Title: Mevacor Daily lovastatin Tablets, 20 mg Merck
1Mevacor Daily (lovastatin)Tablets, 20 mgMerck
Co., Inc.
- Joint Meeting of the Nonprescription Drugs and
- Endocrinologic and Metabolic Drugs Advisory
Committees - Silver Spring, Maryland
- December 13, 2007
- Shewit Bezabeh MD, MPH
- Division of Drug Risk Evaluation,
- Office of Surveillance and Epidemiology
Center for Drug Evaluation and Research
2Introduction/Background
- Regulatory background
- January, 2005 AC meeting, the NDA was
non-approvable due to several deficiencies - One deficiency - lack of adequate safety data of
lovastatin in patients with asymptomatic liver
disease - Applicant submitted A study of potential
hepatotoxicity of Lovastatin in Kaiser
Permanente Northern California (KPNC) liver
disease population
3Introduction/Background
- Objective of presentation
-
- Review of the submitted population study
- Discuss the studys strengths
- Major limitations
- Channeling bias
- Heterogeneity of baseline liver disease
- Surveillance bias
- Study results
- FDAs Review and Conclusions
-
-
-
4KPNC study
- A retrospective cohort observational study
- KPNC is an integrated health plan with gt 3.2
million members and an electronic information
system
5KPNC study
- Population HMO members with liver disease or
evidence of hepatic dysfunction - Study Period January 1, 1995 to June 30, 2004
6KPNC study
- Inception Cohort
- Adults with evidence of liver disease at
baseline or - High risk for liver disease due to either
elevated liver test or presence of a selected
diagnosis - 13 months of continuous membership
- No statin exposure 1 yr prior to study entry
7KPNC study
- Inclusion Criteria
- Presence of at least one of the following
- Elevated ALT and/or AST at least on two occasions
6 to 18 months apart - Diagnosis of liver disease
- Chronic Viral hepatitis
- Metabolic disorders (Wilsons disease,
hemochromatosis) - Other Chronic Liver Diseases e.g. chronic liver
disease, alcoholic fatty liver, biliary cirrhosis
8KPNC study
- Exclusion Criteria
- Past Hx of the following diagnosis
- Drug-induced liver disease
- Disorders of bilirubin excretion
- Cancer (except non-melanoma skin)
9Study Outcomes
- Primary
- Hys Rule (ALT gt 3x ULN, Total Bili gt 2x ULN
and AP 1.5 ULN) - Secondary
- Liver injury (ALT 3X ULN or Bili 3x ULN)
- Cirrhosis (automated medical records)
- Liver Failure (automated medical records)
10Data analysis
- Major Analysis Primary and secondary outcomes
- Incidence rate ratios (IRR)
- IRR IR (exposed) / IR (unexposed)
- Multivariate analysis
- adjustments for potential confounders
- Age
- Gender
- General health status
- Concomitant medications
11Data analysis
- Secondary and sensitivity analyses
- Liver disease etiology subgroup
- Dose-response (cumulative amount or cumulative
number of days) - Lovastatin discontinuation
- Channeling bias sub-study
- Analysis for confounders age, gender, general
health status and concomitant medications)
12Study Results Demographic Data
13Primary Outcome
- Based on per person-days exposure
- unadjusted for all potential confounders
- Exposure to Lovastatin was associated with
72 decrease in the risk of achieving Hys Rule
endpoint.
14Secondary Outcome
- Laboratory diagnosis
- unadjusted for all potential confounders
15Secondary Composite Outcome
- Combined secondary outcome are not mutually
exclusive (1st occurrence of a liver injury,
cirrhosis, or liver-failure) - unadjusted for all potential confounders
16Conclusions by Study Authors
- Fewer adverse outcomes in KPNC patients with
baseline liver disease who were exposed to
lovastatin compared to non exposed. - Lovastatin exposure appears to be protective of
adverse outcomes in liver disease patients.
17Authors Conclusions - Continued
- Sub-study and sensitivity analyses
- Little evidence for channeling bias (or
confounding by contraindication) - No significant evidence of lovastatin
discontinuation in liver disease patients - Sensitivity analyses did not alter the conclusions
18FDA Analysis Study Limitations
- Channeling Bias
- - lovastatin preferentially prescribed (or
channeled) to individuals at low risk for liver
disease due to prominent labeling for risk of
hepato-toxicity - - lovastatin preferentially avoided in patients
at high risk for liver toxicity
19FDA Analysis Study Limitations
- Channeling Bias (continued)
- - unable to address / adjust for this issue
given limitations of administrative data - - results in residual confounding
20FDA Analysis Study Limitations
- Channeling Bias (continued)
- - Of 6391 patients with both
hypercholesterol-emia (LDL gt 160 mg/dL) and
liver disease, only 2746 ( 43) were treated with
lovastatin. - No explanation was provided why 57 of
hypercholesterolemic (LDL gt 160 mg/dL) patients
did not receive lovastatin suggesting
confounding by diagnosis.
21FDA Analysis Study Limitations
- 2. Baseline liver disease
- The cohort of patients with liver disease
consisted multiple potentially disparate clinical
entities, resulting in clinical heterogeneity and
disparate outcomes - Misclassification
- Stratification
22FDA Analysis Study Limitations
23FDA Analysis Study Limitations
- 3. Surveillance bias
- Tendency to discontinue lovastatin after a
positive test or after a liver disease diagnosis - Surveillance bias Sub Study showed increased
frequency of liver enzyme testing in subjects
with liver disease - lovastatin exposed had 46 more LFT testing
compared to non-exposed - Impact of differential surveillance on study
outcome is not clear
24Overall Conclusions
- A well conducted large study
- Most of the limitations are associated with
observational studies - Evidence of channeling
- Disparate baseline disease
- Surveillance bias
25Summary Conclusions
- The studys findings are consistent with results
of other published studies and suggest that
exposure to lovastatin in patients with baseline
liver abnormalities do not appear to increase the
risk of hepatotoxicity.
26Summary Conclusions
- Because of the limitations and nature
(retrospective observational) of the study, a
clinically significant hepatotoxic effect of
lovastatin cannot be ruled out. - Furthermore, it is not possible to determine if
there is a protective role for lovastatin in the
setting of baseline liver disease.
27Acknowledgements
- CDER Review Team
- Mark Avigan, M.D., C.M. Division of Drug Risk
Evaluation - Allen Brinker M.D., M.S. Division of Drug Risk
Evaluation - Eileen Craig, M.D., Division of Metabolic and
Endocrine Products