The€Placebo€Effect€in€Epilepsy€Trials: Where€Does€it€Come...
Transcript of The€Placebo€Effect€in€Epilepsy€Trials: Where€Does€it€Come...
The Placebo Effect in Epilepsy Trials:Where Does it Come From?
The Placebo Effect in Epilepsy Trials:Where Does it Come From?
Institute of Neurology and Clinical Pharmacology Unit,University of Pavia, Pavia, Italy
AED XI, 28 April 2011
Emilio PeruccaEmilio Perucca
OutlineOutline
v Definitions
v Magnitude of the improvement onplacebo in epilepsy trials
v Causes, correlates and predictors
DefinitionsDefinitions
Placebo effect
Any effect attributable to a pill, potion or procedure, butnot to its pharmacodynamic and specific properties1
Placeboassociated improvement
Any clinical improvement which occurs duringadministration of placebo (for example, in the placeboarm compared with baseline)
1Wolf S, Pharmacol Rev 1959;11:689704
0
10
20
30
40
How Large is PlaceboAssociated Improvement?Responder Rates on Placebo in Recent RCTs in Adults with
Refractory Partial Seizures
How Large is PlaceboAssociated Improvement?Responder Rates on Placebo in Recent RCTs in Adults with
Refractory Partial Seizures
Trials published in 20082009
13%17%18% 18%
Plac
ebo
resp
onde
r rat
e (%
)
29%
21% 21%26% 26%
14%
20%19% 19%
23%
32%36%
39%
9%
0
10
20
30
40
50
How Large is PlaceboAssociated Improvement?Responder Rates on Placebo in Recent RCTs in Adults with
Primarily Generalized Tonic Clonic Seizures
How Large is PlaceboAssociated Improvement?Responder Rates on Placebo in Recent RCTs in Adults with
Primarily Generalized Tonic Clonic Seizures
17%
32%
Plac
ebo
resp
onde
r rat
e (%
) 45%49%*
* calculated overmaintenance phase
0102030405060708090
100
How Large is PlaceboAssociated Improvement?An Analysis of Three Epilepsy Trials in 28 Dogs
How Large is PlaceboAssociated Improvement?An Analysis of Three Epilepsy Trials in 28 Dogs
Decrease in Seizurecounts vs Baseline
67%
Plac
ebo
resp
onde
r rat
e (%
)
0%
60%
79%
100%
36%
50% SeizureReduction Rates
Munana et al, J Vet IntMed 2010;24:16670
Possible Causes of PlaceboAssociatedImprovement in Epilepsy
Possible Causes of PlaceboAssociatedImprovement in Epilepsy
v Chance
v Regression to the mean, natural history
v Patient’s bias (tendency to “please” caregivers, desire toenter or to remain in the trial)
v Observer’s bias (unconscious tendency to find what oneexpects, or hopes for)
v Higher level of care during the trial
v Emotional / psychological influences on seizuresusceptibility
Changes in Seizure Frequency on PlaceboOccur in Both Directions
Changes in Seizure Frequency on PlaceboOccur in Both Directions
Between 3 and 9% of patients overall showed agreater than 100% increase in seizure frequency(vs baseline)
Tiagabine trials (n=275) 49%Topiramate trials (n=216) 48%Levetiracetam trials (n=310) 45%
E. Somerville, Neurology 2002;59:7983
Percentage of Patients with Increased Seizures in thePlacebo Arm (Trials in Partial Epilepsy)
Factors Reported to Influence Placeboassociated Improvement in Epilepsy Trials
Factors Reported to Influence Placeboassociated Improvement in Epilepsy Trials
v Seizure type (generalized tonicclonic vs partial) orfrequency
v Age
v Number of underlying AEDs
v Duration of the assessment period
v Method used to calculate response
v Year in which trial was conducted
v Location where trial was conducted
Placebo Responder Rates in Partial Epilepsy are Greaterin Children Than in Adults: A Metanalysis of 32 RCTs
Placebo Responder Rates in Partial Epilepsy are Greaterin Children Than in Adults: A Metanalysis of 32 RCTs
05
101520253035404550
Placebo arm
Mea
n re
spon
der r
ate
(%)
Rheims et al, PLoS Medicine 2008;5:e166
Adults Children Adults Children
AED arm
p=0.00130%37%
10%
19%
Factors Correlating with PlaceboAssociatedResponder Rates: A Metanalysis of 63 Trials in
>14,000 Adults with Partial Epilepsy
Factors Correlating with PlaceboAssociatedResponder Rates: A Metanalysis of 63 Trials in
>14,000 Adults with Partial Epilepsy
Responder rates on placebo:
v Tend to increase with increasing duration of treatment(p=0.059)
v Are higher during maintenance than during entire treatmentperiod (p=0.005)
v Increased over the years (p=0.001).
v Are higher with LOCF than with completer’s analysis(p<0.001)
v Seem to show geographical variation
Rheims et al, Epilepsia 2011;52:219233
Time Course of SeizureFree Days in Active and Placebo Armsfrom a Pooled Analysis of Pregabalin Trials in Partial Epilepsy
Ramsay et al, Epilepsia 2009;50:18918
Placebo trials: n =75, r =0.45, p<0.001
Has Placebo Response in Epilepsy TrialsIncreased over the Years?
Rheims et al, Epilepsia Epilepsia 2011;52:219233
Has Effect Size in Active Treatment Arms of EpilepsyTrials Decreased over the Years?
Rheims et al, Epilepsia 2011;52:219233
Is Increase In Placebo Responder Rates Over TimeRelated to Increase in Number of Study Sites?
An Analysis of Trials in Pediatric Depression
Is Increase In Placebo Responder Rates Over TimeRelated to Increase in Number of Study Sites?
An Analysis of Trials in Pediatric Depression
Bridge et al, Am J Psychiatry 2009; 166:42–49
Does Geographical Variation Contribute?Observations from a Recent Addon Trial comparing
Pregabalin, Lamotrigine and Placebo
Does Geographical Variation Contribute?Observations from a Recent Addon Trial comparing
Pregabalin, Lamotrigine and Placebo
“… trial sites in one country enrolled many patients (91out of the 433 total population) new to this level ofmedical care and showed an unusually enhancedresponse in the placebo group. … . their responses toplacebo were nearly three times that seen in patientsfrom other countries (53% vs 19% seizure reduction...)”
“… The factoring of this unusually large placebo effectwith regression to mean effect and Hawthorne effecthas undoubtedly contributed to lack of efficacy versusplacebo for these two established AEDs… ”
Baulac et al, Epilepsy Res. 2010; 91:109
0
10
20
30
40
How Large is PlaceboAssociated Improvement?Responder Rates on Placebo in Recent RCTs in Adults with
Refractory Partial Seizures
How Large is PlaceboAssociated Improvement?Responder Rates on Placebo in Recent RCTs in Adults with
Refractory Partial Seizures
Trials published in 20082009
13%17%18% 18%
Plac
ebo
resp
onde
r rat
e (%
)
29%
21% 21%26% 26%
14%
20%19% 19%
23%
32%36%
39%
9%
Does a true placebo effect exist inepilepsy?
Does a true placebo effect exist inepilepsy?
“We found little evidence in general that placebos hadpowerful clinical effects. Although placebo had no significanteffects on objective or binary outcomes, they had possible smallbenefits in studies with continuous subjective outcomes and for thetreatment of pain.”
Some Neurotransmitters Associated withPlacebo Responses Based on Data from
Imaging Studies
Some Neurotransmitters Associated withPlacebo Responses Based on Data from
Imaging Studies
v Opioid transmitters (pain studies)
v Dopamine (studies in Parkinson’s disease)
v Serotonin (studies in depression)
Diederich and Goetz, Neurology 2008;71:677684
ConclusionsConclusionsv The causes underlying placeboassociated improvement
in epilepsy trials are multifactorial
v Available evidence does not allow to disentangle therelative contribution of individual factors
v No reliable predictors exist of being responder on placeboat individual level –some correlates identified atpopulation level are intriguing
v Better understanding of factors influencing placeboassociated improvement is important to improve thedesign of epilepsy trials
Sagital view of cingulate increases during placebo administration in hemodynamicimaging studies in analgesics studies ( ), acupuncture analgesia ( ), emotionalprocessing/anxiolysis ( ).
Faria et al, Eur Neuropsychopharmacology 2008; 18:47385
Placebo trials: n =75, r =0.45, p<0.001
Relative Risks for Being a Responder (Drug vs Placebo)for 5 AEDs: Partial Epilepsy Trials, Children vs Adults
Rheims et al, PloS Medicine 2068; 5:e166
All trials (U.S. trials in black) U.S. trials onlyn =31, p<0.0018 n= 23, p=0.0018
Has Placebo Response Increased over the Years?Trials in Schizophrenia
Chen et al, Pharm Statistics 2010;9:21729
Time Course of SeizureFree Days in Active and Placebo Armsfrom a Pooled Analysis of Levetiracetam Trials in Partial Epilepsy
French et al, Epilepsia 2005;46:13047
Comparison between Responders and Nonresponders in the Placebo Arms of Levetiracetam
Trials in Adults with Partial Seizures (n=904)
Comparison between Responders and Nonresponders in the Placebo Arms of Levetiracetam
Trials in Adults with Partial Seizures (n=904)
v Age of epilepsy onset was higher in responders thanin nonresponders (20.8 vs 15.2 years, p = 0.019)
v Responder rates were higher when there was onebaseline AED compared with >2 or more (69.0% vs45.6%, p=0.056)
Has Placebo Response Increased over the Years?Trials in Major Depression
Walsh et al, JAMA 2002:287:18407
Placebo trials: n=75, r=0.45, p<0.001