Long-term Therapy With Adefovir Dipivoxil for HBeAg-Negative Chronic Hepatitis B for up to 5 Years
Gastroenterology Dec 2006
Stephanos J. Hadziyannis_, Nicolaos C. Tassopoulos, E. Jenny Heathcote, Ting-Tsung Chang, George Kitis, Mario Rizzetto#, Patrick Marcellin__, Seng Gee Lim, Zachary Goodman, Jia Ma, Carol L. Brosgart, Katyna Borroto-Esoda, Sarah Arterburn, Steven L. Chuck
_ Department of Medicine and Hepatology, Henry Dunant Hospital, Athens, Greece
Metropolitan Hospital, Athens, Greece
Toronto Western Hospital, University of Toronto, Toronto, Canada
Department of Internal Medicine, National Cheng Kung University Hospital, Tainan, Taiwan
Georgios Papanikolaou Hospital, Thessaloniki, Greece
# Azienda Ospedaliera San Giovanni Battista, Turin, Italy
__ Service d'Hepatologie, INSERM Unite 481, Centre de Recherche Claude Bernard sur les Hepatites Virales, Hopital Beaujon, Clichy, France
Division of Gastroenterology, National University Hospital, Singapore
Gilead Sciences, Foster City, California
Background & Aims: Treatment with adefovir dipivoxil for 48 weeks resulted in clinical improvement in patients with hepatitis B e antigen (HBeAg)-negative chronic hepatitis B that was lost when treatment was discontinued. We investigated the efficacy, safety, and resistance profile of adefovir dipivoxil treatment for up to 240 weeks.
Methods: HBeAg-negative patients were treated double blind with placebo or adefovir dipivoxil 10 mg once daily for 48 weeks, followed by adefovir dipivoxil from week 49 to 96. At week 97, 125 patients enrolled in a 144-week, open-label phase. Patients received adefovir dipivoxil for up to 192 or 240 weeks.
Serum hepatitis B virus (HBV) DNA levels were less than 1000 copies per milliliter in 67% of patients, and alanine aminotransferase (ALT) levels normalized in 69% after 240 weeks.
After 192 or 240 weeks of treatment, over 83% of patients had improvement in necroinflammation, and over 73% had improvement in fibrosis.
Ishak fibrosis scores improved compared with baseline in 35%, 55%, and 71% of patients after 48, 192, and 240 weeks of adefovir dipivoxil, respectively.
After 240 weeks, the cumulative probability of HBV polymerase mutations was 29%, but the cumulative probability of mutations with virologic resistance was 20% and of mutations, virologic resistance, and ALT elevations was 11%.
Slight elevations in creatinine were confirmed in 4 (3%) patients.
Conclusions: Treatment with adefovir dipivoxil for up to 240 weeks was well tolerated and produced significant, increasing improvement in hepatic fibrosis, durable suppression of HBV replication, normalization of liver enzymes, and delayed development of resistance.
Chronic infection with the hepatitis B virus (HBV) is a major public health problem worldwide.1, 2 Two of its complications, cirrhosis and hepatocellular carcinoma, result in the deaths of over a half million people annually.1, 2 Progression to these complications is more frequent among persons with hepatitis B e antigen (HBeAg)-negative disease who harbor replicating HBV with mutations in its precore or core promoter regions.3 Patients with HBeAg-negative, chronic hepatitis B have severe liver necroinflammation; negligible rates of spontaneous, sustained remission; and require continuing treatment.4, 5
Currently approved therapies for HBeAg-negative chronic hepatitis B include standard and pegylated interferon-ƒ¿,6, 7, 8, 9 lamivudine,10, 11, 12 adefovir dipivoxil,13, 14 and entecavir.15 Finite courses of interferon therapy can produce sustained responses and even loss of hepatitis B surface antigen (HBsAg), but the rates are generally low, and treatment can be poorly tolerated. Oral nucleoside or nucleotide analogues can effectively suppress HBV replication, but this response is rarely sustained after cessation of treatment. Long-term therapy with lamivudine is initially effective but is compromised by resistance developing in 23% of patients after 1 year, 46% after 2 years, and 71% after 4 years.16 Both adefovir dipivoxil and entecavir are active against wild-type and lamivudine-resistant HBV, but the long-term safety and efficacy of entecavir are unknown. We studied patients with HBeAg-negative chronic hepatitis B treated for up to 240 weeks with adefovir dipivoxil.
Materials and Methods
This 240-week study consisted of a double-blinded phase of 96 weeks followed by an open-label, long-term safety and efficacy study of 144 weeks (Figure 1). This international, prospective trial enrolled 185 patients between January 10 and June 7, 2000, at 32 sites.13, 14 Patients were randomly assigned to receive 10 mg adefovir dipivoxil (n = 123) or placebo (n = 62; 1 patient did not take study drug) once daily in a double-blind manner for the first 48 weeks. At week 49, 119 patients initially treated with adefovir dipivoxil were randomized and continued treatment (adefovir-adefovir group; n = 79) or switched to placebo (adefovir-placebo group; n = 40), and patients initially given placebo were switched to adefovir dipivoxil (placebo-adefovir group; n = 60), all in a double-blind manner. From weeks 49-96, 5 patients receiving adefovir dipivoxil discontinued treatment (1 for treatment-related creatinine elevation). At week 96, 125 patients being treated with adefovir dipivoxil continued therapy in the long-term safety and efficacy study (70 from adefovir-adefovir and 55 from placebo-adefovir groups), 9 patients chose not to continue, and 40 patients in the adefovir-placebo group entered another study. This report details long-term efficacy and safety in the 125 patients enrolled in the long-term safety and efficacy study but describes resistance and hepatocellular carcinoma in all 183 patients13, 14 treated with adefovir dipivoxil at any time during the 240-week study. Treatment ended at study week 240; therefore, patients in the adefovir-adefovir group received adefovir for up to 240 weeks and those in the placebo-adefovir group for up to 192 weeks.
All patients gave written informed consent. The study was conducted in compliance with the 1975 Declaration of Helsinki and approved by local regulatory bodies. All laboratory data were assessed by Covance Laboratories, and clinical data were monitored and stored by Quintiles. The sponsor held all data and conducted statistical analyses.
All inclusion criteria have been described previously.13, 14 Key entry criteria included detection of HBsAg for at least 6 months, HBeAg-negative/anti-HBeAg-positive serology, compensated liver disease, serum HBV DNA level greater than or equal to 100,000 copies per milliliter, and alanine aminotransferase (ALT) between 1.5 and 15 times the upper limit of normal.
In the open-label phase, the end points included the percentage of patients with serum HBV DNA less than 1000 copies per milliliter, normalization of ALT levels, HBsAg seroconversion, and histologic improvement (evaluated by ranked assessments: improved, no change, or worse; Knodell scores17, 18; and Ishak fibrosis score19).
Adverse events and laboratory tests were assessed every 4 weeks to week 96 and then every 12 weeks thereafter. The Roche Amplicor PCR assay (lower limit of quantification [LLQ] 400 copies per milliliter until January 2002 then 1000 copies per milliliter) was used until May 2004 when the assay was discontinued by the manufacturer (Roche Diagnostics, Indianapolis, IN). HBV DNA levels were then determined by the Roche TaqMan PCR assay (LLQ 169 copies per milliliter; 5.82 copies/mL = 1 IU/mL). Liver biopsies were required from all patients within 6 months of screening and at week 48 but were optional at weeks 96, 144, 192, and 240. All liver biopsy specimens were assessed by one independent histopathologist, who was unaware of treatment assignment or biopsy dates.
Genotypic analysis of the entire HBV polymerase20 was performed for all patients at baseline and for patients with serum HBV DNA greater than the LLQ of the PCR assay in use on samples at study weeks 48, 96, 144, 192, and 240.
All tests for significance were 2-sided (except for the trend test) with P < .05 considered to be statistically significant. HBV DNA and ALT levels were analyzed by intention-to-treat, missing equals failure (ITT; M = F), and intention-to-treat, missing equals failure for resistance or hepatocellular carcinoma, missing is excluded for other reasons (ITT; M = F R/HCC) to provide a more realistic view of efficacy because dropouts unrelated to efficacy are expected in a 5-year trial. In this second type of analysis (ITT; M = F R/HCC), patients were considered failures if they (1) harbored HBV with an adefovir-resistance mutation and either terminated the study or had lamivudine added (if a patient had HBV with a resistance mutation and remained in the study on adefovir dipivoxil monotherapy, his or her serum HBV DNA and ALT values were included in analyses rather than being deemed failures) or (2) were diagnosed with hepatocellular carcinoma; the missing values from patients who left the study for other reasons were excluded. Cumulative probabilities of resistance were calculated using the life-table method.21 The Wald χ2 test was used to assess statistical significance in the stepwise logistic regression. Since logistic regression assumes the same length of follow-up for all patients, potential predictors of resistance were evaluated for patients treated up to 192 weeks to maximize the number of analyzed patients.
All patients who received at least 1 dose of adefovir dipivoxil were included in the safety analysis. Efficacy and safety data are reported from each patient's first dose of adefovir dipivoxil.
Patient Disposition and Characteristics
After 96 weeks of randomized, double-blind treatment with adefovir dipivoxil and/or placebo, 125 patients enrolled in the long-term safety and efficacy study. In the placebo-adefovir group, 87% of patients (48/55) received adefovir dipivoxil for 192 weeks (1 patient discontinued the study because of HBsAg seroconversion), and, in the adefovir-adefovir group, 86% of patients (60/70) received adefovir dipivoxil for 240 weeks. The disposition of patients throughout the entire 240-week study is shown in Figure 1.
The characteristics of the 125 patients at study baseline were similar; however, patients given placebo for the first 48 weeks of the study had lower HBV DNA and ALT levels immediately prior to treatment with adefovir dipivoxil (Table 1), most likely because of fluctuations in viral replication and disease activity observed in HBeAg-negative patients.3, 4 Pretreatment median Knodell necroinflammation and Ishak fibrosis scores were 7.0 and 2.0, respectively, for all patients (Table 1).
Virologic and Biochemical Responses
In the double-blind phase, 72% of the patients who eventually enrolled in the open-label phase had undetectable serum HBV DNA after treatment with adefovir dipivoxil for 48 weeks (Table 2). As expected in a long-term study, analysis by ITT (M = F) showed a decline in the percentage of patients with HBV DNA less than 1000 copies per milliliter to 53% after 240 weeks of treatment. However, the proportion of patients with serum HBV DNA less than 1000 copies per milliliter was generally sustained through 240 weeks of treatment when the effect of patients dropping out for reasons other than resistance or hepatocellular carcinoma is considered (ITT, M = F R/HCC; Table 2). In Table 2, a consistent threshold of 1000 copies per milliliter is used to show the virologic response over 5 years despite changes in the LLQ of the PCR assay used from 400 to 1000 to 169 copies per milliliter (see Materials and Methods section). At week 240, when the Taqman assay was used for all samples, 67% of patients had serum HBV DNA less than 1000 copies per milliliter, 62% of patients had values less than 400 copies per milliliter, and 58% of patients had values less than 169 copies per milliliter (ITT, M = F R/HCC). Table 2 also shows the distribution of virologic suppression in patients over 5 years. An increasing proportion of patients demonstrated virologic responses through 96 weeks; the effects of resistance became evident after 144 weeks.
Similarly, normalization of ALT levels after 48 weeks of adefovir dipivoxil occurred in 75% of these patients (Table 2). The proportion of patients with ALT normalization was generally sustained through 240 weeks when the effect of patients dropping out for reasons other than resistance or hepatocellular carcinoma is considered. Six patients (5%) had HBsAg loss after a median of 196 weeks (range, 20-260) of adefovir dipivoxil; 5 of these patients had developed antibody to hepatitis B surface antigen (anti-HBs) at their last measurement.
Significant improvements in liver histology were observed following long-term treatment with adefovir dipivoxil. By ranked assessment, liver biopsies after 192 (placebo-adefovir group) or 240 weeks (adefovir-adefovir group) of treatment showed that 86% and 83% of patients had improvement in necroinflammation and 73% and 75% had improvement in fibrosis compared with their pretreatment biopsy, respectively. The median change in Knodell necroinflammation score from the time patients started on adefovir dipivoxil was -4.5 points at 192 weeks and -5.0 points at 240 weeks, and the median change in Ishak fibrosis score was -1.0 point for both groups. The proportion of patients with at least a 1-point improvement in Ishak fibrosis score increased from 35% after 48 weeks of adefovir dipivoxil to 55% and 71% after 192 and 240 weeks of treatment, respectively (Figure 2). Seven of 12 (58%) patients with pretreatment bridging fibrosis or cirrhosis improved in their Ishak fibrosis scores by at least 2 points, and 3 of the 4 patients with cirrhosis improved by 4 points.
Importantly, patients who underwent liver biopsy after 192 or 240 weeks of adefovir dipivoxil were representative of all treated patients. Among the patients enrolled in the open-label phase, those who did or did not have liver biopsies at study week 240 had similar baseline characteristics, and there were no statistically significant differences in the baseline Knodell necroinflammation or Ishak fibrosis scores (Table 3). Furthermore, among patients who completed study drug for the entire open-label phase and either did or did not have a liver biopsy at week 240, there were no statistically significant differences in the percentages with ALT normalization after 48 or 240 weeks of adefovir dipivoxil (Table 3). The percentage of patients with serum HBV DNA <1000 copies/mL at treatment week 48, however, was significantly lower in the group who underwent liver biopsy at the end of the study (Table 3). At study week 240, there were no statistically significant differences in the percentages with serum HBV DNA <1000 copies/mL between patients who did or did not have a liver biopsy at week 240 (Table 3). Finally, the percentage of patients with at least a 1-point improvement in Ishak fibrosis score after 48 weeks of adefovir dipivoxil was 35% in the adefovir-adefovir group (Figure 2) compared with 34% among all patients (n = 113) initially treated with 48 weeks of adefovir dipivoxil in the double-blind phase.22 Thus, patients who underwent liver biopsy at the end of the study were representative of the patients in the open-label phase who did not have a biopsy and of all patients treated with adefovir dipivoxil.
The incidence of 3 types of events were investigated in all patients treated with adefovir dipivoxil throughout the 240-week study (n = 183): (1) adefovir-resistance mutations (N236T or A181V) regardless of HBV DNA and ALT levels ("mutations"), (2) adefovir-resistance mutations with HBV DNA increased from nadir by at least 1-log10 copies per milliliter (confirmed or last measurement) or never suppressed to less than 3-log10 copies per milliliter ("virologic resistance"), and (3) adefovir-resistance mutations with virologic resistance and ALT elevations (ALT greater than the upper limit of normal after normalizing ALT; "clinical resistance").
Twenty-nine patients had HBV harboring the N236T and/or A181V adefovir-resistance mutations detected during the 240-week study. The first mutation(s) detected were N236T (n = 15), A181V (n = 9), or A181V + N236T (n = 5). Five patients with N236T initially detected later developed A181V + N236T. One patient had HBV with A181T detected at week 192; however, this patient is not included with the 29 patients who developed adefovir-resistance mutations because A181T does not confer phenotypic resistance to adefovir (the IC50 of the A181T mutant is increased only 1.3-fold compared with wild-type HBV, and the double-mutant N236T/A181T has a 3-fold lower IC50 compared with the N236T mutant).23 One patient entered the study with a mixture of wild-type HBV and HBV carrying I233V. This patient had undetectable serum HBV DNA within the first 48 weeks that was sustained. No patient developed I233V during the 5 years of the study. The median time to development of mutations was 192 weeks.
Twenty-one patients with mutations (21/29) experienced virologic resistance. These 21 patients include 17 patients with serum HBV DNA increased from nadir by at least 1-log10 copies per milliliter, 1 patient whose serum HBV DNA was increased from nadir by at least 1-log10 copies per milliliter 1 day after discontinuing adefovir dipivoxil, 1 patient whose serum HBV DNA was never suppressed to less than 3-log10 copies per milliliter,24 and, to be conservative, 2 patients with serum HBV DNA that increased by at least 0.6-log10 copies per milliliter above 1000 copies per milliliter because the PCR assay used earlier in the study had a lower limit of quantification of 400 copies per milliliter (which is 0.4 log lower than 1000 copies per milliliter). The median time of follow-up on adefovir dipivoxil monotherapy after detection of a mutation was 36 weeks in patients with virologic resistance compared with 12 weeks in those who had mutations without manifesting virologic resistance.
Thirteen (13/29) patients had clinical resistance (adefovir-resistance mutations with virologic resistance and ALT elevations). The median time of follow-up on adefovir dipivoxil monotherapy was 36 weeks in these patients. Three patients had ALT elevations greater than 10 times the upper limit of normal (2 patients discontinued the study, and 1 patient received lamivudine plus adefovir dipivoxil and subsequently had undetectable serum HBV DNA and normal ALT). The cumulative probabilities of developing mutations, virologic resistance, and clinical resistance were calculated for all patients on adefovir dipivoxil and at risk for each of the 3 events every 48 weeks (Table 4). At 240 weeks, the cumulative probabilities of mutations, virologic resistance, and clinical resistance were 29%, 20%, and 11%, respectively.
Adefovir-resistance mutations occurred more frequently among patients who had serum HBV DNA greater than 1000 copies per milliliter at 48 weeks. In a stepwise logistic regression analysis that included demographics, genotype, and baseline fibrosis, only detectable serum HBV DNA at week 48 was a significant predictor of resistance over 192 weeks (P = .0003). Forty-nine percent (49%; 17/35) of patients with serum HBV DNA greater than 1000 copies per milliliter after 48 weeks of adefovir dipivoxil developed mutations on up to 192 weeks of treatment compared with 6% (5/89) of patients with serum HBV DNA less than 1000 copies per milliliter at 48 weeks. Eleven patients who developed adefovir-resistance mutations were subsequently treated with lamivudine-as monotherapy or added to adefovir dipivoxil-which resulted in reductions of 2 to 6 log10 in HBV DNA after a median follow-up of 24 weeks.
Liver biopsies were obtained from 13 patients a median of 45 weeks (range, 8-153) after detection of a mutation. Five of these patients had clinical resistance a median of 1 year prior to liver biopsy. Improvements relative to pretreatment histology in both necroinflammatory (≥2-point decrease) and Ishak fibrosis scores (≥1-point decrease) were seen in 8 of 13 patients including 2 patients with clinical resistance, improvement in necroinflammation with no change in fibrosis was observed in 2 of 13 patients (both patients with clinical resistance), and 3 of 13 patients had worsening in 1 of the scores (the 1 patient with worsening fibrosis had clinical resistance).
Adverse events occurring in at least 10% of the 125 patients are shown in Table 5. In the 144-week open-label phase, 3 patients discontinued because of an adverse event. Two of these patients (2/125 or 1.6%) had an adverse event judged by investigators to be related to adefovir (elevated creatinine in 1 patient; elevated ALT in another patient in the setting of mutations and virologic resistance). Among 125 patients enrolled in the open-label phase, 4 patients (3%) had confirmed increases in serum creatinine greater than or equal to 0.5 mg per deciliter above pretreatment values. The maximum value was 1.5 mg per deciliter, and the maximum increase was 0.8 mg per deciliter. No patient had confirmed serum hypophosphatemia while on adefovir dipivoxil therapy. All serious adverse events (n = 24; 19%) were considered unrelated to adefovir dipivoxil.
Six of the 183 patients (3%) who were treated with adefovir dipivoxil in the 240-week study developed hepatocellular carcinoma. With the exception of 1 patient whose tumor was diagnosed after 48 weeks of therapy, patients with hepatocellular carcinoma were diagnosed after 112-219 weeks of treatment with adefovir dipivoxil. Five of the 6 patients had substantial pretreatment necroinflammation (Knodell necroinflammatory score 7-11 points) and fibrosis (bridging fibrosis or cirrhosis). At the last on-treatment sample before diagnosis of their tumors, HBV DNA was approximately 5-log10 copies per milliliter in 1 patient, between 3- and 4-log10 copies per milliliter in 4 patients, and was undetectable in 1 patient. A resistance mutation was detected in 1 patient 10 weeks prior to diagnosis of hepatocellular carcinoma.
Treatment with adefovir dipivoxil for up to 240 weeks in this study produced sustained suppression of serum HBV DNA levels, normalization of ALT levels, and increasing improvement in liver fibrosis. In addition, most patients with advanced liver disease prior to treatment, including cirrhosis, improved markedly. The first oral antiviral therapy to show reversal of fibrosis with 3 years of treatment was lamivudine; however, improvement was reversed after the frequent development of resistance.25 Optional, end-of-study liver biopsy specimens were obtained in 46 of the 125 patients (37%) enrolled in our long-term safety and efficacy study. These are the first 240-week, prospective set of liver biopsy specimens, and they demonstrated increasing histologic improvement. Furthermore, 6 patients (5%) experienced HBsAg loss and 5 seroconverted to anti-HBs. This study substantiates the efficacy of long-term treatment with adefovir dipivoxil for HBeAg-negative chronic hepatitis B.
Long-term therapy with antiviral drugs can select for mutations associated with resistance. Definitions of resistance have varied widely, making it difficult to compare the incidences for different anti-HBV compounds. We defined 3 subsets of patients harboring HBV with adefovir-associated mutations in the polymerase gene. After 240 weeks of adefovir dipivoxil therapy, the cumulative rates of virologic resistance (20%) and clinical resistance (11%) were much lower than the cumulative rate of mutations (29%). Over the 5 years of this study, mutations were detected by sequencing; presently, tests are available that are more sensitive for detection of viral quasispecies. A more sensitive assay might have detected a higher incidence of genotypic resistance. It should be emphasized that mutations were detected in this study by annual surveillance of all patients rather than just of patients with virologic resistance. If mutations were sought only in patients with virologic breakthrough or lack of suppression, the cumulative probability over 240 weeks would have been 20%. Furthermore, patients in our study with detectable serum HBV DNA at 48 weeks were not discontinued. Since these patients are at increased risk for the development of resistance, discontinuation of such patients from a trial yields a lower rate of resistance. These 3 definitions of resistance and annual surveillance of all treated patients should be used to provide meaningful comparisons of drug regimens.
As expected, follow-up was longer for patients with manifestations of resistance-whether virologic breakthrough or ALT elevations-than for those with mutations only. These data suggest that, over several months, patients with HBV encoding N236T and/or A181V mutations can eventually progress to virologic resistance with or without ALT elevations. Patients with adefovir-resistant HBV were treated effectively with the addition of lamivudine while continuing adefovir dipivoxil or by switching from adefovir dipivoxil to lamivudine.
The necessity of prolonged therapy in HBeAg-negative hepatitis B underscores the need for long-term safety. Adefovir dipivoxil was well tolerated for up to 240 weeks. Among 125 patients, only 2 patients (1.6%) discontinued adefovir dipivoxil because of adverse events judged to be related to treatment. Thus, the long-term efficacy and low rates of resistance of adefovir dipivoxil are not offset by adverse effects.
Hepatocellular carcinoma was diagnosed in 6 (6/183) patients. It is likely that some patients developed tumors prior to the initiation of adefovir dipivoxil therapy. Five patients developed tumors in the setting of marked baseline liver necroinflammation and fibrosis. Patients with a long history of infection with HBV have a significant risk of hepatocellular carcinoma despite antiviral treatment late in the course of the disease. This increased risk of hepatocellular carcinoma remains even if clearance of HBsAg occurs.26 It remains to be investigated whether treating patients at earlier stages of liver disease with adefovir dipivoxil will decrease the incidence of this tumor.
Treatment with adefovir dipivoxil for up to 240 weeks produces sustained suppression of HBV DNA levels, durable normalization of ALT, and continued improvement in liver histology. The selection of mutations in HBV polymerase is delayed, and virologic resistance or elevation in ALT levels occurs even later. With its safety, durable efficacy, and delayed resistance for up to 5 years, adefovir dipivoxil is a primary therapeutic option for the treatment of HBeAg-negative chronic hepatitis B.