1. 1 FDA Review Perspective – Entecavir for Hepatitis B Linda L. Lewis, M.D. Medical Officer Division of Antiviral Drug Products

2. 2 Outline of Presentation Overview Overview Efficacy Efficacy Safety Safety Virology/Resistance Virology/Resistance Risk-benefit assessment Risk-benefit assessment Pharmacovigilance plan Pharmacovigilance plan Questions for the committee Questions for the committee

3. 3 Overview

4. 4 Overview Current treatment for chronic HBV Current treatment for chronic HBV –Interferon (IFN) Approved 1992 Approved 1992 Requires parenteral administration, significant side effect profile Requires parenteral administration, significant side effect profile –Lamivudine (LVD) First effective oral therapy, approved 1998 First effective oral therapy, approved 1998 Emergence of resistance limits effectiveness Emergence of resistance limits effectiveness –Adefovir (ADV) Approved 2002 Approved 2002 Known renal toxicity may limit use in some populations Known renal toxicity may limit use in some populations

5. 5 Overview Entecavir (ETV) development program included diverse population Entecavir (ETV) development program included diverse population –Multi-national sites in North and South America, Europe, and Asia (U.S. subjects about 10%) –25% women, mix of Asian/non-Asian (Black/African Americans under- represented - 2%) –Different stages of disease and treatment (insufficient data to review in patients with decompensated liver disease)

6. 6 Overview Key studies – all compare ETV to LVD Key studies – all compare ETV to LVD –022 - Nucleoside-naïve, e antigen positive, 0.5 mg –027 - Nucleoside-naïve, e antigen negative, 0.5 mg –026 - Persistent viremia despite LVD treatment (LVD-refractory), e antigen positive, 1 mg –014 - LVD-refractory, dose-finding Phase 2 study Studies 022, 027, and 026 – primary endpoint histologic improvement in liver biopsy after 48 weeks of treatment Studies 022, 027, and 026 – primary endpoint histologic improvement in liver biopsy after 48 weeks of treatment

7. 7 Overview Supportive studies in special populations Supportive studies in special populations –015 – Post-liver transplant pilot study –038 – HIV/HBV co-infected patients –048 – Patients with decompensated liver disease, compares ETV to ADV, still enrolling No histologic endpoints; used virologic, serologic, and biochemical endpoints No histologic endpoints; used virologic, serologic, and biochemical endpoints

8. 8 Efficacy Review

9. 9 FDA statistical review confirmed BMS’ primary efficacy analysis FDA statistical review confirmed BMS’ primary efficacy analysis Secondary endpoint analyses were in agreement with BMS’ conclusions Secondary endpoint analyses were in agreement with BMS’ conclusions Multiple sensitivity analyses and subgroup analyses were performed that supported the primary analysis Multiple sensitivity analyses and subgroup analyses were performed that supported the primary analysis

10. 10 Histologic Endpoints in Phase 3 ETV Studies – Primary Efficacy Endpoint Study 022Study 027Study 026 ETV 0.5 mg (N=314) LVD 100mg (N=314) ETV 0.5 mg (N=296) LVD 100mg (N=287) ETV 1 mg (N=124) LVD 100mg (N=116) Overall histologic improvement 72%62%70%61%55%28% Fibrosis no worse89%82%84%79%87%70% Necroinflammatory > 2 point decrease 74%64%73%64%55%32% Ishak fibrosis score improvement 39%35%36%38%34%16% Primary analysis: Only patients with evaluable baseline biopsy included, missing/inadequate Week 48 biopsy counted as failures

11. 11 Sensitivity Analysis of Primary Endpoint Study 022Study 027Study 026 ETV 0.5 mg LVD 100 mg ETV 0.5 mg LVD 100 mg ETV 1 mg LVD 100 mg Overall histologic improvement 72%62%70%61%55%28% FDA sensitivity analysis C 77%72%78%70%62%33% FDA sensitivity analysis D 64%55%64%56%48%22% C: Missing/inadequate baseline or Week 48 biopsy excluded D: Includes all treated patients, missing/inadequate Week 48 biopsy counted as failures

12. 12 Selected Secondary Efficacy Endpoints Study 022Study 027Study 026 ETV 0.5 mg LVD 100 mg ETV 0.5 mg LVD 100 mg ETV 1 mg LVD 100 mg HBV DNA PCR < 400 copies/mL 72%42%95%77%22%1% Log HBV DNA by PCR (mean change from baseline) -7.0-5.5-5.2-4.7-5.1-0.5 HBeAg seroconversion 21%18%NA 8%3% ALT normalization (< 1 x ULN) 69%61%78%71%65%17%

13. 13 Efficacy – Subgroup Analysis The treatment effect measured by the primary endpoint was comparable across the following strata: The treatment effect measured by the primary endpoint was comparable across the following strata: –Gender –Race –Age (by quartiles) –Geographic region –HBV subtype –Baseline ALT (by quartiles) –Baseline bDNA or PCR (by quartiles) –Prior LVD or IFN

14. 14 Efficacy – Subgroup Analysis The treatment effect measured as proportion of patients with ALT normalization or HBV DNA by PCR < 400 copies/mL at Weeks 24 or 48 was similar according to: The treatment effect measured as proportion of patients with ALT normalization or HBV DNA by PCR < 400 copies/mL at Weeks 24 or 48 was similar according to: –Gender –Race –Age (quartiles)

15. 15 Subgroup Analyses – Primary endpoint by baseline covariates Favors ETV Favors LVD

16. 16 Safety Review

17. 17 Safety Review FDA clinical review confirmed general safety assessment of ETV FDA clinical review confirmed general safety assessment of ETV –No significant differences in rates or pattern of common AEs or laboratory abnormalities compared to LVD –Rates of SAEs (8% both treatment groups), discontinuations due to AEs (1% ETV, 4% LVD), and deaths (< 1% both groups) were low –ALT flares, CNS adverse events, and malignancies reviewed in detail

18. 18 Safety Review – ALT Flares In nucleoside-naïve subjects, mean ALT values decreased from baseline to Week 48 in both treatment groups In nucleoside-naïve subjects, mean ALT values decreased from baseline to Week 48 in both treatment groups On-treatment ALT flares uncommon - 15/679 (2%) ETV, 27/668 (4%) LVD On-treatment ALT flares uncommon - 15/679 (2%) ETV, 27/668 (4%) LVD Study design allowed only patients who met Response criteria to discontinue treatment and be followed off therapy; more subjects met the criteria in Study 027; analysis of off-treatment flares represents selected subjects Study design allowed only patients who met Response criteria to discontinue treatment and be followed off therapy; more subjects met the criteria in Study 027; analysis of off-treatment flares represents selected subjects Off-treatment ALT flares slightly more common both groups - 15/414 (4%) ETV, 30/377 (8%) LVD Off-treatment ALT flares slightly more common both groups - 15/414 (4%) ETV, 30/377 (8%) LVD

19. 19 Safety Review – ALT Flares In LVD-refractory subjects, on-treatment flares in 4/183 (2%) ETV subjects and 19/190 (10%) LVD subjects In LVD-refractory subjects, on-treatment flares in 4/183 (2%) ETV subjects and 19/190 (10%) LVD subjects Smaller proportion of LVD-refractory subjects met the Response criteria, discontinued therapy, and were followed off-treatment; selected subgroup Smaller proportion of LVD-refractory subjects met the Response criteria, discontinued therapy, and were followed off-treatment; selected subgroup Off-treatment flares occurred in 3/56 (5%) ETV subjects and 0/31 LVD subjects Off-treatment flares occurred in 3/56 (5%) ETV subjects and 0/31 LVD subjects

20. 20 Safety Review – Nervous System Adverse Events in Study 005 Incidence of grouped CNS events increased with increasing doses, trend toward more frequent AEs of dizziness and insomnia with 0.5 mg

21. 21 Safety Review – On-Treatment Nervous System Adverse Events in Pivotal Studies Nucleoside-naiveLVD-Refractory ETV 0.5 mg (N=679) LVD 100 mg (N=668) ETV 1.0 mg (N=183) LVD 100 mg (N=190) Percent with any CNS AE 33%32%36%32% Anxiety 2%<1%3% Dizziness 6% 8%6% Headache 20%19%21%19% Insomnia 4%5% 6% Migraine <1% 2%1% Paresthesia 1% <1%2% Somnolence 1%2% Syncope or Syncope vasovagal <1% 1%0 Percent with Grades 2-4 CNS AEs 9% 15%9%

22. 22 Safety Review – Malignancies Reported in Clinical Trials Malignancies were tracked in all clinical trials Malignancies were tracked in all clinical trials 37 subjects reported malignancies in ETV development program 37 subjects reported malignancies in ETV development program –19/1497 (1.3%) ETV subjects –9/899 (1%) LVD subjects –From special population studies 3/83 receiving ETV alone 3/83 receiving ETV alone 2/28 receiving ADV alone 2/28 receiving ADV alone 4/849 receiving ETV + LVD 4/849 receiving ETV + LVD

23. 23 Safety Review – Malignancies Reported in Clinical Trials Malignancies reported in more than one subject in either treatment group included in the NDA safety database (1497 ETV, 899 LVD): Malignancies reported in more than one subject in either treatment group included in the NDA safety database (1497 ETV, 899 LVD): –HCC (7 ETV subjects, 4 LVD subjects) –Basal cell carcinoma (2 ETV subjects, 1 LVD subject) –Breast cancer (1 ETV subject, 2 LVD subjects including one with carcinoma in situ) –Prostate cancer (3 ETV subjects) Six subjects reported to have malignancies were known to have had previous malignancies Six subjects reported to have malignancies were known to have had previous malignancies

24. 24 Virology/Resistance Review

25. 25 ETV Resistance Profile – Nucleoside- Naïve Subjects (N=434) No genotypic or phenotypic evidence of ETV-resistance detected among 434 nucleoside-naive subjects analyzed at 48 weeks of ETV treatment (Studies 022 and 027) No genotypic or phenotypic evidence of ETV-resistance detected among 434 nucleoside-naive subjects analyzed at 48 weeks of ETV treatment (Studies 022 and 027) –2 subjects in Study 022 experienced confirmed virologic rebound but no resistance mutations identified Follow-up needed after 48 weeks to determine the ETV resistance pathway in naïve subjects Follow-up needed after 48 weeks to determine the ETV resistance pathway in naïve subjects

26. 26 ETV Response in LVD-refractory Subjects (N=189) LVD-refractory subjects less likely than naïve subjects to achieve HBV DNA < 400 copies/mL (21% vs 83% when data pooled) LVD-refractory subjects less likely than naïve subjects to achieve HBV DNA < 400 copies/mL (21% vs 83% when data pooled) > 2 log reductions in viral load and suppression HBV DNA 2 log reductions in viral load and suppression HBV DNA < 400 copies/mL can occur in subjects with LVD-resistant HBV at baseline when treated with 1 mg ETV LVD-resistance substitutions L80V, L180M, M204V or I can emerge in the HBV of subjects on 1 mg ETV by week 48. LVD-resistance substitutions L80V, L180M, M204V or I can emerge in the HBV of subjects on 1 mg ETV by week 48. –These substitutions often arise in the context of mixtures at these sites and other LVD-resistant mutations at baseline.

27. 27 ETV Resistance Profile – LVD-refractory Subjects ETV-associated resistance substitutions in HBV polymerase: ETV-associated resistance substitutions in HBV polymerase: I169, T184, S202, and/or M250 Emerged only when LVD-resistant mutations at L180 and/or M204 were present at baseline Emerged only when LVD-resistant mutations at L180 and/or M204 were present at baseline 14/189 (7.4%) of evaluated LVD-refractory subjects treated with ETV developed resistance mutations at 48 weeks 14/189 (7.4%) of evaluated LVD-refractory subjects treated with ETV developed resistance mutations at 48 weeks Mutations can be associated with virologic rebound during prolonged therapy (3/14 at Week 48) Mutations can be associated with virologic rebound during prolonged therapy (3/14 at Week 48)

28. 28 Cross-Resistance LVD-resistant HBV clinical isolates showed 3- to 52-fold reduced susceptibility to ETV by in vitro assays LVD-resistant HBV clinical isolates showed 3- to 52-fold reduced susceptibility to ETV by in vitro assays HBV developing ETV-associated resistance substitutions in the clinical trials were susceptible to ADV in vitro but remained resistant to LVD HBV developing ETV-associated resistance substitutions in the clinical trials were susceptible to ADV in vitro but remained resistant to LVD ADV-resistant HBV was susceptible to ETV in vitro ADV-resistant HBV was susceptible to ETV in vitro

29. 29 ETV Resistance – Summary No ETV resistance has been detected in nucleoside-naïve subjects treated with ETV through 48 weeks; longer term data are needed No ETV resistance has been detected in nucleoside-naïve subjects treated with ETV through 48 weeks; longer term data are needed ETV resistant mutations can emerge on ETV treatment when LVD mutations are present; emerge at a rate of <10% at 48 weeks ETV resistant mutations can emerge on ETV treatment when LVD mutations are present; emerge at a rate of <10% at 48 weeks These ETV resistance mutations are associated with virologic rebound These ETV resistance mutations are associated with virologic rebound ETV is cross-resistant with LVD but not ADV by in vitro assays ETV is cross-resistant with LVD but not ADV by in vitro assays

30. 30 Risk-benefit Assessment and Pharmacovigilance Plan

31. 31 Risk-benefit assessment Patients with chronic HBV have increased risk of HCC and probably other malignancies Patients with chronic HBV have increased risk of HCC and probably other malignancies Accumulating evidence that treatment of chronic HBV may decrease the progression of disease Accumulating evidence that treatment of chronic HBV may decrease the progression of disease Efficacy of ETV as measured by liver histology, HBV DNA, or other endpoints better or equivalent to that of LVD over 48 weeks of treatment Efficacy of ETV as measured by liver histology, HBV DNA, or other endpoints better or equivalent to that of LVD over 48 weeks of treatment General safety and tolerability profile of ETV similar to that of LVD General safety and tolerability profile of ETV similar to that of LVD

32. 32 Risk-benefit assessment Positive carcinogenicity findings in animal studies are not rare and are described in approved product labels Positive carcinogenicity findings in animal studies are not rare and are described in approved product labels Animal carcinogenicity studies identify hazard signal not level of risk Animal carcinogenicity studies identify hazard signal not level of risk Quantifying the human cancer risk is difficult Quantifying the human cancer risk is difficult Mechanism of carcinogenicity is likely to be different for different drugs Mechanism of carcinogenicity is likely to be different for different drugs Risk-benefit assessment has been made on a case-by-case basis Risk-benefit assessment has been made on a case-by-case basis

33. 33 Pharmacovigilance plan Increased monitoring and analysis of post-marketing safety reports and reporting to FDA Increased monitoring and analysis of post-marketing safety reports and reporting to FDA Continued tracking of subjects in clinical trials (through ongoing rollover and observational studies) Continued tracking of subjects in clinical trials (through ongoing rollover and observational studies) Proposed large simple safety study to evaluate occurrence of major events in broader clinical use Proposed large simple safety study to evaluate occurrence of major events in broader clinical use

34. 34 Pharmacovigilance plan – Proposed post-marketing study Strengths of proposed study Strengths of proposed study –Study design with randomization, active control, stratification by prior treatment, pertinent endpoints and planned analyses –Will evaluate international population, “real-life” use, allow enrollment of patients with concomitant HCV and HIV and spectrum of HBV disease –Size of study (12,500), enrollment through many local physicians each following relatively small number of patients

35. 35 Pharmacovigilance plan – Proposed post-marketing study Potential limitations of proposed study Potential limitations of proposed study –Length of study may not be adequate to identify malignancies with long latency –Subjects may switch from original assigned treatment to comparator group –Number lost to follow-up may be higher than anticipated –No specific tumor type can be targeted –No way to stratify for all possible co- factors for malignancy in population

36. 36 Pharmacovigilance plan – Proposed post-marketing study Study would be similar in size and scope to some others that have been requested by FDA or that have identified other risk factors Study would be similar in size and scope to some others that have been requested by FDA or that have identified other risk factors Study might identify changes in 5-8 year risk of HCC or other tumors in patients receiving treatment for HBV Study might identify changes in 5-8 year risk of HCC or other tumors in patients receiving treatment for HBV Negative findings at the end of the study may not equate to a conclusion of “no risk” Negative findings at the end of the study may not equate to a conclusion of “no risk”

37. 37 Pharmacovigilance plan – Animal carcinogenicity findings in context Antiviral drugs with positive animal carcinogenicity findings; risk-benefit decisions Antiviral drugs with positive animal carcinogenicity findings; risk-benefit decisions –Zidovudine – Consequences of untreated HIV, uninfected infants exposed perinatally followed in long-term outcome study (PACTG 076/219) –Ritonavir – Consequences of untreated HIV –Ganciclovir – Consequences of untreated CMV, boxed warning in label –Cidofovir – Consequences of untreated CMV, boxed warning in label –Famciclovir – Treatment for HSV, weak hazard signal based on animal data

38. 38 Pharmacovigilance plan – Animal carcinogenicity findings in context Drugs with positive animal carcinogenicity findings approved for other indications Drugs with positive animal carcinogenicity findings approved for other indications –Lipid-lowering drugs –Anticonvulsants –Osteoporosis –ADHD –Gastroesophageal reflux

39. 39 Pharmacovigilance plan The FDA has requested post-marketing studies to assess the risk of human cancer for some approved drugs The FDA has requested post-marketing studies to assess the risk of human cancer for some approved drugs –Long-term prospective observational study of a drug compared to an appropriate control group –Registry of patients using the drug long term –Post-marketing surveillance program –Retrospective cohort study to measure the incidence of a specific tumor and the contribution of a drug

40. 40 FDA Summary In well-conducted clinical trials ETV was shown to provide superior efficacy compared to LVD in multiple analyses of histologic, virologic, biochemical, and composite endpoints In well-conducted clinical trials ETV was shown to provide superior efficacy compared to LVD in multiple analyses of histologic, virologic, biochemical, and composite endpoints Treatment benefit of ETV over LVD greatest in LVD-refractory subjects Treatment benefit of ETV over LVD greatest in LVD-refractory subjects General safety and tolerability of ETV was similar to LVD in all populations studied General safety and tolerability of ETV was similar to LVD in all populations studied Safety and tolerability profile of ETV similar in nucleoside-naïve and LVD-refractory subjects Safety and tolerability profile of ETV similar in nucleoside-naïve and LVD-refractory subjects

41. 41 FDA Summary Pre-clinical studies identified ETV as carcinogenic in mice and rats Pre-clinical studies identified ETV as carcinogenic in mice and rats Clinical relevance of animal carcinogenicity data is unknown Clinical relevance of animal carcinogenicity data is unknown To date, no increase in human malignancies has been identified in the clinical trials To date, no increase in human malignancies has been identified in the clinical trials BMS has proposed a large simple safety study designed to identify increased cancer risk in patients receiving ETV as part of their pharmacovigilance program BMS has proposed a large simple safety study designed to identify increased cancer risk in patients receiving ETV as part of their pharmacovigilance program

42. 42 Questions for the Committee

43. 43 Question 1 How would you assess the risk-benefit of ETV in the context of the available clinical safety, efficacy, resistance, and non-clinical carcinogenicity data?

44. 44 Question 2 A.Does the risk-benefit assessment for entecavir support the approval of entecavir for the treatment of chronic HBV in adult patients? B.If the answer to #2A is no, what information would be needed to support a resubmission?

45. 45 Question 3 A.If the answer to #2A is yes, discuss whether the results of the rodent carcinogenicity studies should impact the Indication and Usage section of the product labeling. B.Based on the available data, discuss the potential role of entecavir in the HBV treatment armamentarium.

46. 46 Question 4 A.Assess the potential risks and benefits of proceeding with development of entecavir for the treatment of chronic HBV in pediatric patients. B.What, if any, additional information is needed in order to proceed?

47. 47 Question 5 Discuss the applicant’s proposed pharmacovigilance plan to address human cancer risk, including comments on the design of the proposed large simple study.

48. 48 Question 6 Are there other issues that you would like to see addressed through post- marketing commitments?