Transcript Talk Title

What Accounts for Placebo Group
Improvements and Can They Be Reduced?
John T. Farrar, MD, PhD
Departments of Epidemiology
Anesthesia (Secondary)
and Neurology (Secondary)
University of Pennsylvania
CCEB
“Placebo Response” Versus
Placebo Group Response
• Response in a placebo treated group
–Natural history of disease
–Regression to the mean
–Brain-body effect (placebo effect)
• Response in a drug treated group
–Natural history of disease
–Regression to the mean
–Brain-body effect (placebo effect)
–Specific effect of the therapy
CCEB
Factors That Affect
Individual Response
• Natural history
• Regression to the mean
• Brain-body effect
–This is the only truly individual
characteristic “placebo” effect
• Specific effect of therapy
CCEB
Descending
Pain
Modulation
Pathways
Kandel and Schwartz,
Principles of Neural Science
Issues in Considering
Placebo Response Reduction
• Physiologic placebo response
–Appropriate to control
» Independent from treatment response
» No major overlap with treatment response
–In-appropriate to control
» Similar or related to treatment mechanism
» Facilitates treatment response
CCEB
Potential Reasons to Reduce
Placebo Group Response Rate
• Statistical – Comparison of change
in group levels is harder to detect the
closer the underlying group response
is to 0.5 or 50%
• Measurement – Ceiling affect
• Reduction in size of the detectable
difference between groups (more
efficient)
CCEB
Does the Placebo Response
Affect the Success Rate of RCTs
• Larger placebo response is associated
with lower likelihood of a statistically
positive study*
• This does not prove that the higher group
placebo rate causes the study failure
• Also does not prove that excluding
placebo responders would change the
results
*Katz, Finnerup, Dworkin: Neurology 2008
CCEB
Thoughts About Reducing
Placebo Response Rate
• Exclude placebo responders? (which type?)
• Conduct longer trials – placebo response
may not last as long (controversial)
• Select patients with worse pain
–Lower physiologic placebo response (maybe)
–Larger placebo group response (probably)
» (Regression to the mean?)
1.Dworkin RH, Turk DC, Peirce-Sandner S, et al. Placebo and treatment group
responses in postherpetic neuralgia vs. painful diabetic peripheral neuropathy
clinical trials in the REPORT database. Pain. Jul 2010;150(1):12-16.
CCEB
Placebo Rates in Neuropathic
Pain Clinical Trials
• Placebo response level increased
with time up to 19 weeks
• Calendar time (since 1990s )
– placebo rates have been flat
• Placebo response by disease:
–PHN average 15%
–DPN average 26%
*Quessy, Rowbotham: Pain 2008
CCEB
Problems in Placebo “Responder”
Exclusion
• Placebo run-in responder
–Does not definitively identify brain-body
placebo responder or non-responder
• Natural history/ Regression to the mean
–If you remove those getting better will be
left with only those getting worse
–Those may be likely to get better again in
the next period
CCEB
Other Issues to Think About
• Even if know it is placebo may respond *
• Ranking of placebo – pill vs shot**
• Nocebo effect of delayed treatment group
–Knowing not getting the new treatment
• Proof of concept studies
– If can’t show an effect with unblinded
administration, then
unlikely to be an appreciable effect
CCEB
* Kaptchuk TJ, Friedlander E, Kelley JM, et al. Placebos without deception: a
randomized controlled trial in irritable bowel syndrome. PLoS One.
2010;5(12):e15591.
**Kaptchuk TJ, Stason WB, Davis RB, et al. Sham device v inert pill: randomised
controlled trial of two placebo treatments. BMJ. Feb 18 2006;332(7538):391-397.
Design Considerations
• Placebo run in
–Has not made a difference in depression
• Suppressing placebo response may be
unhelpful or even counter productive*
*Quessy, Rowbotham: Pain 2008
CCEB
Osteo-Arthritis Trials
• Did study to look at M-F response to
Cox-2 vs Non-specific vs Placebo
• Harmonized 6 OA and 5 RA studies
• Analyzed to look at prediction of
response in placebo and treatment
CCEB
Data Shown – Not Yet Published
• Demographics of 9 NSAID Trials
• Comparison of Male and Female
Response
• Predictive Models
CCEB
Neuropathic Pain Studies
• Harmonized 8 studies with pain diaries
–Neuropathic Pain Group
» 3 Diabetic Neuropathy
» 1 Post-herpetic neuralgia
» 1 Mixed neuropathic pain
– 3 Fibromyalgia Studies
CCEB
Data Shown – Not Yet Published
•
•
•
•
•
•
CCEB
Demographics for Neuropathic Pain
Univariate Model Neuropathic Pain
Multivariate Model Neuropathic Pain
Demographics for Fibromyalgia
Univariate Model Fibromyalgia
Multivariate Model Fibromyalgia
BOCF Response Rate in DPNP Studies
Variation of Baseline Daily Pain Rating
DPNP
SD<1.2 (N=847)
SD ≥1.2 (N=177)
60
50% BOCF (%)
30% BOCF (%)
70
*
50
40
30
SD <1.2
SD ≥1.2
*
*
70
60
50
40
30
20
20
10
10
0
0
DLX
PBO
Zhang S, Skljarevski V, Dworkin RH Poster @ADEPT 2011
CCEB
Future Considerations
• Look at other patient characteristics
–Baseline SF-36 subscales
–Depression measures
• Look at other study characteristics
–Number of arms
–Countries included
–Number of recruits per site
CCEB
Thank you
•
•
•
•
•
CCEB
Andrea Troxel
Kevin Haynes
Ann Tierney
FDA
ACTION