Back to Conference Reports
Jay H. Hoofnagle
Division of Digestive Diseases and Nutrition
National Institute of Diabetes and Digestive and Kidney Diseases
National Institutes of Health, Bethesda, USA
Several antivirals have shown benefit as therapy of hepatitis B in large clinical trials. More difficult is to translate these results into recommendations for therapy. One difficulty has been lack of clear definitions and firm endpoints in chronic hepatitis B and therapeutic trials of therapy. Chronic infection with hepatitis B virus (HBV) has several forms and outcomes: typical hepatitis B e antigen (HBeAg)-positive chronic hepatitis B, variant HBeAg-negative chronic hepatitis B, the inactive hepatitis B surface antigen (HBsAg) carrier state, and virological resolution or recovery. In most trials of therapy of hepatitis B, the endpoint of successful treatment has been the loss of HBeAg and development of antibody (anti-HBe). This endpoint has several shortcomings, no the least of which is that it is a surrogate endpoint that does not guarantee long-term, durable improvement in the underlying liver disease. A more appropriate approach to evaluation of therapy and design of trials would be to use separate definitions for reponses as either biochemical (fall of aminotransferase levels to normal), virological (loss of HBeAg and/or decrease of HBV DNA), histological (improvement in necroinflammation with no worsening of fibrosis) response or combined (all three). Furthermore, responses must be categorized as either initial (within 6 months), end-of- treatment (while still on therapy), sustained (at 6 or 12 months after therapy), and maintained (at the time of last contact in a patient treated long-term). The optimal endpoint for therapy would be a combined biochemical, virological and histological response that is maintained or is sustained and durable. Ultimately, loss of HBsAg should be the goal of therapy of chronic hepatitis B. Unfortunately, none of these endpoints are reached within a short period of time and all require long-term follow up in this chronic liver disease.
Twenty five years after the first description of interferon therapy of chronic hepatitis B (1) the optimal treatment of this disease remains unsettled and controversial. Currently, two antiviral therapies are approved for use in chronic hepatitis B in the United States, and a third will probably be approved within the next few months. Nevetheless, there remains no clear guidelines for use of antiviral therapy, and investigators differ markedly in recommendations and approaches (2). The task of developing a consensus of therapy of this disease is challenging, as there is little firm medical evidence to support specific recommendations. Yet, for all the difficulty, it is clear that medical control of this chronic liver disease is potentially within reach. New and potent antivirals, innovative and reliable serological assays, and accumulation of information on the natural history and immunopathogenesis of hepatitis B are beginning to allow for important inroads into management of this disease.
This summary will discuss the clinical challenges of clinical trials of therapy of hepatitis B and the best approaches to addressing these challenges. The essential issues are: diagnostic criteria and definitions, optimal endpoints for treatment, use of serological and virological markers, and indications for therapy. These are interrelated issues. Thus, rigorous diagnostic criteria and definitions will not only help with defining indications for treatment but also with defining endpoints. In addition, serological and virological markers are essential in definitions and diagnostic criteria and will ultimately define endpoints for treatment.
Back to top
Diagnostic Criteria and Definitions
Chronic Hepatitis B: Chronic hepatitis B is usually defined by the presence of hepatitis B surface antigen (HBsAg) in serum and histological evidence of chronic necroinflammatory disease on liver biopsy or persistently elevated serum alanine aminotransferase (ALT) activities. Furthermore, chronic hepatitis B is separated into two forms based upon hepatitis B e antigen (HBeAg) and antibody (anti-HBe) status: either HBeAg-positive or HBeAg-negative (sometimes referred to as anti-HBe positive) (2). HBeAg-positive chronic hepatitis B is considered “typical” and is characterized by stable, high levels of circulating hepatitis B virus (HBV) DNA in serum (generally between 107 and 1010 viral copies per mL). HBeAg-negative chronic hepatitis B is attributable to a variant strain of HBV and is characterized by more modest and often fluctuating levels of HBV DNA (generally between 104 and 108 copies per mL). Most patients with HBeAg-negative chronic hepatitis B harbor a variant HBV with mutations in the pre-core region or the basic core promoter of the HBV genome. There are genotype and geographic variations in these two forms of chronic hepatitis B. HBeAg-positive chronic hepatitis B is most typical of North American and European patients who are infected with HBV genotype A. The HBeAg-negative forms of chronic hepatitis B are found more commonly in Southern Europe, the Middle East and Asia in patients infected with HBV genotypes B, C and D.
Inactive HBsAg Carrier State: An important distinction is the separation of chronic hepatitis B from the inactive chronic HBsAg carrier state, which is sometimes referred to as the “healthy carrier state” or “asymptomatic carrier state”, terms that are not particularly accurate. Persons with the inactive HBsAg carrier state have HBsAg in serum, but normal ALT levels and little or no necroinflammatory activity on liver biopsy. These patients have either no detectable or only low levels of HBV DNA in serum (generally less than 104 and rarely as high as 105 copies per mL) (3,4). A large proportion of persons with HBsAg in serum are inactive carriers and appear to have a favorable prognosis with little risk of developing progressive liver injury or cirrhosis.
Resolved Hepatitis B: Finally, full recovery from hepatitis B is defined by clearance of HBsAg and development of antibody (anti-HBs). Actually, HBV often persists in low levels in persons who appear to recover from the infection. HBV DNA can usually be detected in liver tissue and in rare instances in serum as well (5). Furthermore, patients who have recovered can suffer a reactivation of hepatitis B in situations of marked immunosuppression, such as after bone marrow transplantation or with cancer chemotherapy (2). While most persons with resolved hepatitis B have no further liver injury or complications of disease, the infection is latent and not completely eradicated or cured. The mechanisms that allow for latency and that underlie the inactive HBsAg carrier state are unknown and represent a major challenge to our understanding of this disease.
Thus, provisional diagnostic criteria and definitions for chronic hepatitis B can be made (Table 1). Unfortunately, the complexity of hepatitis B does not allow for simple categorization of patients. Some patients with HBsAg and raised serum aminotransferase levels do not have chronic hepatitis B, but are inactive HBsAg carriers and their liver disease is due to another agent or condition, such as hepatitis C or D, hemochromatosis, or nonalcoholic steatohepatitis. Furthermore, patients with HBeAg-negative chronic hepatitis B may have normal serum ALT levels and low levels of HBV DNA for prolonged periods, followed by episodes of marked elevations in virus and ALT levels (a relapsing course occurs in approximately 25% of patients) (2). Finally, patients can spontaneously change from one form of hepatitis B to another: either seroconversion in a patient with HBeAg-positive chronic hepatitis B or sero-reversion (reactivation) in an HBeAg-negative patient. Thus, the definitions shown in the table rely upon repeat testing over at least a six month period.
Back to top
Endpoints of Antiviral Therapy
Loss of HBeAg as an Endpoint for Success of Therapy: Definitions of the various forms of chronic hepatitis B provide the framework for definitions of endpoints of antiviral therapy. For the majority of studies in chronic hepatitis B, the endpoint of successful treatment has been the loss of HBeAg and decrease of HBV DNA to below the level of detection by hybridization assay (~ 105 copies per mL). This endpoint was chosen because studies of the natural history of typical chronic hepatitis B had shown that loss of HBeAg was typically followed by seroconversion to anti-HBe, a marked decrease in HBV DNA levels, fall of serum aminotransferase levels into the normal range and a sustained remission in disease (6). While patients usually remained HBsAg-positive, the accompanying liver disease improved and thus a transition was made from chronic hepatitis B to the inactive carrier state. Thus, when alpha interferon was found to induce loss of HBeAg in ~33% of patients treated, this endpoint was taken as evidence of benefit (7). Long-term follow up of patients treated with alpha interferon showed that the loss of HBeAg and remission in disease was usually durable and, indeed, many patients went on to lose HBsAg and develop anti-HBs (8-10).
Shortcomings of Using HBeAg as an Endpoint: Loss of HBeAg as an endpoint of successful therapy, however, has many limitations, which have become more apparent with development of newer agents for this disease. Perhaps the major shortcoming is that this endpoint is only applicable to patients who have typical, HBeAg-positive chronic hepatitis B. In patients with HBeAg-negative hepatitis B, who often have severe and progressive disease, this endpoint cannot be used. This form of chronic hepatitis B appears to account for an increasing number of cases, not only in Asia and Southern Europe, but more recently in Northern Europe and North America (2). For these cases, suppression of HBV DNA below the level that typically occurs after loss of HBeAg (below 105 copies per mL) has been used as an alternative endpoint (2,11). This endpoint, however, was weakened by the fluctuating course of HBV DNA levels that typically occur in the natural history of chronic hepatitis B and has been problematic until recently because of the lack of standardized, sensitive and accurate measurements of HBV DNA in serum.
Another important shortcoming of the use of HBeAg in defining the endpoint for treatment is that loss of HBeAg may not be have the same significance in patients with different genotypes. The early studies of interferon a were carried out largely among patients with HBV genotype A (8,9), which rarely mutates into a pre-core variant (because of the molecular sequence of this genotype) and thus rarely evolves into HBeAg-negative chronic hepatitis B. In contrast, patients with genotypes B, C and D are susceptible to evolving into HBeAg negative chronic hepatitis B. For these reasons, loss of HBeAg may not be a reliable marker for long-term improvement in disease in patients with these genotypes. These differences may also explain the variability in responses to interferon a in different geographical areas (2).
Problems with the use of HBeAg as an endpoint have also become evident with the development of newer nucleoside analogues with potent activity against HBV. While a one-year course of lamivudine or adefovir dipivoxil lead to marked reductions in HBV DNA and improvements in serum aminotransferase activities and liver histology in the majority of patients, only a small proportion become HBeAg-negative with treatment (12-16). Furthermore, the loss of HBeAg did not appear to be as durable after lamivudine as after a interferon therapy (17). Thus, in recent studies, loss of HBeAg has not had the same long-term clinical implications that were identified in initial studies of the natural history of chronic hepatitis B and long-term follow up studies after interferon therapy.
Timing of Assessment of Loss of HBeAg: A further difficulty that has arisen in trials of newer antiviral agents is that responses were defined during rather than after therapy. In studies of a interferon, the endpoint of therapy was defined as absence of HBeAg six months or more after discontinuation of therapy (7). In contrast, in studies of nucleoside or nucleotide analogues, important endpoints (loss of HBeAg, improvements in ALT, improvements in histology) were measured at the end of treatment (12-16). This difference has major implications, because it is not clear whether the responses are sustained once therapy is stopped. Indeed, discontinuation of lamivudine is almost always followed by a rise in HBV DNA levels and relapse in disease in patients who remain HBeAg-positive on therapy (~80%) (12). Even in patients who become HBeAg-negative, this loss was not always durable once therapy was stopped, relapse being most frequent within the next six months (17). Unfortunately, follow up in many of these studies has been informal and the durability of improvements in HBV DNA, HBeAg, serum aminotransferase levels and histology has not been well defined.
Back to top
Continuous, Maintenance Therapy: Because lamivudine and adefovir dipivoxil have minimal side effects and result in marked suppression of HBV DNA levels with accompanying improvements in serum aminotransferase levels and liver histology, these agents have been increasingly used as continuous, maintenance therapy. The difficulty with prolonged therapy with lamivudine has been the development of resistance in 20 to 25% of patients each year (18,19). Viral resistance is associated with rises in serum HBV DNA levels in association with mutations in the HBV polymerase gene and often with increases in serum ALT and worsening of liver histology. To date, viral resistance to adefovir dipivoxil has not been reported (20). Nevertheless, the risks and benefits of prolonged, continuous therapy with lamivudine and adefovir have not been well defined.
Definition of Endpoints: For these reasons, measured endpoints of therapy need to be modified in trials of therapy with nucleoside analogues and deserve modification in future trials of interferon as well. A reasonable approach to definition of endpoints was developed in the analyses of trials of antiviral therapy for hepatitis C (21). Thus, responses can be defined as biochemical (normal ALT levels), virological (disappearance of virus or fall to a low level), and histological (improvements in histological scoring of necroinflammation and fibrosis) (Table 2). In addition responses are defined as either initial (within six months of starting therapy), end-of-treatment (on therapy), or sustained (measured 6 or 12 months after stopping therapy). In hepatitis B, where long-term continuous therapy is also considered, a fourth category is needed: a maintained response (measured at the last point of follow up). These definitions will allow for a more reasonable follow up of trials of lamivudine and adefovir dipivoxil and future studies of newer nucleoside analogues and combination therapies.
Biochemical Response: A biochemical response is defined by fall of serum aminotransferase levels into the normal range. A difficulty arises when patients with normal ALT levels are treated (13). These patients obviously cannot be evaluated for a biochemical response. In addition, data from these patients should not be used in statistical analyses of changes in ALT levels. This limitation is particularly acute when median rather than geometric mean levels of ALT are shown in descriptions of results (12-14). An occasional problem in definition of a biochemical response is when ALT levels fall markedly but do not become completely normal, or when ALT levels become normal but asparate aminotransferase levels remain elevated (as can occur in patients with cirrhosis). Exact definitions of a biochemical response need to accommodate these special situations.
Virological Response: The usual definition of a virological response based upon HBeAg has many limitations as discussed above. Unfortunately, use of HBV DNA levels to define a virological response is also problematic. Current asssays for HBV DNA can measure levels to as low as 100 to 500 copies per ml. Thus, one definition of a virological response is HBV DNA levels below detection even by sensitive polymerase chain reaction methods. However, in most studies, serum aminotransferase levels and liver histology improves once HBV DNA levels are maintained below 104 to 105copies per mL, so that suppression to a lower level may not be needed for long-term clinical improvement. Furthermore, inactive HBsAg carriers may harbor HBV DNA levels up to as high as 105 copies per mL, but have inactive and non-progressive disease (3,4). For the present time, both definitions could be used.
Histological Response: Improvements in hepatic histology with therapy are usually evaluated using a scoring system, such as the histology activity index (HAI) or the metavir score (22). These systems score necroinflammatory disease activity separately from fibrosis. In most trials, histological improvement is defined by decrease in HAI score by 2 points or more (13-16). Unfortunately, this definition has not been rigorously assessed and correlated with long-term clinical improvement. Furthermore, sampling error on liver biopsy can easily account for improvements or worsening of HAI scores by 2 points or more. For statistical analyses of large numbers of patients, a two point or more improvement in HAI score may be adequate, but in an individual patient, improvements in HAI scores should be far greater or by a proportion of the initial score, such as a decrease of 50% compared to baseline. Worsening of fibrosis is sometimes not included in the definition of improvement and ultimately progression of fibrosis is the most important intermediate endpoint in disease progression. Any worsening of fibrosis scores should negate histological improvement in necroinflammatory scores. Because these systems rely upon subjective readings of liver biopsies and the difficulties of sampling error, the accuracy of the scoring is not great. Clearly, further studies are needed to define a reasonable algorithm for defining histological improvements using the various scoring methods.
Back to top
Combined Response: Improvements in ALT levels without decreases in HBV DNA levels or improvement in liver histology must be interpreted cautiously. Similarly, improvements in histology with no change in ALT or HBV DNA levels are more likely to reflect sampling error than improvement (or recent change in biochemical and virological markers that are not yet reflected in the liver histology). For reasons such as these, definition of beneficial outcome of therapy might best employ combined results from aminotransferase results, HBV DNA levels and histological findings. Until there are better virological markers for a beneficial response to antiviral therapy, use of a combined response is most appropriate. In chronic hepatitis C, a combined response was used until it became clear that a sustained loss of hepatitis C virus (HCV) RNA was a reliable surrogate marker for long-term remission in disease (21).
Complete Response: The loss of HBsAg with development of anti-HBs is a reliable marker for resolution of chronic hepatitis B and might be used as a definition of a complete response to therapy. In studies of a interferon therapy, an average of 8% of patients cleared HBsAg as well as HBeAg with therapy (7). In studies of lamivudine and adefovir dipivoxil, loss of HBsAg occurred in only 1 to 2% of patients. In studies of continuous, maintenance therapy, HBsAg with development of anti-HBs allows for withdrawal of antiviral therapy and is reliably followed by long-term, durable remission in disease (19).
Initial Response: In comparing antiviral regimens, a definition of an initial response is often helpful. Fall of serum aminotransferases into the normal range and loss of detectable HBV DNA within 6 months of starting therapy has been used to define an initial response.
End-of-Treatment Response: Defining responses based upon laboratory or histological results at the end of therapy has been common in recent studies of antiviral therapies of hepatitis B. However, because most antiviral agents suppress rather than eradicate HBV replication, end-of-treatment responses are not very predictive of long-term benefit or remission in disease. Studies that use end-of-treatment response rates to assess efficacy should clearly state that responses are defined while therapy is still being administered.
Maintained Response: In evaluating continuous, maintenance antiviral therapy, a maintained response is perhaps a better means of assessing response rates. Since liver biopsies are done only at defined and infrequent time points, a maintained histological response should be defined by the timing of the sample. Because patients are enrolled into studies over time, a maintained response may be defined by biochemical and virological results at 1 year for some patients and 2 or 3 years for others. For this reason, 1-year, 2-year, and 3-year maintained response rates may be the appropriate approach to reporting results.
Sustained Response: In studies of chronic hepatitis C, a six-month post-treatment sustained response has been used to define response rates to antiviral therapy (21). This definition has been helpful and found to be reliable in chronic hepatitis C, even with introduction of new antiviral agents and new antiviral regimens. In chronic hepatitis B, a 6- to 8-month post-treatment sustained response was used to define rates of response to a interferon. Whether 6 months or 12 months after treatment is the most reliable point for such analyses requires further study. However, relapses in hepatitis B can occur 12 or more months after therapy is stopped. Thus, even after using a six or twelve month point of time to define “sustained”, responses must then continue to be assessed in the future for “durability.” In chronic hepatitis C, a 6-month post-treatment sustained loss of HCV RNA has been shown to be a reliable definition for a beneficial outcome of therapy and has subsequently been shown to be durable in over 98% of patients. The durability of 6 and 12-month post-treatment responses for hepatitis B have yet to be defined.
Back to top
The optimal approach to antiviral therapy of chronic hepatitis B remains undefined. The variable and unpredictable course of the disease, the lack of understanding of its immunopathogenesis, the variability in response to antiviral agents, and the lack of long-term information on the outcome of treatment all contribute to the difficulties in making reliable recommendations regarding therapy. Helpful would be the use of standard definitions both in diagnosis as well as in setting endpoints to therapy. Because chronic hepatitis B is a chronic disease with variable long-term outcomes, the most reliable information will come from long-term, longitudinal follow up of treated and untreated patients with all forms and degrees of disease.
Back to top
- Greenberg HB, Pollard RB, Lutwick LI, Gregory PB, Robinson WS, Merigan TC. Effect of human leukocyte interferon on hepatitis B virus infection in patients with chronic active hepatitis. N Engl J Med 1976;295:517-522.
- Lok AS, Heathcote EJ, Hoofnagle JH. Management of hepatitis B: 2000- summary of a workshop. Gastroenterology 2001;120:1828-1853.
- Manesis EK, Papatheodoridis GV, Hadziyannis SJ. Serum HBV-DNA levels in inactive hepatitis B virus carriers. Gastroenterology 2002;122:2092-2093.
- Martinot-Peignoux M, Boyer N, Colombat M, Akremi R, Pham BN, Ollivier S, Castelnau C, et al. Serum hepatitis B virus DNA levels and liver histology in inactive HBsAg carriers. J Hepatol 2002;36:543-546.
- Fong TL, Di Bisceglie AM, Gerber MA, Waggoner JG, Hoofnagle JH. Persistence of hepatitis B virus DNA in the liver after loss of HBsAg in chronic hepatitis B. Hepatology 1993;18:1313-1318.
- Di Bisceglie Am, Waggoner JG, Hoofnagle JH. Hepatitis B virus deoxyribonucleic acid in liver of chronic carriers: correlation with serum markers and changes associated with loss of hepatitis B e antigen after antiviral therapy. Gastroenterology 1987;93:1236-1241.
- Wong DKH, Cheung AM, O'Rourke K, Naylor CD, Detsky AS, Heathcote J. Effect of alpha-interferon treatment in patients with hepatitis B e antigen-positive chronic hepatitis B. Ann Intern Med 1993;119:312-323.
- Niederau C, Heintges T, Lange S, Goldmann G, Niederau CM, Mohr L, Häussinger D. Long-term follow-up of HBeAg-positive patients treated with interferon alfa for chronic hepatitis B. N Engl J Med 1996;334:1422-1427.
- Lau D-Y, Everhart J, Kleiner DE, Park Y, Vergalla J, Schmid P, Hoofnagle JH. Long-term follow-up of patients with chronic hepatitis B treated with interferon a. Gastroenterology 1997;113:1660-1667.
- Lok ASF, Chung H-T, Liu VWS, Ma OCK. Long-term follow-up of chronic hepatitis B patients treated with interferon alfa. Gastroenterology 1993;105:1833-1838.
- Fattovich G, Farci P, Rugge M, Brollo L, Mandas A, Pontisso P, Giustina G, et al. A randomized controlled trial of lymphoblastoid interferon-alfa in patients with chronic hepatitis B lacking HBeAg. Hepatology 1992;15:584-589.
- Dienstag JL, Perrillo RP, Schiff ER, Bartholomew M, Vicary C, Rubin M. A preliminary trial of lamivudine for chronic hepatitis B infection. N Engl J Med 1995;333:1657-1661.
- Lai C-L, Chien R-N, Leung NWY, Chang T-T, Guan R, Tai D-I, Ng K-Y, et al. A one-year trial of lamivudine for chronic hepatitis B. Asia Hepatitis Lamivudine Study Group. N Engl J Med 1998;339:61-68.
- Dienstag JL, Schiff ER, Wright TL, Perrillo RP, Hann HWL, Goodman Z, Crowther L, et al. Lamivudine as initial treatment for chronic hepatitis B in the United States. N Engl J Med 1999;341:1256-1263.
- Marcellin P, Goodman Z, Chang T-T, Lim SG, Tong M, Sievert W, Shiffman M, et al. Histological improvement in HBeAg positive chronic hepatitis B patients treated with adefovir dipivoxil. J Hepatol 2002;36(Suppl 1): 8. [abstract]
- Hadziyannis S, Tassopolous N, Heathcote E, Chang T-T, Kitis G, Rizzetto M, Marcellin P, et al. GS-98-438. A double-blind, randomized, placebo-controlled study of adefovir dipivoxil (ADV) for presumed precore mutant chronic hepatitis B: 48 week results. J Hepatol 2002;16(Suppl 1):4. [abstract]
- Song BC, Suh DJ, Lee HC, Chung Y-H, Lee YS. Hepatitis B e antigen seroconversion after lamivudine therapy is not durable in patients with chronic hepatitis B in Korea. Hepatology 2000; 32:803-806.
- Tipples GA, Ma MM, Fischer KP, Bain VG, Kneteman NM, Tyrrell DLJ. Mutation in the HBV RNA dependent DNA polymerase confers resistance of lamivudine in vivo. Hepatology 1996;24:714-717.
- Lau DT-Y, Khokhar F, Doo E, Ghany MG, Herion D, Park Y, Kleiner DE, et al. Long-term therapy of chronic hepatitis B with lamivudine. Hepatology 2000; 32:838-834.
- Heathcote E, Jeffers L, Perrillo R, Wright T, Sherman M, Namini H, Xiong S, et al. Sustained antiviral response and lack of viral resistance with long term adefovir dipivoxil (ADV) therapy in chronic HBV infection. J Hepatol 2002;36 (Suppl 1):110. [abstract]
- Lindsay KL. Therapy of hepatitis C: overview. Hepatology 1997; 26 (Suppl 1): 71S-77S.
- Desmet VJ, Gerber M, Hoofnagle JH, Manns M, Scheuer PJ. Classification of chronic hepatitis: diagnosis, grading and staging. Hepatology 1994;95:1513-1520.
Back to Conference Reports