Long-term efficacy and safety of adefovir dipivoxil for the treatment of hepatitis B e antigen–positive chronic hepatitis B


  • Potential conflict of interest: Drs. Arterburn, Borroto-Esoda, Frederick, and Rousseau own stock in Gilead Sciences. Dr. Lim advises Novartis, Bristol-Myers Squibb, and Schering-Plough. He is on the speakers' bureau of GlaxoSmithKline. Dr. Marcellin advises, is on the speakers' bureau of Roche, Schering-Plough, Gilead, Bristol-Myers Squibb, GlaxoSmithKline, and Idenix-Novartis. He is also a consultant and advises Vertex, Valeant, Cytheris, Intermune, Wyeth, and Tibotec. Dr. Sievert is on the Speakers' bureau of Gilead Sciences.


Treatment of 171 patients with hepatitis B e antigen (HBeAg)-positive chronic hepatitis B (CHB) with adefovir dipivoxil (ADV) 10 mg over 48 weeks resulted in significant histological, virological, serological, and biochemical improvement compared with placebo. The long-term efficacy and safety of ADV in a subset of these patients was investigated for up to 5 years. Sixty-five patients given ADV 10 mg in year 1 elected to continue in a long-term safety and efficacy study (LTSES). At enrollment, the 65 LTSES patients were a median 34 years old, 83% male, 74% Asian, 23% Caucasian, median baseline serum hepatitis B virus (HBV) DNA 8.45 log10 copies/mL, and median baseline alanine aminotransferase (ALT) 2.0 × upper limit of normal. At 5 years on study, the median changes from baseline in serum HBV DNA and ALT for the 41 patients still on ADV were 4.05 log10 copies/mL and −50 U/L, respectively. HBeAg loss and seroconversion were observed in 58% and 48% of patients by end of study, respectively. Fifteen patients had baseline and end of follow-up liver biopsies; improvements in necroinflammation and fibrosis were seen in 67% and 60% of these patients, respectively. Adefovir resistance mutations A181V or N236T developed in 13 LTSES patients; the first observation was at study week 195. There were no serious adverse events related to ADV. Conclusion: Treatment with ADV beyond 48 weeks was well tolerated and produced long-term virological, biochemical, serological, and histological improvement. (HEPATOLOGY 2008;48:750–758.)

Chronic infection with the hepatitis B virus (HBV) affects an estimated 400 million people worldwide and continues to be an important cause of morbidity and mortality, as well as a source of potential new infections.1 An estimated one million people die annually of complications of hepatitis B.2 Chronic hepatitis B (CHB) will require prolonged if not lifelong therapy in a significant number of patients; therefore, long-term safety and efficacy data for antiviral medications are indispensable to guide the risk–benefit assessment for a patient.

Adefovir dipivoxil (ADV) is an oral prodrug of adefovir, a phosphonate nucleotide analog of adenosine monophosphate with activity in vitro against a variety of viruses, including hepadnaviruses such as HBV. Adefovir is metabolized intracellularly to adefovir diphosphate, which is a competitive inhibitor of HBV DNA polymerase that results in termination of the growing DNA chain. In hepatitis B e antigen (HBeAg)-negative CHB patients treated for up to 240 weeks, ADV has been shown to be well tolerated and produced significant improvement in hepatic fibrosis, durable suppression of HBV replication, normalization of liver enzymes, and delayed development of resistance.3

Study GS-98-437 was designed to evaluate ADV 10 and 30 mg once daily compared with placebo with respect to changes in liver histology, HBeAg seroconversion, serum HBV DNA, and alanine transaminase (ALT) levels in patients with HBeAg-positive CHB infection. The study was to be conducted in three sequential periods: year 1 (study weeks 1–48) and year 2 (study weeks 49–96) were double-blind, randomized, parallel group, placebo-controlled treatment periods; years 3 to 5 (study weeks 97–240), the long-term safety and efficacy study (LTSES), was added later to provide open-label treatment with ADV 10 mg for those patients who did not have durable loss or seroconversion for e antigen. The safety and efficacy findings of the double-blind portion of the study through year 1 have been reported.4 The current report presents efficacy results through week 240 for all patients who received ADV 10 mg during year 1 of the study, and safety results for the subset of patients who were enrolled in the LTSES and received up to 5 years of ADV.


ADV, adefovir dipivoxil; ADV10-Y1, patients who received at least one dose of ADV 10 mg in year 1 of the study; AE, adverse event; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CHB, chronic hepatitis B; HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; LTSES, long-term safety and efficacy study; OL, open-label; ULN, upper limit of normal.

Patients and Methods

Study Design.

The study was conducted in compliance with the Declaration of Helsinki and approved by appropriate local regulatory bodies. All patients provided written informed consent. Enrollment timelines, study centers, randomization procedures, and inclusion and exclusion criteria have been described previously.4

Initially patients were randomized equally to treatment with ADV 30 mg once daily, ADV 10 mg once daily, or placebo. At the beginning of year 2 (study week 49), patients who had taken ADV 30 mg once daily were to be converted to treatment with placebo, and patients who had taken placebo were to be converted to treatment with ADV 10 mg once daily. Patients who had taken ADV 10 mg were to be re-randomized equally to treatment with either ADV 10 mg once daily or placebo.

However, during year 2, it was learned that study medication had been misallocated during the second year of the study due to an error in the interactive voice response system used to assign blinded treatment. No misallocation errors were found in any year 1 treatment assignments. A total of 444 of the 466 patients (95%) who were treated in year 2 of the study received at least one bottle of study medication that either contained a medication or dose that was different from what they received as their first dose in year 2, or contained a medication or dose that, together with their year 1 treatment assignment, constituted a treatment sequence that was not allowed per protocol, which resulted in variable dosing in the second year of the double-blind phase of the study. Because of this, a decision was made to terminate the blinded phase of the study, and all patients were offered treatment with open-label (OL) ADV 10 mg for up to 4 months (Table 1). Later, a 3-year LTSES was added, and 65 patients chose to continue long-term treatment with ADV 10 mg.

Table 1. Study Treatment Timeline
Year 1Year 2Year 5
Weeks 1–48 (Blinded)Weeks 49–96 (Blinded Until Discovery of Misallocation)Open-Label (OL) PeriodLong-Term Safety and Efficacy Study (LTSES)
  • *

    Patients initially randomized to placebo or ADV 30 mg were discontinued from the study after the OL period.

ADV 30 mgPlaceboADV 10 mgOff study*
ADV 10 mgRe-randomized to:ADV 10 mgADV 10 mg
ADV 10 mg ADV 10 mg or Placebo  
PlaceboADV 10 mgADV 10 mgOff study*

Patients received misallocated medication as early as their first dose in year 2 (week 49). The median duration that patients initially randomized to ADV 10 mg received misallocated medication was 28 weeks (range, 4–44 weeks; n = 154). review times for Institutional Review Boards and Independent Ethics Committees, some patients suspended treatment for a period until approval was obtained to start OL ADV 10 mg. The median duration of time that patients were off-treatment before the OL period was 16 weeks (n = 91). By study week 120, 85% of patients were back on treatment (open-label ADV 10 mg daily), and by study week 169, all patients were back on treatment. Of the 65 patients who enrolled into LTSES, 64 had misallocation of study drug with a median duration of 31.2 weeks (range, 4.1–43.7 weeks). All 65 of the LTSES patients were off treatment at the end of blinded phase for a median of 16 weeks (range, 3.1–70.3 weeks), and 5 patients were off treatment at the end of the open-label phase for a median of 0.3 weeks (range, 0.1 to 16 weeks).

Requirements for a patient to enter the LTSES included: (1) randomization to ADV 10 mg once daily in the first year of the study and on study up to the discovery of misallocation of study medication; (2) completion of at least 4 months of open-label ADV 10 mg treatment; and (3) no evidence of durable HBeAg loss (defined as two consecutive HBeAg−, HBV DNA undetectable, and ALT normalization at visits 12 weeks apart) or confirmed HBeAg seroconversion or protocol-defined treatment-limiting toxicity in the study. Patients initially randomized to placebo or ADV 30 mg were discontinued from the study but had the option to consider an open-label rollover protocol.

In the LTSES, patients were treated with ADV 10 mg once daily, preferably at the same time each day. Assessments were scheduled at each visit (every 3 months) during the LTSES. Patients who were not on open-label ADV 10 mg treatment and who entered the LTSES were required to return to the clinic every 4 weeks for the first 12 weeks. A liver biopsy was recommended annually where possible and at end of study or early termination (within 2 weeks of the last ADV dose). Patients were followed for 4 months after receiving their last dose of ADV to monitor for flares, unless they were receiving ADV treatment outside the setting of this clinical study.

The objectives of the LTSES were to investigate the safety, tolerability, and efficacy of long-term treatment with 10 mg ADV. The extent and severity of liver damage on each biopsy were assessed using two scoring systems, the Knodell/Histological Activity Index score5, 6 and the Ishak system.7 Additionally, potential emergence of ADV-related resistance mutations in the HBV polymerase gene was evaluated as reported previously,4 as were the virological, biochemical, and serological outcomes of CHB patients who developed confirmed resistance to ADV.

Patients who received at least one dose of ADV 10 mg in year 1 of the study are referred to as the ADV10-Y1 efficacy population. Patients in the LTSES safety subset included patients from the ADV10-Y1 efficacy population who received at least one dose of ADV 10 mg in the LTSES period of the study (i.e., the LTSES safety population is a subset of the ADV10-Y1 efficacy population).

All patients who received ADV 10 mg at any time in the double-blind portion of the study (n = 309) were screened for confirmed (two consecutive values) serum creatinine increases from baseline to 0.5 mg/dL or greater and for confirmed phosphorus values less than 2.0 to evaluate renal parameters in the broadest possible sample.

Statistical Methods.

Study weeks were calculated based on weeks from the first dose of ADV 10 mg. Duration of misallocation of dosing was defined as months from patient receiving their first dose of medication different from that received on the first dose in year 2 (or from time of first dose in year 2 if the double-blind treatment sequence was not one that was specified in the protocol) to the minimum of [date of start of OL minus 1 day, date of start of LTSES minus 1 day, or end of study]. HBV DNA, ALT, HBeAg loss and seroconversion, and hepatitis B surface antigen loss and seroconversion analyses were conducted across the entire duration of the study in the 171 patients who received at least one dose of ADV 10 mg in year 1 of the study. Percent of patients with HBV DNA less than 1000 c/mL and ALT normalization were calculated by study visit, for patients with HBV DNA 1000 c/mL or greater, and ALT greater than the upper limit of normal (ULN) at baseline, respectively. Data collected after discontinuation of ADV or after adding an additional antiviral was excluded. A missing = failure rule was used if data were missing because of death, discontinuation due to hepatocellular carcinoma, or resistance (otherwise missing data were excluded). Changes from baseline in HBV DNA and ALT were calculated using available data.

The number and proportion of patients who achieved confirmed (two consecutive or last available value) HBeAg loss or seroconversion by the end of the study is presented. Off-treatment follow-up data collected after discontinuation of ADV were included; data after adding another antiviral were excluded. The estimated Markov probabilities of HBeAg loss and seroconversion were calculated using a two-state Markov illness–death model and plotted. The Markov model allows transitions between response and nonresponse during the time of interest, whereas the Kaplan-Meier method only allows the transition from nonresponse to response and ignores the transition from response to nonresponse. The probability estimated by a two-state Markov illness–death model can be interpreted as the response at a given time point, whereas the probability from the Kaplan-Meier method is interpreted as the response probability by a given time point.

A Wilcoxon signed-rank test was used to test for significant changes in biopsy results between visits for the 15 LTSES patients who had a baseline and LTSES biopsy result.

Safety analyses were generally assessed over the duration of the study in the 65 LTSES patients. Renal safety was evaluated in the 309 patients who were on ADV 10 mg in year 1 or who took placebo in year 1 followed by ADV 10 mg up to the point of discovery of misallocation of drug in year 2; graded laboratory toxicities (including serum creatinine and phosphorus) were evaluated in the 65 LTSES patients. On-treatment ALT flares were defined as an ALT value greater than 10 times the ULN value and greater than twice the patient's baseline value at the start of ADV treatment.


Figure 1 shows patient disposition for the ADV10-Y1 efficacy population and the LTSES safety subset from enrollment through week 240. Of the 171 patients originally randomized to ADV 10 mg in year 1, 42 (25%) completed the 240-week study period; and an additional 27 of 171 (16%) were terminated early per protocol for durable HBeAg loss or seroconversion. The most common reason for not completing the full 240 weeks was that patients chose to not continue the study after the end of the blinded period when the misallocation of dosing error was discovered (58/171; 34%). Demographics and baseline disease characteristics for the 171 patients included in the ADV10-Y1 efficacy population and the 65 patients from the ADV10-Y1 efficacy population who continued in the LTSES are presented in Tables 2 and 3.

Figure 1.

Patient Disposition Study GS-98–437: ADV 10 mg dose group.

Table 2. Demographics at Study Baseline
CharacteristicADV10-Y1 Efficacy Analysis Set (N = 171)LTSES Safety Analysis Set (N = 65)
Sex: n (%)  
 Male130 (76%)54 (83%)
 Female41 (24%)11 (17%)
Age (years)  
 Mean ± SD34 ± 11.236 ± 11.3
 Min, Max16 to 6518 to 63
Ethnicity (n, %)  
 White60 (35%)15 (23%)
 Black8 (5%)2 (3%)
 Asian102 (60%)48 (74%)
 Other1 (<1%)0
Weight (kg)  
 Mean ± SD72 ± 15.973 ± 15.1
 Min, Max43–11843–115
Table 3. Disease Characteristics at Study Baseline
CharacteristicADV10-Y1 Efficacy Analysis Set (N = 171)LTSES Safety Analysis Set (N = 65)
  1. ULN, upper limit of normal range.

HBV DNA (log10copies/mL)  
 Mean ± SD8.25 ± 0.908.26 ± 0.81
 Min, max5.24–10.166.13–9.61
ALT (IU/L)  
 Mean ± SD139 ± 153.5130 ± 162.5
 Min, max20–128646–1270
Baseline ALT as Multiples of ULN  
 Mean ± SD3.4 ± 4.03.1 ± 3.8
 Min, max0.6–37.81.1–29.5

Efficacy data are reported for all 171 patients on ADV 10 mg in year 1, whereas safety data are generally presented for those patients who took at least one dose of study drug in the LTSES phase of the study (a subset of the 171 patients in ADV 10 mg in year 1).

Virological and Biochemical Response.

Median serum HBV DNA concentrations decreased rapidly in the first 12 weeks of the study and continued to decline gradually until week 48, the end of correctly allocated blinded treatment (Fig. 2). At week 48, the median change from baseline in HBV DNA was −3.44 log10 copies/mL. After week 48, patients who received ADV 10 mg were randomized (1:1) to either placebo or ADV (in other words, by design half the patients came off treatment during the second 48 weeks of the study), and thus viral rebound would be expected in those patients switching from ADV to placebo. However, the misallocation of dosing created a situation in which a patient who was randomized to either placebo or ADV at the end of 48 weeks actually received monthly pill allocations that were incorrect approximately 50% of the time during the second year, until the dosing error was detected. At study week 96, median serum HBV DNA change from baseline was −2.15 log10copies/mL, reflecting inconsistent antiviral effect, whereas patients received misallocated dose regimens or were off-treatment pending Institutional Review Boards and Independent Ethics Committees approval of the open label or LTSES study protocols (when ADV 10 mg was taken by all patients). At study weeks 144, 192, and 240 (all in the LTSES), the median changes in HBV DNA from baseline were −3.69, −3.55, and −4.05 log10copies/mL, respectively; 36%, 38%, and 39% had serum HBV DNA 1000 copies/mL at study weeks 144, 192, and 240.

Figure 2.

Median serum HBV DNA concentration by study visit.

Similar to the pattern of HBV DNA concentrations, median serum ALT concentrations decreased rapidly in the first year of the study (Fig. 3). Between weeks 48 and 108, while patients were receiving misallocated study medication or were off treatment waiting for Institutional Review Board/ethics committee approval of the open-label or LTSES periods, median serum ALT concentrations increased, then decreased again after week 108 and remained decreased through week 240. Median change values from baseline in serum ALT concentrations were −43, −18, −49.5, −41, and −50 IU/L at study weeks 48, 96, 144, 192, and 240, respectively; 66% had normalized (ULN) serum ALT concentrations at study week 240.

Figure 3.

Median serum ALT concentration by study visit.

Serological Response.

During the course of the study (weeks 0–240), 58% of patients (99/171) had confirmed (two consecutive or last observed value) HBeAg loss, and 48% (82/171) had confirmed HBeAg seroconversion. Figure 4 presents a graph of the estimated Markov probabilities of HBeAg seroconversion over the course of the study. The proportion of patients with e-seroconversion increased steadily from baseline until approximately week 120, then leveled out subsequently (patients meeting this endpoint terminated from study and enrolled into an off-treatment follow-up protocol). The estimated Markov probability values of HBeAg seroconversion were 13%, 29%, 37%, 35%, and 30% for weeks 48, 96, 144, 192, and 240, respectively).

Figure 4.

Plot of probability of HBeAg seroconversion by two-state Markov model (ADV10-Y1 analysis set).

Forty-five patients with confirmed HBeAg seroconversion enrolled in a long-term follow-up study to assess the durability of seroconversion after stopping ADV [median off-treatment follow-up time was 150 weeks (range, 13–252)].8 Of these 45 patients, 41 (91%) maintained durable seroconversion at the last two assessments. Four patients (9%) seroreverted at follow-up week 12 (n = 3) and week 16 (n = 1).

Four of the 171 patients (2%) had both confirmed hepatitis B surface antigen loss and hepatitis B surface antigen seroconversion.

Histological Response.

Fifteen patients (9%) had assessable liver biopsies before first dose of ADV and at the end of the L LTSES; a significant histological improvement was observed in this subset of patients.

Median changes in scores from baseline were 4 for the Knodell inflammation scores (scale 0–18) (P = 0.014) and −1 for the Ishak fibrosis scores (scale 0–6) (P = 0.022). Histological improvement (defined as a reduction of at least 2 points in Knodell necroinflammatory score with no worsening of Knodell fibrosis score) was seen in 60% of patients (9/15). Fifty-three percent of patients (8/15) improved from baseline to week 48, and 53% of patients (8/15) improved from week 48 to end of follow-up; four patients improved in both periods.

In the ranked assessment scores from paired biopsies, 67% of patients (10/15) had improvement in necroinflammation, and 60% (9/15) showed improvement in fibrosis from baseline to the end of follow-up. Two patients had worsened necroinflammation, and one patient had worsened necroinflammation and fibrosis scores; none of these three patients had evidence of adefovir-associated resistance.

Six of 15 (40%) subjects had undetectable HBV DNA, and 5 of 15 (33%) had ALT normalization, respectively. Two subjects had confirmed seroconversion at the time of biopsy, four had HBeAg loss without confirmation of anti-HBe positivity, and nine remained HBeAg positive. All but one subject had their final biopsy taken after week 240 (one subject had the final biopsy taken at week 179). Five of the subjects in this analysis set developed adefovir resistance during the LTSES. Three of these subjects had more improvement in the Knodell necroinflammatory score from week 48 to LTSES than was average for the analysis set, one had average improvement, and one had an increase in the score.

Viral Resistance Analysis.

In the 65 LTSES patients with a median (range) of 235 (110–279) weeks of ADV exposure at the time of the last genotyping sample on ADV monotherapy, 13 (20%) had developed adefovir-associated resistance mutations, rtN236T or rtA181V. The first resistance mutation was observed at week 195 on study (after 135 weeks of ADV). Twelve of the 13 patients with these resistance mutations also had viral load rebound (confirmed increase of at least 1 log10 copies/mL from the nadir value), and two also had increases in ALT to greater than 10 times the ULN. The 95% confidence interval around the proportion of LTSES patients developing genotypical resistance (20%) is 10% to 30%, which is in the range of rates reported in other ADV studies.

In addition, the single rtA181T mutation was observed in three patients (5%). However, there was no clear association of the emergence of the rtA181T mutation with serum HBV DNA rebound. After the emergence of the rtA181T mutation, the serum HBV DNA remained unchanged in one patient, there was a decline (confirmed 1 log10) in the serum HBV DNA in one patient, and there was a transient rebound (one value 1 log10 from nadir) in one patient. In vitro phenotypical testing on two patient-derived HBV isolates demonstrated that this mutation alone resulted in only a 1.3-fold to 1.9-fold decrease in susceptibility to ADV.


Treatment with ADV 10 mg for up to 240 weeks was generally well tolerated in this study; the median weeks on study drug (including misallocation of dosing, placebo, and off-treatment periods) was 289.9 (range, 152.1–331.9) for LTSES patients.

Sixteen patients (25%) reported adverse events (AEs) that were graded as severe. The most commonly reported severe AE was ALT increased (9%); severe AEs occurring in two patients (3%) each were abdominal pain, asthenia, fever, lack of drug effect, liver damage, nausea, aspartate aminotransferase (AST) increased, and pharyngitis. During the 5-year period, no single severe AE occurred with an incidence of more than 3% except ALT increased (9% overall, 5% excluding misallocation of dosing, placebo, and off-treatment data). Of the severe events reported by at least 3% of patients in the 5-year analysis in LTSES patients, ALT increase was seen by week 48 in four patients (2%), AST increase and pharyngitis were seen in two patients (1%) each, and fever was seen in a single (<1%) patient. The increase in the percentage of patients reporting a severe AE over the 5 years compared with the week 48 results presented in the interim report was most likely attributable to the longer duration of follow-up.

A total of 65% (42/65) of patients in the LTSES analysis set reported AEs that were considered to be possibly or probably treatment-related (Table 4). The most commonly reported treatment-related AEs were lack of drug effect, asthenia, and ALT increased (each in 18% of patients); headache (14%); and abdominal pain (11%). When treatment-related AEs reported during the off-treatment periods (7 days), placebo, and misallocation of dosing periods were excluded, the numbers of patients reporting treatment-related AEs were generally the same, except for ALT increased (6 versus 12 patients) and AST increased (6 versus two patients).

Table 4. Treatment-Related Adverse Events Reported in at Least 5% of LTSES Patients
No. (%) of Patients Reporting Treatment-Related Adverse Event* by Preferred TermLTSES Analysis Set All Data (N = 65)LTSES Safety Subset Excluding Misallocation, Placebo, Off-treatment Datab (N = 65)
  • *

    Each patient is counted only once per preferred term. If a patient discontinued study medication prematurely, data collected to 30 days post-discontinuation were included.

  • †These data excluded AEs that occurred during misallocation of dosing, placebo, and off-treatment periods of at least 7 days.

Any treatment-related adverse event42 (65%)38 (58%)
Body as a whole  
Lack of drug effect12 (18%)12 (18%)
Asthenia12 (18%)10 (15%)
Headache9 (14%)8 (12%)
Abdominal pain7 (11%)5 (8%)
Pain3 (5%)3 (5%)
Anorexia4 (6%)2 (3%)
Diarrhea4 (6%)3 (5%)
Dyspepsia4 (6%)3 (5%)
Nausea4 (6%)3 (5%)
Flatulence3 (5%)3 (5%)
Metabolic and nutritional  
ALT increased12 (18%)6 (9%)
AST increased6 (9%)2 (3%)
Liver function tests abnormal4 (6%)2 (3%)
Creatinine increased5 (8%)5 (8%)
Albuminuria3 (5%)2 (3%)

Overall, the profile of treatment-related AEs in the first 48 weeks4 was similar to that over the 5 years of the study. More patients reported lack of drug effect over the 5-year period. In addition, treatment-related ALT and aspartate aminotransferase increases were reported by relatively higher proportions of patients in the 5-year period (ALT: 4% by week 48 and 18% by the end of the study; AST: 1% by week 48 and 9% by the end of the study). However, in half of these cases, the adverse events occurred only during the misallocation of dosing, placebo, and off-treatment periods.

No patient treated with ADV 10 mg died during the study. Two LTSES patients became pregnant and ended their pregnancies with a therapeutic abortion. Nine patients (14%) experienced one or more serious adverse event. No individual serious adverse event was seen in more than one patient. None of these events was considered to be related to study treatment by the investigator, and ADV was not permanently discontinued in any of these patients as a result. Five patients had permanent discontinuation of study medication because of AEs in the LTSES period: virological failure with resistance in three patients and increase in creatinine concentration in two patients.

Grade 3 or 4 abnormalities were seen in several of the clinical laboratory analytes in at least two patients (3%) of the LTSES analysis set (Table 5). When data from the periods when patients were not on active treatment were excluded, the percentages of patients with grade 3 or 4 laboratory abnormalities decreased substantially for ALT, AST, and creatine kinase. Elevations in serum ALT were the most common. Two patients had grade 3 or 4 abnormalities in phosphorus concentration.

Table 5. Grade 3 or 4 Laboratory Abnormalities Reported in at Least 3% of Patients (LTSES Analysis Set)
Analyte With Grade 3 or 4 AbnormalityAll LTSES Safety Subset Data (N = 65)Excluding Misallocation, Placebo, Off-Treatment Data* (N = 65)
  • NA, not available.

  • *

    These data excluded laboratory data collected during misallocation of dosing, placebo, and off-treatment periods of at least 7 days.

ALT50 (77%)31 (48%)
AST33 (51%)17 (26%)
GGT2 (3%)2 (3%)
Serum amylase5 (8%)5 (8%)
Creatine kinase14 (22%)8 (12%)
Hyponatremia2 (3%)2 (3%)
Phosphorus2 (3%)2 (3%)
Hematuria12 (18%)11 (17%)

Five patients had grade 3 serum amylase abnormalities. In only one patient, the event was recorded as an AE (mild in severity). No patient permanently discontinued study medication because of this abnormality. This abnormality occurred in year 1 of the study in four patients, in the OL period in two patients, and in the LTSES in one patient. All grade 3 serum amylase abnormalities resolved within three visits.

Renal Safety.

Patients in the ADV-ADV and placebo (PLA)-ADV groups (n = 309 patients) during the double-blind period were screened for renal events. Six patients (all from the LTSES safety subset) had confirmed (two consecutive values) increases of 0.5 mg/dL from baseline in serum creatinine (the first event occurred after 170 weeks on study). Two of the six subjects permanently discontinued study medication because of the AE of confirmed creatinine elevation. The remaining four subjects completed the study without dose reduction or interruption. Two patients had confirmed phosphorus values less than 2.0 mg/dL that occurred during or immediately after off-treatment periods. No grade 3 or grade 4 abnormalities (>3 mg/dL) in serum creatinine were seen in the LTSES patients. Two LTSES patients had Grade 3 abnormalities in serum phosphorus (1.0 to 1.5 mg/dL); none had a Grade 4 abnormality.

Hepatic Safety.

Abnormalities in liver tests were commonly seen in conjunction with the misallocation of dosing and off-treatment periods and when study treatment was intermittent or withdrawn. On-treatment ALT flares, defined as an ALT value greater than 10 times the ULN value and greater than twice the patient's baseline value, were reported in 29 patients; 15 patients when misallocation of dosing and off-treatment data were omitted. In eight of the 15 patients, the ALT flare occurred immediately after the misallocation of dosing or off-treatment period. Three patients developed adefovir-associated resistance mutations around the same time as an ALT flare. Posttreatment ALT flares occurred in three LTSES patients at 60, 91, and 113 days after stopping ADV at the end of the LTSES phase of the study.


In this study we have shown that treatment with ADV 10 mg daily for up to 240 weeks in HBeAg-positive CHB patients achieved important virological, biochemical, and histological endpoints including effective suppression of HBV DNA replication, normalization of serum ALT concentrations, and HBeAg seroconversion. Antiviral efficacy results in patients who participated in the LTSES period are arguably conservative estimates of the long-term effect of ADV in HBeAg+ patients with compensated liver function because patients who had durable HBeAg loss or seroconversion at the end of the open-label phase or during the LTSES phase were discontinued from the study and not permitted to continue in the LTSES phase. Most of the patients in this study were Asian and therefore likely to have acquired HBV infection early in life. Recently concerns have been raised regarding the long-term benefit of HBeAg seroconversion for such patients.9, 10 Although some observational studies suggest that most Asian patients experience some clinical benefit after HBeAg seroconversion,11, 12 it is an incomplete marker of immune control. HBeAg seroconversion associated with incomplete viral suppression may result in the emergence of precore mutant hepatitis B and attendant chronic sequelae. In regard to liver histology, although only a small number of patients had paired liver biopsy specimens available for evaluation, significant improvement in hepatic necroinflammation and fibrosis was observed in this group at the end of the study compared with pretreatment biopsies. In terms of the representativeness of the sample of 15 subjects with assessable liver biopsies, this subgroup was found to have slightly poorer efficacy results (proportions of subjects with HBV DNA <1000 copies/mL and normalized ALT values) than did the rest of the study sample.

Recent natural history studies have shown an association between active viral replication and adverse clinical outcomes such as cirrhosis and hepatocellular carcinoma,13–15 supporting that lifelong viral suppression will be necessary for most CHB patients.9 In this context, we have shown that treatment with ADV 10 mg for up to 240 weeks was well tolerated. There were no ADV-related serious adverse events. Liver test abnormalities were relatively frequent during the periods of drug misallocation and drug withdrawal but not when the events occurring in these off-treatment periods were excluded from analysis. Most patients showed ALT normalization during the LTSES.

Serum creatinine elevations occurred in six patients during long-term ADV therapy; however, only two patients required permanent drug discontinuation. Our data suggest that significant renal impairment is uncommon during long-term ADV therapy; however, monitoring of renal function during treatment remains necessary.

The development of antiviral resistance negates treatment efficacy compared with patients in whom HBV replication remains suppressed.16 Signature adefovir resistance mutations developed in 20% of treated patients during the LTSES; however, as seen in previous studies,3, 4 the development of resistance was delayed and occurred in a lower proportion of patients in comparison with that seen with lamivudine and telbivudine over 2 years.17 Virological breakthrough developed in most patients with genotypic resistance.

The main limitation of our study arises from the drug misallocation that occurred during the second year of the study, which necessitated significant protocol revisions and led to significant attrition. This resulted in long-term evaluation of a small number of patients. The full effect of the drug misallocation period and the off-treatment period(s) on the efficacy, safety, and resistance development of ADV 10 mg is impossible to assess. However, in those patients who remained in the study and restarted ADV 10 mg treatment in the LTSES period, the drug was well tolerated, and continued improvement in efficacy was observed, as has been previously shown for patients with HBeAg-negative chronic HBV infection.3