Professional Documents
Culture Documents
Placebo in Epilepsy
Placebo in Epilepsy
Placebo Effect i n C a n i n e E p i l e p s y T r i a l s
K.R. Muñana, D. Zhang, and E.E. Patterson
Background: The placebo effect is a well-recognized phenomenon in human medicine; in contrast, little information exists
on the effect of placebo administration in veterinary patients.
Hypothesis: Nonpharmacologic therapeutic effects play a role in response rates identified in canine epilepsy trials.
Animals: Thirty-four dogs with epilepsy.
Methods: Meta-analysis of the 3 known prospective, placebo-controlled canine epilepsy trials. The number of seizures per
week was compiled for each dog throughout their participation in the trial. Log-linear models were developed to evaluate
seizure frequency during treatment and placebo relative to baseline.
Results: Twenty-two of 28 (79%) dogs in the study that received placebo demonstrated a decrease in seizure frequency
compared with baseline, and 8 (29%) could be considered responders, with a 50% or greater reduction in seizures. For the 3
trials evaluated, the average reduction in seizures during placebo administration relative to baseline was 26% (P 5 .0018), 29%
(P 5 .17), and 46% (P 5 .01).
Conclusions and Clinical Importance: A positive response to placebo administration, manifesting as a decrease in seizure
frequency, can be observed in epileptic dogs. This is of importance when evaluating open label studies in dogs that aim to assess
efficacy of antiepileptic drugs, as the reported results might be overstated. Findings from this study highlight the need for more
placebo-controlled trials in veterinary medicine.
Key words: Clinical trials; Dog; Epilepsy; Statistical modeling.
having generalized seizures based on owners’ descriptions of seizure For a parallel study, the study period is divided into 2 periods:
episodes. Some of the dogs were also reported to have occasional period 1 for the baseline and period 2 for the treatment or the pla-
partial seizures. No attempt was made to further characterize the cebo. Using the above notations, we consider the following log-
seizure type with video monitoring or electroencephalography. The linear model for mij:
therapeutic interventions evaluated included a surgical implant,6 di-
etary modification,a and a novel antiepileptic drug (unpublished logðmij Þ ¼ b0 þ b1 IðtreatmentÞ þ b2 Iðperiod2Þ; ð2Þ
data). All studies were placebo controlled and blinded such that
both the owner and veterinary personnel involved in evaluating the where b0 has the same interpretation as before, b1 is the effect of the
dog were unaware of the treatment being administered. The novel treatment relative to the placebo, and b2 is the placebo effect plus
the possible time effect of period 2 relative to period 1.
interventions were evaluated as add-on therapies, and as such, dogs
The assumption was made that the variance of Yij has the form of
were maintained on their previously prescribed conventional anti-
epileptic medication throughout the study. The surgical implant and Poisson variance function with a possible overdispersion. To ac-
drug study were cross over in design, and consisted of an 8-week count for the correlation in the longitudinal data in each patient, the
inference on effect parameters was carried out by the generalized
baseline period followed by 2 treatment periods of 13–16 weeks in
estimating equation (GEE) methodology with equal correlation ma-
length, with a 4-week washout period between the 2 treatment pe-
riods. The diet trial was designed as a parallel study, and included a trix as the working correlation matrix.7 A commercially available
baseline period of 3–6 months followed by a treatment period of 6 software programb was used to implement the analysis.
months. In each study, the treatment group received the experimen-
tal therapeutic intervention in addition to conventional antiepileptic
drug treatment, and the control group received a matching placebo Results
in addition to conventional antiepileptic drug treatment.
Seizure monitoring was based on owner observations. Informa- Thirty-four dogs were included in this study: 9 dogs as
tion was recorded on study forms that were adapted from those used participants in the study evaluating the surgical implant,
for human seizure monitoring. Owners were instructed to complete 14 as participants in the study evaluating novel drug
a daily diary noting medication administered, the presence of any therapy, and 11 as participants in the study evaluating
seizure activity, and whether any signs of illness, change in activity
dietary modification. In the latter study, which was de-
or attitude were noted. Owners were asked to complete an addi-
tional form immediately after any observed seizure that captured
signed as a parallel study, 5 of 11 dogs were randomized
information on specific manifestations of the episode. Based on to placebo treatment while the remaining 6 dogs were
these observations, a weekly seizure frequency was compiled randomized to the active treatment.
for each dog throughout the study. The average weekly seizure Data on seizure frequency during active treatment and
frequency for each of the study periods (baseline, placebo adminis- placebo administration relative to baseline for dogs in
tration, and active treatment) was calculated. As is the convention the 3 trials are summarized in Table 1 and Figure 1.
for human epilepsy trials, a positive treatment response was defined Log-linear analysis of weekly seizure frequency for the
as a 50% or greater reduction in seizure frequency compared with population in dogs in each trial identified a reduction in
baseline. To assess for any changes in drug levels that might affect seizures during placebo administration of 26–46%. The 2
seizure frequency, mean serum antiepileptic drug concentrations
cross-over studies involving the surgical implant and the
obtained at the end of baseline, placebo administration and active
treatment were calculated and compared for each trial.
drug trial demonstrated a reduction in seizures during
placebo administration compared with baseline of 26%
(P 5 .0018) and 29% (P 5 .17), respectively. For the di-
Statistical Analysis etary modification trial, which was designed as a parallel
Log-linear models were developed for each of the study designs study, dogs administered placebo had a 46% (P 5 .01)
(cross-over and parallel) to evaluate weekly seizure frequency dur- decrease in seizure frequency from baseline.
ing treatment and placebo relative to baseline. Such models account There were no significant changes in serum concentra-
for the treatment effect as well as the effect of time. tions of either phenobarbital or bromide throughout the
For a cross-over study, the study period can be divided into 3 study, although dogs in both the surgical implant trial
periods: period 1 corresponding to the baseline followed by 2 peri-
and the drug trial demonstrated a trend toward increas-
ods for the treatment and placebo. Denote by Yij the seizure count
for the ith dog at the jth week and let mij 5 E(Yij) be the average
ing serum bromide levels during placebo administration
seizure count per week. Because the dogs are randomized to the relative to baseline that was not associated with a change
treatment and placebo, we then consider the following log-linear in potassium bromide dosage. Similarly, mild increases in
model for mij: both phenobarbital and bromide levels from baseline
were seen in dogs in the placebo arm of the diet trial
logðmij Þ ¼ b0 þ b1 I ðtreatment at period 2Þ þ b2 I ðtreatment at period 3Þ (Table 2).
þ b3 Iðplacebo at period 2Þ þ b4 Iðplacebo at period 3Þ
ð1Þ Discussion
where I( % ) is an indicator function. With this model specification, b0 Results from this study suggest that a positive response
is the log average seizure count per week at period 1 (assumed to be to placebo administration can be observed in epileptic
independent of treatment sequence because of randomization), (b1
dogs. A decrease in seizure frequency was seen in more
1 b2)/2 can be interpreted as the average treatment effect relative to
baseline and (b3 1 b4)/2 is the average placebo effect relative to
than half of the dogs in the trials during treatment with
baseline. Then b1 & b3 is the treatment effect relative to placebo at placebo when compared with baseline values. Eight of 28
period 2, b2 & b4 is the treatment effect relative to placebo at period dogs (29%) evaluated in this study could be classified as
3, and (b1 & b3 1 b2 & b4)/2 is the average treatment effect relative responders, based on a 50% or greater reduction in sei-
to the placebo. zure frequency. Futhermore, placebo administration was
168 Muñana, Zhang, and Patterson
Table 1. Seizure frequency relative to baseline during administration of active treatment and placebo for dogs in
epilepsy trials.
Number (%) of Dogs with Number (%) of Dogs Number (%) of Dogs
Decrease in Seizures Classified as Respondersa Seizure-free
associated with an estimated reduction in seizure fre- An obvious difference between the current study and
quency of 26–46% in the population of dogs studied. those involving human subjects involves the inherent na-
This response was identified across the different treat- ture of the placebo response. In human studies, placebo
ment modalities evaluated in these trials, which included effects are primarily because of expectations of an indi-
a surgical implant, a novel antiepileptic drug, and a di- vidual as a patient being treated in a double-blind
etary modification, suggesting that this effect is not manner. In veterinary studies, the placebo response can
specific with respect to treatment type. be because of effects on the animal, but more importantly
The findings from this study are consistent with results may be a result of expectations of the pet owner regard-
from similar studies evaluating the placebo effect in hu- ing treatment. It is likely that the expectations of the
man epilepsy. A recent meta-analysis of randomized owner might play a greater role in a study such as this,
trials of antiepileptic drugs in human patients identified where the owners are responsible for administration of
a placebo response rate of 9.3–16.6%,8 while an earlier treatment and outcome measures (ie, seizure frequency)
study comparing data from trials involving 5 antiepilep- are derived solely from owner observations.
tic drugs demonstrated placebo response rates of 4–18%, It is possible that the effect observed in the present
with an average rate of 10% over all studies.9 study is because of nonspecific antiepileptic properties of
Fig 1. Box and whisker plots of the weekly seizure frequency during baseline, placebo administration, and active treatment for dogs in the
surgical implant trial (a), novel drug trial (b), and dietary modification trial (c). Central lines of the box represent the median, upper and lower
limits of the box represent the 75th and 25th percentiles, and whiskers represent the 95th and 5th percentiles. Mean values are denoted by &.
Placebo Effect in Epilepsy 169
Table 2. Serum antiepileptic drug concentrations during baseline, active treatment, and placebo administration for
dogs in epilepsy trials.
Phenobarbital Serum Concentration (mg/mL) Bromide Serum Concentration (mg/dL)
the placebo, but this is considered unlikely. Although a outcome measure, with larger effects observed in trials
decrease in seizure frequency was observed during pla- reporting subjective outcomes.12
cebo administration, the change in seizure frequency Finally, it is possible that improved adherence to the
relative to baseline did not assume significance in all of established antiepileptic drug regimen during participa-
the trials (P 4 .05). Rather, it is more likely that the ob- tion in the trial played a role in the present study findings.
served effect is because of what has been termed the Although prescribed dosages of conventional antiepilep-
‘‘perceived placebo effect,’’ or a change that is observed tic medications stayed constant throughout the studies, a
after placebo administration that is not directly caused to slight increase in bromide serum concentrations relative
placebo.10 The perceived placebo effect includes any to baseline was identified during placebo administration
changes associated with the administration of placebo, in all of the trials, and a similar trend in phenobarbital
including the effect of time. It is important to note serum levels was observed in the diet trial. The differ-
that the statistical models utilized in this study do not ences in serum drug concentrations were relatively small
differentiate the treatment effect from the effect of time, and did not approach statistical significance; nonetheless
and consequently the analysis performed provides a mea- this may have contributed to the reduction in seizure fre-
sure of multiple potential factors that can contribute to quency identified during placebo administration.
treatment response. These factors may include regression A greater placebo response was identified in the paral-
to the mean, investigator bias, client bias, the potential lel study compared with the cross-over studies. Sixty
for a higher level of care during the study, and improved percent of dogs in the parallel study were classified as
adherence to treatment with active medication that is placebo responders, compared with response rates of 0%
being administered in addition to the placebo during the and 36% in the cross-over studies. Similarly, the overall
study. decrease in seizure frequency during placebo administra-
Regression to the mean is a statistical term used to de- tion compared with baseline for the population of dogs
scribe the fluctuations of biological variables that occur was 46 % for the parallel study, compared with values of
over time and take the form of a sine wave around the 26% and 29% for the cross-over studies. Although the
mean.11 Epilepsy is a waxing and waning disorder, and total number of dogs in the trials is too low to draw any
fluctuations in seizure frequency are common over the conclusions about these percentages, it seems plausible
course of the disease. Owners are most likely to seek a that owner bias might assume a greater role in a parallel
change in therapy for their pet when seizures are under study design than in a cross-over design. Owners that
poor control. Over the short term, improvement in the choose to enroll their dogs in trials tend to have a positive
seizure frequency is probable, regardless of the treatment attitude toward the new intervention being evaluated and
administered. However, this improvement is often erro- its potential to help their animal. This desire to have the
neously attributed to a recently instituted change in treatment be effective can influence owners’ observations
therapy, whereas in fact it is because of an effect of time. and data collection for the study. The cross-over study
The potential for bias by individuals involved in the design, in which each dog receives the therapeutic agent
study cannot be overlooked. Even in blinded studies, ex- being tested, might minimize this bias. In contrast, the
pectations of a response can influence the subjective parallel study design, in which each participant receives
interpretation of study data by the investigator. The po- either the test agent or placebo, has the potential to high-
tential for participant bias, known as the Hawthorne light this source of bias.
effect, is a well-recognized phenomenon in clinical trials. The beneficial effects of placebo administration are
The act of participating in a trial provides individuals well documented in laboratory animals.1 Animal studies
with a better experience because of the focus of interest have served as the basis of investigation into the mecha-
toward them, which is rewarding for its own sake. nism of the placebo effect for most of the past century.
Consequently, participants document better results re- However, to the authors’ knowledge there is only one
gardless of the change provided or the treatment other report that specifically addresses the placebo effect
experienced. Because canine epilepsy trials rely on owner in clinical veterinary medicine.5 Jæger and colleagues
observations to quantify seizure frequency, an owner’s performed a controlled parallel study to evaluate gold
expectations of a trial can influence the study results for bead therapy as a treatment for canine hip dysplasia. The
their dog. Indeed, it has been shown that the size of the main outcome parameter in the study was a change in
placebo effect may be associated with the nature of the pain signs as assessed by the owner. Owners were also
170 Muñana, Zhang, and Patterson