Statistical Issues and Challenges in Combination Vaccines Ivan S. F. Chan, Ph.D. Merck Research Laboratories Sang Ahnn, Ph.D. CBER, FDA 2005 FDA/Industry Statistics Workshop Washington D.C. September 14-16, 2005 Note: Sang Ahnn's opinions are his own and do not necessarily reflect FDA policy
27
Embed
Statistical Issues and Challenges in Combination Vaccines Ivan S. F. Chan, Ph.D. Merck Research Laboratories Sang Ahnn, Ph.D. CBER, FDA 2005 FDA/Industry.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Statistical Issues and Challenges in Combination Vaccines
– Exact (e.g., Chan and Zhang 1999, Chan 2002, 2003)
• Confidence interval (CI) for – Test-based methods provide consistency with p-value
– CI lower bound is > (-) if and only if the non-inferiority is demonstrated
• Regression (ANCOVA) for GMT analysis
16
Power and Sample Size Determination
• Asymptotic (Farrington and Manning 1990) or exact (Chan 2002) methods
• Power is sensitive to the true responses– best case scenario is PT = PC (or GMTT = GMTC )
– useful to assess power assuming PT (or GMTT) slightly less than PC (or GMTC)
• Need to evaluate overall power for demonstrating success for all components
17
Consistency Lots Study• For licensure of a new vaccine, consistency of
the manufacturing process must be supported by clinical trials using at least 3 lots of the vaccine
• Objective of study is to show equivalence of clinical response among consistency lots– Evaluate both immune responses and safety
• For combination vaccine, an active control may be included to demonstrate “assay sensitivity”– First demonstrate lot consistency
– Then combine data from consistency lots to show noninferiority to active control
18
Analysis of Consistency Lots Study• Hypothesis of interest:
H0: |Pi – Pj| ≥ for at least one pair (i, j) vs. H1: |Pi – Pj| < for every pair
• Same as testing 3 pairs of noninferiority hypotheses H0: |P1 – P2| ≥ H0A: P1 – P2 ≥ or H0B: P2 – P1 ≥
• Test each hypothesis at one-sided 5% level– i.e., requiring 90% CI completely within [-, ]
– Control overall type I error rate at 5%
19
Example:Consistency Lots Study for MMRV
• Study enrolled ~4000 subjects– 3 lots of MMRV and an active control (MMR+V)
• 1st step: showed consistency of 3 MMRV lots– Antibody response rates and GMTs for measles,
mumps, rubella, and varicella– 24 pairwise comparisons
• 2nd step: showed the combined responses of MMRV is noninferior to those of MMR + V– 8 comparisons
20
Some Challenges of Developing Combination Vaccines
• Statistical challenge with multiple endpoints
• Clinical challenge with potential interaction among components on immunogenicity and safety
• Formulation issues– Compatibility of components/Stability?
21
Multiplicity Challenge with Combination Vaccines
• Major impact on power and sample size because of the need to show success on all components
• Impact even more in consistency lots study
• Accounting for correlation among components only increases power slightly
• May consider multivariate analysis (e.g. T2 test)?
22
Multiplicity Hit on Power and Sample SizeCombination Vaccine Studies
# of Vaccine Components
( = 0.1)
Power (%) with a fixed sample size
Sample Size Per Group Needed to Maintain 90% overall power for study
1 90 205
2 81 250 (22%↑)
4 66 293 (44%↑)
8 43 335 (63%↑)
12 28 357 (74%↑)
23
Challenges of Potential Interaction with Combination Vaccines
• Potential interactions among components are difficult to predict– Immune responses may be lower– Safety concern may arise
• A new dose-response study may be needed to re-establish the optimal dosing
24
Examples of Interactions:MMRV Vaccine
• Interaction observed in early clinical studies– Suboptimal varicella responses
• A new dose-response study (N>1500) established optimal potency range for varicella– Acceptable varicella responses– But more fever (≥102 F) and measles-like rash
• A large database (N>5800) confirms that MMRV is well tolerated – Both fever and measles-like rash were transient and
resolved with no long-term complications
25
Formulation Issues with ProQuadTM
• M-M-R™ II is refrigerated product• VARIVAXTM is frozen product• European market needs refrigerated
combination vaccine – how to develop?– First, develop frozen ProQuad to gain regulatory
approval– Then, introduce refrigerated ProQuad via
manufacturing supplement (a 4-6 months regulatory review time)
26
Summary
• Combination vaccines provide substantial public health benefits
• Immune marker and selection of endpoints and margins are important to the development of combination vaccines
• Special considerations for studies of consistency lots
• Potential interactions among components and multiplicity of endpoints pose special challenges
27
References• Blackwelder WC. Similarity/equivalence trials for combination vaccines. Combined
Vaccines and Simultaneous Administration (ed by Williams JC,Goldenthal KL, Burns DL, Lewis BP), Ann New York Acad Sciences 1995;754:321-328.
• Chan ISF and Zhang Z. Test-based exact confidence intervals for the difference of two binomial proportions. Biometrics 1999; 55: 1202-1209.
• Chan ISF. Power and sample size determination for non-inferiority trials using an exact method. Journal of Biopharmaceutical Statistics 2002; 12: 457-469.
• Chan ISF. Proving non-inferiority or equivalence of two treatments with dichotomous endpoints using exact methods. Statistical Methods in Medical Research 2003; 12 (1): 37-58.
• Chan ISF, Wang WWB, Heyse J. Vaccine clinical trials. Encyclopedia of Biopharmaceutical Statistics, 2nd Edition, 2003, 1005-1022.
• Farrington CP and Manning G. Test statistics and sample size formulae for comparative binomial trials with null hypothesis of non-zero risk difference or non-unity relative risk. Statistics in Medicine 1990; 9, 1447-1454.
• FDA: Guidance for industry for the evaluation of combination vaccines for preventable diseases: production, testing and clinical studies. 1997. http://www.FDA.GOV/CBER/gdlns/combvacc.pdf.
• Kuter B, Hartzel J, Schodel F. The Challenges of Developing a combination Measles, Mumps, Rubella & Varicella Vaccine (ProQuad™), ESPID symposium, Valencia, Spain, May 2005.
• Miettinen O and Nurminen M. Comparative analysis of two rates. Statistics in Medicine 1985; 4, 213-226.