MEMORANDUM
Date:
To: Antiviral Drug Products Advisory Committee Members and Guests
From: Entecavir Review Team
Through: Debra Birnkrant, M.D.
Director, Division of Antiviral Drug Products
Mark Goldberger, M.D., M.P.H.
Office of Drug Evaluation IV
Subject: Briefing document for NDA 21-797, entecavir 0.5 and 1 mg tablets and NDA 21-798, entecavir oral solution 0.05 mg/mL
1.
Executive
Summary: Regulatory Issues and Purpose
of Meeting
This briefing document provides background information and
the FDA perspective on the New Drug Applications (NDAs 21-797 and 21-798)
submitted by Bristol-Myers Squibb (BMS) for entecavir (ETV), a nucleoside
analogue intended for the treatment of chronic hepatitis B virus (HBV)
infection in adults. The information
presented in this document represents the preliminary findings and opinions of
the primary reviewers from each discipline based on their review of the
submitted material. The material
included in this briefing document and other material presented by the
applicant will be the subject of a meeting of the Antiviral Drug Products
Advisory Committee to be held on
Entecavir is a guanosine nucleoside analogue with selective activity against HBV; it has no activity against other hepatitis viruses or HIV. The applicant’s clinical development program was designed to assess the safety and efficacy of ETV in a variety of distinct patient populations recruited from a global network of investigators. The Advisory Committee will be asked to review and discuss issues related to the strength and completeness of the clinical database, the risk/benefit assessment of the drug in the context of non-clinical data, and the appropriateness for further post-marketing development and pharmacovigilance.
2.
Summary of
Clinical Development
The applicant has conducted an extensive global development program for ETV in the treatment of chronic HBV. The key clinical studies are summarized in Table 1. At the time these studies were initiated, lamivudine (LVD) was the only approved oral treatment for chronic HBV and was chosen as the appropriate comparator for the Phase 3 studies. The preliminary opinions and discussion to follow are based on review of the 4 designated pivotal trials unless otherwise stated.
Table 1: Summary of Phase 3 and Key Phase 2 Clinical
Trials of Entecavir
|
Study/Sites |
Patient Population |
Number of Patients Treated on Study |
Dose/Duration of Treatment |
Primary Efficacy Endpoint |
|
Pivotal Clinical Trials |
||||
|
AI463022 |
HBeAg positive,
nucleoside naïve, ALT > 1.3 x
ULN |
709 |
ETV 0.5 mg QD or LVD 100 mg QD for 52 weeks, up to 96
weeks total for partial (virologic) responders |
Liver histology at 48 weeks of treatment |
|
AI463027 |
HBeAg negative,
nucleoside naïve, ALT > 1.3 x
ULN |
638 |
ETV 0.5 mg QD or LVD 100 mg QD for 52 weeks, up to 96
weeks total for partial (virologic) responders |
Liver histology at 48 weeks of treatment |
|
AI463014 |
LVD-refractory, HBeAg positive or negative |
181 (87 received either ETV 1 mg or LVD 100 mg) |
ETV 0.1, 0.5, 1.0 mg QD or LVD 100 mg QD for up to 76
weeks, some low-dose patients received ETV 1.0 mg open label after Week 28 |
HBV DNA by bDNA at 24 weeks of treatment |
|
AI463026 |
LVD-refractory, HBeAg positive, ALT > 1.3 x ULN |
286 |
ETV 1 mg QD or continued LVD 100 mg QD for 48 weeks, up to
96 weeks total for partial (virologic) responders |
Liver histology at 48 weeks of treatment |
|
Study/Sites |
Patient Population |
Number of Patients Treated on Study |
Dose/Duration of Treatment |
Primary Efficacy Endpoint |
|
Supportive Clinical Trials |
||||
|
AI463004 Worldwide |
Nucleoside naïve and IFN/LVD-refractory, HBeAg positive or negative |
42 |
ETV 0.05, 0.1, 0.5, or 1 mg QD or placebo for 28 days |
HBV DNA by bDNA and PCR assays |
|
AI463005 Worldwide |
Nucleoside naïve, HBeAg positive or negative |
177 |
ETV 0.01, 0.1, or 0.5 mg QD or LVD 100 mg QD for 24 weeks,
up to 48 weeks in partial responders |
HBV DNA by bDNA and PCR assays |
|
AI463007 Worldwide |
Rollover study for patients completing AI463004 |
28 |
Open label ETV 0.1 mg QD for 24 weeks |
HBV DNA by bDNA and PCR assays |
|
AI463015 Worldwide |
Liver transplant patients with HBV reinfection despite LVD
or HBIG |
9 |
ETV 1 mg for 48 weeks, 48 week extension |
HBV DNA by bDNA and PCR assays |
|
AI463038 Worldwide |
HIV/HBV coinfected patients, LVD-refractory |
68 |
ETV 1 mg or placebo added to LVD-containing HAART regimen
for 24 weeks, open label ETV 1 mg for additional 24 weeks |
HBV DNA by bDNA and PCR assays, HIV PCR |
|
AI463901 Worldwide |
Rollover study for patients who have failed monotherapy in
Phase 2 or 3 study, HBeAg positive or negative |
969 (currently still enrolling) |
ETV 1 mg QD plus LVD for 52 weeks, up to 144 weeks for
partial responders |
HBV DNA by bDNA and PCR assays |
Data were submitted from 19 pharmacokinetic studies that provided a good understanding of ETV’s absorption, distribution, metabolism, excretion, and interactions with other commonly used drugs. These studies are summarized in Appendix A.
3.
Summary of Non-clinical
Toxicology and Carcinogenicity
3.1.
Overview
of General Non-clinical Findings
Entecavir is efficiently phosphorylated to ETV-triphosphate (TP) by cellular nucleoside kinases. By competing directly with the natural deoxyguanosine triphosphate (dGTP), ETV-TP potently inhibits each of the 3 distinct activities of the HBV viral polymerase: priming, reverse transcription of first-strand DNA synthesis, and the DNA-dependent DNA polymerase activity responsible for second-strand DNA synthesis. Entecavir has little activity against the DNA polymerase of mitochondria.
The pharmacokinetic (PK) characteristics of entecavir in mice, rats, rabbits, dogs, and monkeys are comparable to those in humans indicating the acceptability of these species for the toxicological assessment of ETV.
Species-specific, reversible CNS inflammation was seen in dogs administered doses that achieve ~51 times the exposure to ETV in humans at clinically proposed doses. It was concluded that this is not relevant to human safety. Other target organs in repeat-dose studies in animals were the kidneys, liver, lungs, skeletal muscle and testis. Data from a 1-year study in monkeys indicated that there was no target organ toxicity in monkeys at exposures to ETV ~136 times those in humans.
Long-term
dosing of ETV was evaluated in a woodchuck model of chronic HBV. In this study, woodchucks received a daily
dose of ETV equivalent to the 1 mg human dose for 2 months and then were
maintained with weekly dosing for up to 3 years. Viral suppression was maintained through 3
years of treatment with no evidence of emergence of resistant HBV. The applicant reported survival rates of 40%
and 80% for animals treated for 14 and 36 months, respectively, compared to a
survival of 4% in historical controls.
Of most interest, the occurrence of hepatocellular carcinoma was
significantly reduced in the animals treated long-term compared to historical
control animals.
3.2.
Overview of Carcinogenicity Studies
In a battery of genetic toxicology
studies, ETV was an in vitro mutagen
in mouse lymphocytes and clastogenic in
vitro in human lymphocytes (without metabolic activation). However, ETV
was negative in an
Carcinogenicity studies in Sprague Dawley rats and CD-1 mice were
conducted. Increased incidences of tumors were observed in both the studies. The results of these studies were presented to
the Executive Carcinogenicity Assessment Committee (ECAC) on
Rat Carcinogenicity Study: The
oncogenicity potential of ETV was investigated in male rats at oral gavage
dosages of 0.003 (low), 0.02 (mid), 0.2 (high) or 1.4 mg/kg/day (highest) and
in females at dose levels of 0.01 (low), 0.06 (mid), 0.4 (high) or 2.6
mg/kg/day (highest) in comparison with untreated controls for a period of 104
weeks.
The no observed effect level (NOEL) for
neoplasia was 0.2 mg/kg/day for males and 0.06 mg/kg/day for females. At
tumorigenic doses, systemic exposures were 35- and 4-times that in humans (1 mg
daily dose) in male and female rats, respectively.
Treatment-Associated
Tumors:
1.
Hepatocellular adenomas in female rats were significant (p=0.005) at the
highest dose level. Combined adenomas and carcinomas in the female rats were
also significant (p=0.005) at the highest dose. In female rats, the combined
incidence of adenomas and carcinomas was 1% (controls), 4% (low), 5% (mid), 2%
(high) and 18% (highest).
2. Brain
gliomas were significant (p=0.025) at the highest dose in both male and female
rats. In male rats, the incidence was 0% (controls), 2% (low), 2% (mid), 3%
(high) and 7% (highest). In female rats, the incidence was 0% (controls), 0%
(low), 2% (mid), 0% (high) and 5% (highest).
3. The skin
fibromas in female rats were significant (p=0.025) at the high and highest
doses. In female rats, the incidence was 0% (controls), 0% (low), 2% (mid), 3%
(high) and 5% (highest).
Mouse Carcinogenicity Study: The
oncogenicity potential of ETV was investigated in mice at oral gavage dosages
of 0.004 (low), 0.04 (mid), 0.4 (high) or 4.0 mg/kg/day (highest) in comparison
with untreated controls for a period of 104 weeks.
The NOEL for
neoplasia was 0.004 mg/kg/day for males, based on pulmonary adenomas; for all
other tumors in males and females, the NOEL was 0.4 mg/kg/day. At the
tumorigenic dose in male mice, systemic exposure was 3-times that in humans (1
mg daily dose).
Treatment-Associated Tumors:
1. Lung
adenomas were significant (p=0.005) in male mice (mid, high and highest) and in
the female mice at the highest dose (p=0.005); lung carcinomas in both male and
female mice were significant (p=0.005) at the highest dose. Combined lung
adenomas and carcinomas were significant (p=0.005) in male mice at the mid,
high and highest dose levels and in the female at the highest dose level
(p=0.005. In male mice, the combined incidence of adenomas and carcinomas was
12% (controls), 20% (low), 26% (mid), 40% (high) and 58% (highest). In female
mice, the combined incidence of adenomas and carcinomas was 20% (controls), 13%
(low), 10% (mid), 35% (high) and 52% (highest).
2.
Hepatocellular carcinomas in male mice were significant (p=0.005) at the
highest dose level. Combined liver adenomas and carcinomas were also
significant (p=0.005) at the highest dose level in the male mice. In male mice,
the combined incidence of adenomas and carcinomas was 11% (controls), 9% (low),
8% (mid), 16% (high) and 25% (highest).
3. Vascular
tumors in female mice (hemangiomas of ovaries and uterus and
hemangiomas/hemangiosarcomas of spleen) were significant (p=0.005) at the
highest dose level. In female mice, the incidence of vascular tumors was 16%
(controls), 23% (low), 29% (mid), 26% (high) and 64% (highest).
The ECAC found that the carcinogenicity studies in mice and rats were
adequately designed and conducted. The
committee judged the results of ETV carcinogenicity studies. They concluded that ETV was a carcinogen in
rodents. The committee concluded that ETV produced tumors in both species and
both genders, and these results suggest a potential cancer hazard to patients.
At the request of the sponsor, the results of the carcinogenicity studies
were presented to the full FDA CAC (CAC), a committee that has been designated
as the arbiter of disputes between applicants and review divisions regarding
the relevance of results in carcinogenicity studies. The CAC met with the applicant and the Review
Team on
4.
Summary of
Efficacy Data
4.1.
Dose
Selection
The proposed dose of ETV was selected on the basis of reductions in HBV DNA and safety and tolerability of the drug observed during short and long-term Phase 2 dose-ranging studies. In Study 004, reduction in HBV DNA over a 28-day dosing period and 28-day follow-up was greater in treatment groups receiving 0.5 and 1 mg daily than those receiving lower doses. Similarly, in Study 005, the dose of 0.5 mg resulted in greater decreases in HBV DNA over 22 weeks of dosing compared to either ETV 0.1 mg or LVD 100 mg. In these early Phase 2 studies, safety and tolerability of ETV 0.5 mg appeared to be somewhat better than ETV 1 mg, so the 0.5 mg dose was carried forward into Phase 3 trials for nucleoside-naïve patients (Studies 022 and 027).
The applicant anticipated that LVD-refractory patients would require a higher dose based on in vitro data demonstrating that LVD-resistant HBV also had reduced sensitivity to ETV. Dose selection for this population was based on the interim, 24-week results from Study 014. In this study, a dose-response relationship for ETV was observed in HBV DNA reduction and the dose of ETV 1 mg was superior to 0.5 mg or to LVD 100 mg for the proportion of patients achieving HBV DNA < LLOQ by bDNA. In this study, there were no observed differences in safety and tolerability across the treatment arms. Consequently, ETV 1 mg was carried forward into the Phase 3 study in LVD-refractory patients (Study 026).
Pharmacokinetic and ADME studies
of ETV document that the drug is predominately eliminated by the kidneys. Pharmacokinetics in patients with renal
insufficiency, including those requiring hemodialysis or continuous ambulatory
peritoneal dialysis, were studied in Study AI363011. Based on this study and pharmacokinetic
modeling, the sponsor has proposed dose adjustments for patients with moderate
or severe renal impairment. Elderly
patients demonstrated increased ETV exposure but this could be attributed to
decreased renal function and did not warrant a change in dosing based on age
alone.
4.2.
Summary of Study
Designs
Study 014 was a Phase 2, randomized, double-blind, pilot study to evaluate 3 doses of ETV (0.1 mg, 0.5 mg and 1.0 mg) compared to LVD 100 mg in patients with HBV viremia while receiving LVD (“LVD-refractory”). Patients could be HBeAg-positive or negative. Patients were randomized to either continue LVD or receive ETV. Management decisions to discontinue or continue study treatment were made at Week 28 on the basis of the Week 24 assessments. Patients who achieved > 1 log reduction in HBV DNA continued blinded therapy to Week 52. Patients failing to achieve at least 1 log reduction in HBV DNA discontinued study treatment but were eligible for the rollover study (AI463901) or other available therapy. Patients remaining on study treatment were reassessed at Week 48 and management decisions were based on criteria similar to those used in other studies. Patients could continue blinded therapy for up to 76 weeks. Liver biopsy for histology was optional in this study. The primary endpoint was the proportion of patients who achieved HBV DNA levels below the LOQ by the bDNA assay (0.7 MEq/mL) at 24 weeks. This study was intended to be a dose selection study for Study 026. Only the 87 patients receiving ETV 1 mg or LVD 100 mg QD were included in the efficacy analyses and the major safety assessments presented here. The patients receiving doses less than 1 mg daily were included in only selected analyses.
The 3 pivotal Phase 3 studies utilized similar study design and endpoint analyses. Study 022 was a randomized, double-blind comparison of ETV 0.5 mg versus LVD 100 mg in HBeAg-positive, nucleoside-naïve patients. Other key inclusion criteria included: age > 16 years, ALT > 1.3 x ULN, normal renal function, and compensated liver disease. Previous treatment with IFN was not an exclusion. Continuation of blinded study treatment at the end of 52 weeks was based on results of the Week 48 evaluation. Complete Responders (HBV DNA by bDNA assay < 0.7 MEq/mL and loss of HBeAg) stopped study treatment and were followed for 24 weeks off therapy to assess durability of response. Partial Responders (HBV DNA by bDNA assay < 0.7 MEq/mL but still positive for HBeAg) continued blinded therapy for up to 96 weeks or until complete response was achieved. Non-responders (HBV DNA by bDNA > 0.7 MEq/mL) discontinued study treatment but were eligible for the rollover study or other available therapy. The primary endpoint was histologic improvement on liver biopsy at 48 weeks defined as > 2 point decrease in Knodell necroinflammatory score with no worsening in fibrosis score. Secondary endpoints included improvement in Ishak fibrosis score, change in HBV DNA by bDNA assay and by PCR assay, normalization of ALT, HBeAg loss, HBeAg/HBeAb seroconversion, and various composite endpoints.
Study 027 was a randomized, double-blind comparison of ETV 0.5 mg versus LVD 100 mg in HBeAg-negative, nucleoside-naïve patients. Study design was similar to Study 022 with similar management decisions occurring at Week 52. Inclusion criteria were similar to those in Study 022 except for HBeAg status. In the study treatment management algorithm, patients achieving < 0.7 MEq/mL HBV DNA by bDNA assay and ALT < 1.25 x ULN at 48 weeks were considered to have reached the composite efficacy endpoint (Composite Responders) and were eligible to discontinue study treatment and enter the follow-up phase. The primary efficacy endpoint was histologic improvement at 48 weeks defined as > 2 point decrease in Knodell necroinflammatory score with no worsening in fibrosis score. Secondary endpoints were similar except that HBeAg loss and HBeAb seroconversion were not evaluated.
Study 026 was a randomized, double-blind comparison of ETV 1 mg versus LVD 100 mg in HBeAg-positive, LVD-refractory patients with compensated liver function. Patients were randomized to either continue LVD or receive ETV. Continuation of treatment at the end of 52 weeks was based on results of Week 48 evaluation. Criteria for continuation of blinded therapy through 96 weeks were based on Complete, Partial, or Non-response and were similar to Study 022. For this study there were co-primary endpoints at Week 48: histologic improvement on liver biopsy similar to other Phase 3 studies and the proportion of patients with both undetectable HBV DNA by bDNA assay (< 0.7 MEq/mL) and normalization of ALT (defined as < 1.25 x ULN). Multiple secondary endpoints were evaluated as in the other studies.
All of the Phase 2 and 3 studies used 2 assays to measure changes in HBV DNA: HBV DNA using a branched DNA assay (Chiron/Bayer Quantiplex™ v1.0) with a lower limit of quantitation (LLOQ) of 0.7 MEq/mL and HBV DNA using a PCR assay (Roche COBAS Amplicor HBV Monitor™ v2.0) with an LLOQ of 300 copies/mL. Clinical management decisions were based on real-time reporting of HBV DNA by the bDNA assay results while the PCR assays were conducted at a later time. For study defined endpoint calculations, the applicant used a cut-off HBV DNA value of < 400 copies/mL for the PCR assay. Neither the bDNA assay nor the PCR assay has been approved by the FDA and both are considered investigational. However, there is considerable experience with the use of both assays in clinical trials and clinical practice.
4.3.
Patient
Demographics
Patients who participated in the clinical
trials include a representative sampling of patients with chronic HBV and
compensated liver disease. Study
participants were recruited from 31 countries in
Table 2: Patient Demographics
|
Characteristic |
Nucleoside Naïve Studies |
LVD-refractory Studies |
||
|
Study 022 |
Study 027 |
Study 014 |
Study 026 |
|
|
Mean Age (years) |
35 |
44 |
48 |
39 |
|
Gender Male Female |
75% 25% |
75% 25% |
39% 61% |
74% 26% |
|
Race Asian Caucasian Other** |
57% 40% 3% |
39% 58% 3% |
37%* 62% 1% |
30% 63% 7% |
*In Study 014 “Asian” racial
designation includes all Asian/Pacific Islander.
**Other designation includes:
Black/African American, native Hawaiian/other Pacific Islander, Hispanic/Latino,
Filipino, and others not specified.
Table 3: Baseline HBV Disease
Characteristics
|
Characteristic |
Nucleoside Naïve Studies |
LVD-refractory Studies |
||
|
Study 022 |
Study 027 |
Study 014 |
Study 026 |
|
|
Mean HBV Log bDNA (MEq/mL) Log PCR (copies/mL) |
2.59 9.66 |
1.24 7.58 |
2.45 9.18 |
2.50 9.36 |
|
Mean ALT (U/L) |
143 |
142 |
125 |
128 |
|
Knodell necroinflammatory score |
7.8 |
7.9 |
ND |
6.5 |
|
Knodell fibrosis score |
1.7 |
1.9 |
ND |
1.7 |
|
Ishak fibrosis score |
2.3 |
2.4 |
ND |
2.3 |
|
HB e antigen positive |
98% |
< 1% |
97% |
68% |
|
HB e antibody positive |
3% |
99% |
4% |
28% |
ND, not done
4.4.
Primary
Efficacy Analysis
As noted above, the primary efficacy endpoints for the Phase 3 studies were based on the change in liver histology over the initial 48 week study period when patients received blinded study drug. Histologic improvement was defined as > 2 point decrease in the Knodell necroinflammatory score with no worsening in the fibrosis score. All liver biopsy specimens were evaluated by a single reader who was blinded to treatment group and biopsy order and assigned the histologic scores by both the Knodell and Ishak criteria.
The FDA statistical analysis confirmed the applicant’s analysis of the primary efficacy endpoint. The applicant’s analysis included patients with available baseline biopsy data and counted those with missing or inadequate Week 48 biopsy data as treatment failures. Although the studies were originally designed to show non-inferiority of ETV to LVD, this analysis demonstrated that patients receiving ETV experienced superior overall histologic improvement compared to LVD in all 3 studies (P values < 0.02 in all studies). ETV performed better for each of the individual components of the overall histologic improvement, > 2 point decrease in Knodell necroinflammatory score and no worsening in Knodell fibrosis score.
The FDA statistical reviewer also conducted sensitivity analyses of the primary histologic endpoint that included several methods of imputing missing data. A more conservative analysis included all patients who received study drug and counted patients with missing baseline or Week 48 biopsy as treatment failures. This analysis continued to show superiority of ETV compared to LVD (P values < 0.03 in all studies) in both nucleoside-naïve and LVD refractory patients.
ETV was equivalent (non-inferior) to LVD for the secondary histologic endpoint of improvement in Ishak fibrosis score in the nucleoside-naïve Studies 022 and 027 but was superior in the LVD-refractory patient population in Study 026. Sensitivity analyses were also conducted for the secondary histologic endpoints and supported the conclusion that ETV was no worse than LVD as measured by the Ishak score. Histologic efficacy results are summarized in Table 4 below.
Table 4: Histologic Efficacy Assessments at 48 Weeks
in Studies 022, 027, and 026
|
|
Study 022 |
Study 027 |
Study 026 |
|||
|
ETV 0.5 mg (N=354/314) |
LVD 100mg (N=355/314) |
ETV 0.5 mg (N=325/296) |
LVD 100mg (N=313/287) |
ETV 1 mg (N=141/124) |
LVD 100mg (N=145/116) |
|
|
Knodell scores |
|
|
|
|
|
|
|
Overall histologic improvement* |
72%# |
62% |
70%# |
61% |
55%# |
28% |
|
Fibrosis no worse* |
89%# |
82% |
||||