 48 weeks (group A) and, PEG-IFN for 96 weeks (group B) in patients with HBeAg-negative chronic hepatitis B.

Table 1. Numbers of patients reporting adverse events (AEs) and laboratory abnormalities in the PegBeliver study
Outcomes, n (%)48-week PEG-IFN (Group A, N = 51)96-week PEG-IFN (Group B, N = 52)P-valueb (A vs B)
  1. a

    Metastatic hepatocellular carcinoma during follow-up.

  2. b

    Two-sided Fisher's exact test, 48-week vs 96-week PEG-IFN.

  3. ALT, alanine aminotransferase.

One or more treatment-related AE42 (82.4)41 (78.8)0.8041
One or more SAE8 (15.7)3 (5.8)0.1222
Deaths 1a00.4951
Laboratory abnormalities
ALT increased3 (5.9)8 (15.4)0.2008
Neutropaenia12 (23.5)9 (17.3)0.4719
Thrombocytopaenia6 (11.8)6 (11.5)1.000
Anaemia5 (9.8)4 (7.7)0.7412
Dose modification15 (29.4)7 (13.5)0.0573
Overall10 (19.6)12 (23.1)0.8107
As a result of AEs8 (15.7)6 (11.5)0.2043

On-treatment predictors of response and stopping rules

Evidence is accumulating to show that a decrease in serum HBsAg levels is a predictor of off-treatment SVR and/or HBsAg seroclearance [27-29]. In a study of 48 HBeAg-negative patients receiving PEG-IFNα-2a, Moucari et al. has reported for the first time that early on-therapy virological and serological (HBsAg) responses could predict a SVR. A decline of 0.5 and 1 log IU/ml of serum HBsAg levels at weeks 12 and 24 of therapy, respectively, had a 90% negative predictive value (NPV) and a 89% positive predictive value (PPV) for week 12 and a 97% NPV and a 92% PPV for week 24 for the detection of patients who would achieve a SVR [27]. After a retrospective analysis of serum HBsAg levels in the PEG-IFNα-2a registration study, patients who achieved a decline of ≥10% in serum HBsAg from baseline at 12 weeks of treatment had a higher probability of SVR than those with a < 10% decline (47% vs 16%, P < 0.01) [28]. The pattern of HBsAg decline in genotype D patients, who were underrepresented in this study, was investigated in Italy for the relationship between on-treatment HBsAg decline ≥10% and a SVR to PEG-IFN [29]. A ≥ 10% decline of HBsAg at week 24, but not at week 12 of treatment, was significantly associated with a SVR to PEG-IFN administered for 96 weeks only [29].

The absolute HBsAg level during treatment may also predict the response to IFN. One-third of the HBeAg-negative patients with HBsAg levels ≤1500 IU/ml at week 12 of PEG-IFNα-2a±LMV, cleared HBsAg 4 years after treatment compared with 4% of those with HBsAg > 1500 IU/ml (P < 0.001) [30]. Brunetto et al. showed that either an EOT serum HBsAg level <10 IU/ml or an on-treatment reduction >2 log10 from baseline at week 48 were significantly associated with HBsAg loss in the next 3 years [31].

On-treatment HBsAg levels may also help identify patients with a low probability of achieving SVR, leading to PEG-IFN withdrawal before the EOT is achieved. A combination of a lack of decrease in HBsAg and < 2 log10 IU/ml reduction in HBV DNA seems to be the best predictor of non-response in genotype D HBeAg-negative patients treated with PEG-IFNα-2a [32, 33]. In 102 HBeAg-negative patients treated with PEG-IFNα-2a±RBV for 12 months, none of the 20 patients with unmodified HBsAg levels and a < 2 log10 copies/ml decrease in HBV DNA had a long-term response defined as serum HBV DNA <10 000 copies/ml and normal ALT 6 months after the end of treatment (NPV = 100%) [32].

This stopping rule has been recently confirmed in genotype D patients enrolled in the original PARC trial (n = 81) and in a validation dataset (n = 91) that also included patients treated with PEG-IFN for more than 48 weeks [33] (Fig. 3). This stopping rule which was effective in both studies (P = 0.001), had a negative predictive value (NPV) of 100% in genotype D patients so that therapy could be discontinued in 19% of patients while all patients with a SVR were kept on treatment. Also none of the seven (21%) patients with unmodified HBsAg levels and < 2 log10 HBV DNA decline at week 12 treated for 96 weeks achieved a SVR.

Figure 3.

Algorithm showing the chances of sustained viral response in the PARC trial (panel A) and the validation trials (registration and PegBeLiver studies for panel B), based on the presence of any HBsAg decline and HBV DNA decline ≥2 log copies/ml at week 12 compared to baseline for patients infected with HBV genotype D.

These studies provide the rationale for stopping PEG-IFN at week 12 in approximately 20% of HBeAg-negative patients with a < 2log10 HBV DNA decline and no change in HBsAg levels, as these patients have no chance of achieving a SVR even if therapy is extended beyond week 48.

PEG-IFN in NUC responders

Because of the slower kinetics of HBsAg loss during NUC therapy, requiring decades to lose HBsAg compared with the faster decline in HBsAg in patients receiving PEG-IFN monotherapy, an alternative strategy to PEG-IFN therapy in HBeAg-negative patients is add-on PEG-IFN in NUC responders to accelerate the decline in HBsAg-induced seroclearance and discontinue antiviral treatment. Two studies have tested this new strategy.

One study reported HBsAg kinetics in 12 patients (nine HBeAg-negative) on stable oral therapy with undetectable HBV DNA who received add-on PEG-IFN as individualized therapy. Only one (11%) HBeAg-negative patient had a rapid decline in HBsAg with a slope that became evident at week 8 and a reduction in HBsAg of 4 log10 at week 48, with the appearance of anti-HBs at week 40. However, this patient had very low baseline HBsAg levels (16 IU/ml), and low levels of anti-HBs even after cessation of PEG-IFN. In the remaining patients, HBsAg only declined by a mean 0.09 log10 after 8–24 (mean 16) weeks and therefore PEG-IFN was stopped [34].

In the second study, 12 HBeAg-negative patients with undetectable serum HBV DNA and stable HBsAg levels received 6–12 months of add-on PEG-IFN. Following an initial increase in HBsAg levels from 841 IU/ml to 1115 IU at month 1 and 927 IU at month 3, HBsAg later declined to 740, 634 and 509 IU at months 4, 5 and 6. Overall, HBsAg declined by approximately 50% within 6 months and two patients (20%) achieved HBsAg levels below 100 IU (78 and 13 IU). During another 3 months, four patients had a further decline in HBsAg levels, from 951 to 221 IU/ml [35]. Despite preliminary encouraging results, the add-on strategy with PEG-INF in NUC responders cannot yet enter clinical practice, but should be evaluated in large specifically designed clinical trials.


Despite the improvement in possible antiviral treatments, management of HBeAg-negative patients remains complex because of the choice between the two different possible strategies: PEG-IFN vs NUC. To date, a course of PEG-IFN may be the most appropriate first-line treatment strategy when the purpose of treatment is to achieve a SVR after a defined treatment course. Despite good tolerance and the higher rates of off-therapy response compared with NUC, the benefits of PEG-IFN are limited to a subgroup of patients. To increase the rates of patients who can benefit from this treatment and minimize the risk of adverse events, treatment should be optimized by careful patient selection and individualized treatment decisions. In this decision-making process, all the relevant factors for each individual patient should be taken into consideration to evaluate the potential risks and benefits. High ALT levels, low HBV DNA, virus genotype and host genetic factors have been identified as pretreatment predictors of a response. However, the application of these parameters is limited in clinical practice as ALT and HBV DNA levels are time-dependent while HBV genotype or IL28B SNPs require specific tests that are not available in every day clinical practice in non-referral centres. To overcome this limitation a response-guided approach to therapy based on HBsAg levels during the first 3 months of treatment is recommended for HBeAg-negative patients. Thus, early identification (week 12 of therapy) of primary non-responders who can stop PEG-IFN, or to extend treatment for more than 48 weeks in those who are likely to respond to reduce relapse rates may be a solution.


The authors have no disclosure.