Guidance Update: Average, Population, and Individual Approaches to Establishing Bioequivalence

1
Guidance UpdateAverage, Population, and
Individual Approaches to Establishing
Bioequivalence
Advisory Committee for Pharmaceutical Science

• Mei-Ling Chen, Ph.D.Associate Director
• Office of Pharmaceutical Science
• Center for Drug Evaluation and ResearchFood and
  Drug Administration

November 16, 2000
2
Statistical Guidances on Bioequivalence

• 1992 GuidanceStatistical Procedures for
  Bioequivalence Studies Using Standard
  Two-Treatment Crossover Design
• 1997 Preliminary Draft GuidanceIn Vivo
  Bioequivalence Studies Based on Population and
  Individual Bioequivalence Approaches
• 1999 Draft GuidanceAverage, Population and
  Individual Approaches to Establishing
  Bioequivalence

3
The 1999 Draft Guidance

• Incorporates 1992 Guidance
• Updates 1997 Preliminary Draft Guidance
• Focuses on various study designs, statistical
  criteria and methodologies for establishing
  bioequivalence
• Does not deal with the question of when to
  usea specific criterion

4
Regulatory Application

• Replicate designs are recommended for modified
  release dosage forms and highly variable drug
  products.
• Average BE is recommended for comparison of
  bioavailability measures in most BE studies.
• Sponsors/applicants have the option to provide
  rationale for using another criterion, e.g.,
  individual BE for highly variable drug products.
  
• Currently, population BE is used for the in vitro
  testing of nasal and oral inhalation drug
  products.

5
Statistical Guidance
Outline

• Introduction
• Background
• Statistical Model
• Bioequivalence Criteria
• Study Design
• Statistical Analysis
• Miscellaneous Issues
• Appendices

6
Types of Bioequivalence

• Average Bioequivalence (Current)- Population
  means (?T , ?R)
• Population Bioequivalence- Population means (?T
  , ?R) - Total variances (?TT2, ?TR2)
• Individual Bioequivalence - Population means
  (?T , ?R) - Within-subject variances (?WT2 ,
  ?WR2)- Subject-by-formulation interaction (?D2)

7
Bioequivalence Criteria

• Criterion ? BE Limit
• Average BE (mT - mR)2 ? qA2
  (mT - mR)2 sD2
  (sWT2 - sWR2)
• Individual BE -----------------------------
  -------------- ? qI
  sWR2
  (mT - mR)2 (sTT2 - sTR2)
• Population BE ------------------------------
  ----- ? qP
  sTR2

8
Distance Concept

• A key concept for the population and individual
  BE criteria is to compare the distance measure
  between the T and R formulations, T-R, with that
  of the reference formulation against itself,
  R-R. The distance measure is expressed by the
  expected squared difference and this comparison
  is denoted as a ratio of the two measures
  indicated below.
  E (T - R)2
  ---------------------
  E (R - R)2

9
General Form

• Derived from the distance concept, both
  population and individual BE criteria compare the
  distribution (combination of mean and variance)
  between the T and R products, and result in a
  general form as follows (Average Difference)2
  Variance Terms---------------------------------
  --------------------- ? BE Limit
  Reference Variance
• Both criteria are scaled to the variability of
  the reference product - this is referred to as
  reference-scaling.

10
Constraint on Mean Difference

• (Average Difference)2 Variance
  Terms--------------------------------------------
  ---------- ? BE Limit
  Reference Variance
• Concern for a possible large average difference
  due to the mean-variance tradeoff and
  reference-scaling
• Resolution - In addition to the BE limits based
  on confidence bounds, the guidance recommends
  further constraint on the point estimate of
  the geometric T/R mean ratio to fall within
  80-125.

11
Guidance Status
Draft GuidanceAugust 1999
Public Comment Period
Review of Public Comments by Working Group (WG)
Revision of Draft Guidance by WG and BCC
Internal Clearance