Treatment with nucleos(t)ide analogues in chronic hepatitis B: Where does the road map lead us?☆

Despite the introduction of effective vaccination programs in a substantial number of countries, it is anticipated that in the coming decades many hepatitis B virus (HBV)-infected patients will die from liver failure or hepatocellular carcinoma. Effective antiviral therapy may decrease the incidence of this HBV-related mortality, which is currently estimated around one million annually worldwide [1]. Permanent and complete suppression of viral replication is the main treatment goal of antiviral therapy with nucleos(t)ide analogues (NA) for chronic HBV infection [2], [3], [4]. As a sustained off-treatment response seems only possible in a minority of patients, NA have to be administered for very long periods, if not indefinitely [5], [6], [7]. Development of antiviral resistance is a major limitation to long-term efficacy of NA. It leads to reversion of virologic and histological improvement, and enhances the rate of disease progression [8]. In the last few years management of resistance has evolved, and treatment strategies and clinical guidelines have been developed, which are focused on prevention of antiviral drug resistance [4], [9], [10]. First, only potent NA with a high genetic barrier, meaning drugs requiring multiple resistance mutations, should be used as monotherapy. Second, treatment should be adapted at an early stage in case of incomplete viral suppression, as several studies have already shown that an initial virologic response is associated with lower rates of antiviral drug resistance in HBV patients in the long term [11], [12], [13]. The so-called road map concept and the recently published EASL guidelines on the treatment of chronic HBV infection both propose that virologic response should be assessed at week 12 and 24 to identify primary non-response and/or partial virologic response, respectively, and to modify treatment accordingly [4], [9]. Although it is questionable whether this concept also applies to the new potent drugs with low resistance rates, experts suggested treatment modification in patients receiving tenofovir or entecavir with a partial virologic response at week 48 as well [4]. A third option would be to offer de novo combination of NA therapy, although the clinical benefit in HBV-monoinfected patients will be difficult to demonstrate in light of antiviral drugs with excellent resistance profiles.

In this issue of the Journal of Hepatology, Zeuzem et al. [14], report on a post-hoc analysis of the two-year results of the GLOBE trial, to identify the strongest predictor for optimal outcomes at two years of treatment with telbivudine. Key endpoints are HBV DNA less than 300copies/ml, serum ALT normalization, HBeAg seroconversion, and telbivudine resistance at week 104. Resistance was defined as emergence of treatment – associated resistance mutations, identified by direct sequencing in all patients with serum HBV DNA >1.000copies/ml at week 104. Stepwise logistic regression analyses were performed to identify variables associated with treatment outcomes. The most important message of this study is that undetectable HBV DNA at week 24 is the strongest predictor for all outcomes at year 2. Furthermore, as previously known from other studies, low HBV DNA (i.e. <9log10copies/ml) and high ALT levels (i.e. >2× ULN) were associated with favourable outcomes at year 2 in HBeAg-positive patients. In HBeAg-negative patients, baseline HBV DNA and ALT levels were not significant predictors of outcomes at year 2. The study describes an intent-to-treat analysis and therefore has an advantage to the entecavir follow-up studies using a suboptimal design necessitating cumulative confirmed response and the tenofovir studies where a combination of tenofovir and emtricitabine was given to those not responding adequately after 72 weeks [15], [16], [17]. Zeuzem et al. do not address the results of their study on telbivudine within the perspective of the outcome of other recently licensed anti-HBV drugs.

What implications does the outcome of this study on telbivudine have on our clinical practice? The first question is whether week 24 is the optimal time point for treatment evaluation of the initial response to telbivudine, and whether these results can be translated to other NA. The multivariate logistic regression analyses identified undetectable HBV DNA at week 24 as the strongest predictor for all four outcomes at week 104. Yet, one can hardly be surprised by this outcome. Serum HBV DNA is closely related to all four study endpoints, and, more importantly, week 24 is nearer to the time point of outcome (i.e. week 104) than baseline or week 12. It is likely that if undetectable HBV DNA at week 36 or 48 were included in the model, these parameters would appear to be a stronger predictor for all outcomes at week 104 than undetectable HBV DNA at week 24. This would, however, not necessarily indicate that week 36 or 48 would be better time points of evaluation. This is in fact suggested by the current study where it appears that approximately 30% of the HBeAg-positive patients with detectable HBV DNA at week 12 demonstrated resistance to telbivudine at week 104 [14]. Subsequently, most clinicians would probably not wait until week 24 before switching to alternate treatment. With respect to lamivudine which has an inferior resistance profile as compared to telbivudine, a recent report showed that HBV DNA levels measured as early as four weeks were able to predict the ideal response to lamivudine at five years of therapy [18]. So far, no predictors for resistance to either entecavir or tenofovir could be identified due to their extremely low resistance rates [17], [19], which make any of these on-treatment recommendations unapplicable to antiviral agents with a high barrier against resistance. Therefore, only extensive clinical decision analyses will be able to help us to elucidate the optimal treatment strategies using on-treatment serum HBV DNA levels to prevent the emergence of antiviral drug resistance in patients receiving NA-therapy.

The second question is whether telbivudine, which was not given priority as a first-line treatment option in the recent EASL HBV treatment guidelines [4], may still be a suitable treatment option for patients with favourable baseline and on-treatment characteristics. Telbivudine is an l-nucleoside that is highly selective for HBV [20]. Recently, the two-year results of the GLOBE-trial, comparing telbivudine to lamivudine monotherapy in patients with chronic HBV monoinfection, were presented [21]. In HBeAg-positive patients, 56% of telbivudine-treated subjects achieved undetectable HBV DNA, and 30% demonstrated HBeAg-seroconversion. In HBeAg-negative patients, 82% of telbivudine-treated subjects showed HBV DNA levels less than 300copies/ml after two years of treatment. However, the frequency of telbivudine-resistant HBV (rtM204I) in HBeAg-positive patients increased from 5% at year 1 to 25% at 2 years, whereas in HBeAg-negative patients 11% demonstrated genotypic resistance to telbivudine at year 2. The current study shows that in patients with favourable baseline characteristics and undetectable HBV DNA at week 24, emergence of telbivudine-resistant mutations still occurred in approximately 2% of both HBeAg-positive and HBeAg-negative patients [14]. As compared with entecavir-treated subjects, genotypic resistance to entecavir developed in only 1.2% of NA-naı¨ve patients after five years of treatment [17]. Moreover, antiviral drug resistance was not observed in any patient after two years of tenofovir treatment [19]. As previously mentioned, one can comment on the design of the phase III entecavir and tenofovir studies. Nevertheless, the excellent profiles of both agents have now been confirmed in large-scale investigator-initiated studies with variable follow-up, organised by European research consortia [22], [23], [24]. Finally, it is not well known which percentage of patients indeed meet the criteria of the favourable baseline profile (HBV DNA <9log10copies/ml and ALT levels >2× ULN) for response to telbivudine. The authors state that recent survey data indicate that patients with such characteristics predominate in clinical practice, yet only 17% (80/458) of their own HBeAg-positive patients – a population still highly prevalent in eastern Asia and northern Europe – qualified for these criteria.

In conclusion, the current study undoubtedly helps to identify patients who are likely to achieve a long-term maintained response to telbivudine therapy. Nevertheless, one could still question whether telbivudine, due to its inferior resistance profile, should be considered as an optimal treatment for HBV-monoinfected patients, even for those with favourable baseline characteristics and even when virological response will be evaluated at week 24. Furthermore, we need well-powered long-term studies, preferably investigator-driven, to assess the best future strategy with our strongest antiviral agents: long-term monotherapy, add-on therapy with a road map or de novo combination therapy.