Download The placebo effect in neurological disorders

Reviews
The placebo effect in neurological disorders
Raúl de la Fuente-Fernández, Michael Schulzer, and A Jon Stoessl
Recent evidence suggests that the placebo effect is
mediated by the dopaminergic reward mechanisms in the
human brain and that it is related to the expectation of
clinical benefit. On the basis of this theory, we propose
some criteria for the proper investigation of the placebo
effect, and review the evidence for a placebo effect in
Parkinson’s disease, depression, pain, and other
neurological disorders. We also discuss the evidence for
the use of placebos in long-term substitution programmes
for the treatment of drug addiction.
Lancet Neurology 2002; 1: 85–91
Any treatment (physical, pharmacological, or psychological)
for a medical condition can potentially have a dual effect for
the patient: that related to the treatment itself (eg, the
intrinsic pharmacological property of an active drug) and
that inherent in the perception that the treatment is being
received.1,2 The latter is known as the placebo effect.3 Doubleblind randomised placebo-controlled trials were originally
designed to control for such an effect. Since distinction
between the effects of an active psychological treatment and
the placebo effect may be difficult,4 in this review we focus
on placebo responses associated with physical and
pharmacological treatments.
We start off by reviewing the history and definitions of
“placebo” and “placebo effect”, then attempt to describe the
ideal methods to investigate the placebo effect, before
discussing the placebo effect as it relates to neurological
disorders. We use the term neurology here in its broadest
sense; that is, including medical conditions, such as
depression, once believed to be in the field of psychiatry.
Placebos and the placebo effect
Historical background
“Placebo” is Latin for “I shall please”. Traditionally, the
word placebo has had negative connotations, which have
often been reflected in the medical literature. In 1811
(Hooper’s Medical Dictionary), placebo was defined as “an
epithet given to any medicine adapted more to please than to
benefit the patient”.5 More recent definitions tend to avoid
pejorative terms.6
Most medicines, potions, and remedies used in ancient
times would today be labelled as placebos.4 One can assume
that if any of these treatments had benefits, they must have
been related to the placebo effect. This assumption has led
some to speculate that the placebo effect could have evolved
by natural selection.7 Accordingly, placebo responders
would have a higher chance of survival. However, many
THE LANCET Neurology Vol 1 June 2002
studies have been done and none has shown a consistent
placebo-reactor profile.4,5,8,9 Cultural factors should also be
taken into account in assessment of the placebo effect. For
example, the magnitude of the placebo effect in ulcer disease
may vary from one country to another.10
Definitions
It is important to distinguish between placebo and placebo
effect.9 Essentially, any sort of treatment can act as a placebo,
but what determines whether there is a placebo effect is the
response of the patient to the intervention. The magnitude
of the response to the placebo varies according to its
supposed potency.11 For example, placebo surgery seems to
be more effective than a placebo pill.4,5,12
Definitions of the placebo effect are abundant in the
literature. Perhaps one of the most commonly used is that
provided by Wolf who defined the placebo effect as: “any
effect attributable to a pill, potion, or procedure, but not to
its pharmacodynamic or specific properties”.6 One problem
with most definitions is their implication that the placebo
effect is non-specific. Kirsch, however, pointed out that,
whereas the ingredients of a placebo preparation may be
totally non-specific, the effects of placebos can be very
specific.13 The specificity of the placebo effect depends on the
information given to the patient (ie, the expectation). For
example, placebos can have opposite effects on heart rate or
on blood pressure depending on whether they are given as
tranquillisers or as stimulants.13 Thus, precise definition of
“placebo effect” is difficult.
Other treatment effects
There are some effects associated with treatment—such as
the effect of informed-consent procedures, the effect of
medical and nursing care (the “Hawthorne” and “halo”
effects), and the patient-doctor relationship—that can
contribute to the perception of clinical benefit.14 Whether
these effects can be legitimately considered as part of the
placebo effect is a matter of debate.
There are, however, other factors leading to the spurious
perception of benefit from the treatment received that need
to be identified and differentiated from the placebo effect.
RF-F, MS, and AJS are all at the Pacific Parkinson’s Research
Centre, University of British Columbia, Vancouver, BC, Canada.
Correspondence: Dr A Jon Stoessl, Pacific Parkinson’s Research
Centre, University of British Columbia, Purdy Pavilion, 2221
Wesbrook Mall, Vancouver, BC, Canada V6T 2B5.
Tel +1 604 822 7967; fax +1 604 822 7866;
email [email protected]
http://neurology.thelancet.com
85
For personal use. Only reproduce with permission from The Lancet Publishing Group.
Review
The placebo effect in neurological disorders
Investigation of the placebo effect
Study design Groups
Standard
Active drug and placebo
two-group
RCT
Objective
Study of the efficacy of
the active drug (with control
for the placebo effect)
Placebo effect
Not measured directly; comparison
with baseline values gives inaccurate
estimates of the placebo effect
Other twogroup RCT
Placebo and untreated
Study of the placebo effect
Underestimated
Standard
three-group
RCT
Active drug, placebo,
and untreated
Study of both the efficacy Underestimated
of the active drug (with
control for the placebo effect)
and the placebo effect
Multi-group
RCT
Patients allocated to one
of several doses of an
active drug (or to one of
several active drugs), to
the placebogroup, or to
the untreated group
Study of both the efficacy Potentially overestimated
of the active drug (with
control for the placebo effect)
and the placebo effect
These include: the natural course of the disease (eg, the
symptoms of a cold may be over in 3–5 days whether or not
it is treated); regression to the mean (owing to the random
error in the measurements, or the random variability of the
characteristic being measured, patients with initially extreme
scores in a particular scale tend to “regress” to their steadystate values in subsequent follow-up visits);15 and other time
effects, such as changes over time in measurement
sensitivity.16
As long as the influence of these factors and non-specific
effects are similarly distributed between the group receiving
the active treatment and the placebo group, a randomised
controlled trial (RCT) should provide a reasonable estimate
of the difference between the effect of the active treatment
and the placebo effect.
Placebos, active drugs, and their interactions
An important assumption made when evaluating the results
of an RCT is that the effects of the active treatment and the
placebo are additive.3 This additive model may be too
simplistic,17 however, because it does not account for the
possibility of interactions between the effect of the active
treatment and the placebo effect.
For example, the placebo effect associated with the
simple act of ingesting a pill may potentiate the actual
physiological effect of the drug (positive interaction).
Conversely, the placebo effect could diminish the
physiological response to the drug, especially in patients
with a strong placebo effect (negative interaction). The
extent of such interactions may also depend on the design
selected for the RCT (eg, parallel-group versus crossover
design). Such interactions may jeopardise the RCT if they
are ignored in the statistical analysis.17
The situation could be even more complicated: the
placebo effect could be conditioned in either direction by the
presence of an active treatment. Naturally, this is still a
statistical interaction, but, by contrast with the previous
situation, the causal link would be in the opposite direction.
For example, the magnitude of the placebo response
may vary according to the circulating concentration of
the active drug.
86
Although the ideas developed
here are applicable to crossover
studies, in the next section we
concentrate on the parallel-group
design to avoid additional
problems
of
interpretation
associated with carry-over and
period effects.15
Investigation of the placebo
effect
The power of placebos to alleviate a
great variety of medical conditions
has long been recognised3 but has
recently been challenged.18 We
describe here what we consider to
be the requisites for investigating
the placebo effect under ideal
conditions, and which medical conditions or populations of
patients might not be suitable for such investigations. On the
basis of recent evidence, we also emphasise the link between
the expectation of benefit and the placebo effect.
Characteristics of different study designs are summarised in
the table.
Optimum populations of patients
The power of placebos can be conceptualised as the mind’s
healing power. The patient’s “belief” that he or she may be
receiving a beneficial treatment is the factor thought to
determine the placebo effect.1 The simple act of taking a pill,
or of having a sham operation, can be regarded only as a
trigger of the placebo response.
Thus, the patient should meet three requirements if the
researcher is to have the best chance of detecting a placebo
effect: consciousness must be intact;16 mental faculties must
be preserved; and some sort of pathological dysfunction, of
moderate intensity, should be present (ie, the individual has
to be ill). Patients who are comatose, who have severe
mental problems, or whose medical condition is mild, are
unlikely to be good candidates for investigation of the
placebo effect. This does not necessarily mean that the
placebo effect might not be present in such situations;
rather, that our ability to detect it may be substantially
reduced.
Whether children are a suitable population for
investigation of the placebo effect is also debatable. Because
the placebo effect, in our view, depends on expectation, and
that is in turn dependent on many factors, including
experience, children may not manifest the effect to the same
degree as adults.
The importance of study design
Although long hypothesised,5,6,8,13,19,20 only very recently has
evidence been obtained (in Parkinson’s disease) that the
placebo effect is likely to be related to the expectation of
benefit.21 This finding brings out a subtle problem inherent
in many studies aimed at investigating the placebo effect.
Many investigators have assumed that a study design with
only two groups—a placebo group and an untreated
THE LANCET Neurology Vol 1 June 2002
http://neurology.thelancet.com
For personal use. Only reproduce with permission from The Lancet Publishing Group.
The placebo effect in neurological disorders
group—should be ideal for detecting and quantifying the
placebo effect.14,16,18 Paradoxically, however, with this study
design, and adequate informed consent, neither of the two
groups will expect any benefit from the experiment and,
consequently, no full placebo intervention can be
evaluated. One group may receive a pill or some other sort
of placebo treatment, but the pill is devoid of any real
placebo power.
Another approach is the three-group study,22 in which
patients are randomly assigned to three groups: one
receiving the active treatment (in most studies a new
treatment); one receiving the placebo; and one that is
untreated. Some investigators have proposed that
comparison of the first two groups estimates the degree to
which the active treatment differs from the placebo, whereas
comparison of the second and third groups estimates the
placebo effect.18,22 However, even in this study design, the
patients’ expectation of benefit may be too low, because a
fully informed patient may realise that there is only a one in
three chance of getting some benefit. After randomisation,
the degree of expectation may decrease in those patients
included in the untreated group. Arguably, this change
should make detection of the placebo effect easier.
However, patients with such low a priori expectation of
benefit might be expected not to fluctuate much in their
clinical status after being allocated to the untreated group.
In addition, patients who volunteer for studies with low
probability of benefit may have particularly low
expectations, and may therefore not be representative of the
population as a whole.
Unsurprisingly, many studies with these designs (table)
have failed to demonstrate a placebo effect.18 This observation is
interesting in itself, however, because it shows that the simple
act of being exposed to a placebo is not enough to provide
benefit to the patient. Several letters have been published in
response to the claim by Hrobjartsson and Gotzsche that
placebos are powerless (see N Engl J Med 2001; 345: 1276–78).
Many of the correspondents independently expressed
arguments similar to ours about the relation between
expectation of benefit and detection of the placebo effect.
RCTs in which several active drugs (or different effective
doses of an active drug) are compared with a placebo are
likely to evoke a much higher expectation of benefit (and
consequently a greater placebo effect) than those in which
only a single dose of one active drug is used. In such RCTs
with many treatment options, each with different groups of
patients, the inclusion of an untreated group could
potentially inflate the placebo effect if, after randomisation,
the untreated group experienced a negative effect (ie, decline
below baseline; table).
Psychological characteristics of the patient
The psychological characteristics of the patients included in
the study can be another source of confounding. As
explained earlier, patients with no expectation of benefit are
not likely to manifest a placebo effect. Another factor to
consider is the patient’s knowledge—about his or her
disease, the efficacy of the available active drugs, and the
potential for placebos to affect the particular disease.
THE LANCET Neurology Vol 1 June 2002
Review
The patient’s knowledge of whether he or she may be
receiving a placebo during the study might be relevant to the
placebo response. For example, the placebo effect may be
greater in patients who have not been informed that they
might receive a placebo during the study.20 Thus, from a
technical point of view, the best way to detect a placebo
effect may be deliberately not to inform the patients that
they may be receiving an inactive treatment; this approach
would in most circumstances be unethical.
Allocation concealment
Apart from randomisation (ie, patients’ random allocation
to the active drug or placebo group), another requisite
crucial to any RCT is the double-blind strategy (ie, neither
patients nor examiners should know patients’ allocations).15
To preserve the masking of the study, the a priori probability
of drug-induced side-effects has to be similarly distributed
among the groups. When this condition is not met
(eg, chemotherapy produces characteristic side-effects), the
results can be compromised. In this case, both the
investigator and the patient can see through the “masking”,
which results in measurement and reporting biases.1 For
example, patients who believe they are not receiving the
active drug may actually decline below baseline (negative
placebo effect). In addition, the patient’s knowledge of the
treatment he or she is receiving will surely alter his or her
expectations.13
Pain, depression, and Parkinson’s disease
The literature on the placebo effect is mostly based on
standard two-group RCTs, in which changes in the placebo
group with respect to baseline are entirely attributed to the
placebo effect. As previously discussed, such changes are,
however, likely to be confounded by other factors
(eg, regression to the mean). Nevertheless, three
neurological disorders in which a prominent placebo effect
has been repeatedly reported are pain,23 depression,24 and
Parkinson’s disease.25–28 These three disorders are sometimes
associated.29 Interestingly, all three are associated with
dysfunction of neurotransmitters and neuropeptides in the
CNS,30 and may therefore be ideal candidates for exploring
the mechanisms underlying the placebo effect.
Placebo effect and pain
The placebo effect in pain associated with a great variety of
medical disorders is powerful and consistent.3,18,19,23 We
concur with Turner and colleagues23, who concluded that
physicians should use this effect to their (and their patients’)
advantage. Many neurological conditions are associated with
pain, and the placebo effect is applicable to all of them. For
example, de Craen and colleagues’ meta-analysis showed a
placebo effect in 26–32% of patients with migraine.11 The
route of placebo administration influenced the placebo
effect, probably as a consequence of the expected treatment
potency. In addition to the clinical relevance of placebos in
pain disorders, experimental observations on placebo
analgesia provided the first evidence for a biochemical
mechanism underlying the placebo effect.19 This original
observation, that placebo analgesia can be blocked by
http://neurology.thelancet.com
87
For personal use. Only reproduce with permission from The Lancet Publishing Group.
Review
The placebo effect in neurological disorders
disease.25–28 This effect is biochemically
mediated by the activation of the
damaged system (ie, the nigrostriatal
dopaminergic pathway), which leads to
the release of dopamine in the striatum
(figure 1).21 The biochemical placebo
effect in Parkinson’s disease is as
powerful as the effect of an active drug
(apomorphine),21 and also similar in
magnitude to the effect of amphetamine
in healthy people and patients
with schizophrenia.35 The dopaminergic
projection to the nucleus accumbens,
a region especially involved in
reward-related mechanisms,36,37 is also
susceptible to the placebo effect in
Parkinson’s (unpublished observation).
Figure 1. [ C]raclopride PET scans of a patient with Parkinson’s disease. The lower radioactivity
This observation suggests that activation
observed in the striatum after placebo (saline injection) reflects increased occupancy of striatal D
of the dopamine system may also
receptors by dopamine (ie, placebo-induced dopamine release).
mediate the placebo effect in other
naloxone, suggested that placebos can induce the release of medical conditions is unknown. For example, although
endogenous opioids. This theory has gained further support placebo-induced pain relief could be due to the release of
from the recent observation that placebo analgesia is endogenous opioid substances,19,23 Zubieta and colleagues38
associated with patterns of cerebral-blood-flow activation reported that sustained pain induces release of endogenous
(particularly in the rostral anterior cingulate cortex) similar opioids in several cortical and subcortical areas that are also
to those seen after injection of an active opioid.31 Similar known to receive dopaminergic projections. This finding
changes are seen in association with hypnosis-induced suggests that dopamine release may also be involved in the
analgesia.32
placebo effect encountered in several pain disorders
(figure 2), and may, in fact, be a common process in many
medical disorders susceptible to the placebo effect.
Placebo effect and depression
Depression is also susceptible to benefit from placebo
There may additionally be a negative interaction between
interventions. Although Hrobjartsson and Gotzsche18 failed the effect of the active drug and the placebo effect in
to detect a placebo effect in depression in their meta- Parkinson’s disease—as the placebo effect increases, the
analysis, they regarded depression as a binary variable,
which must have reduced the power of the study. Moreover,
“The reward circuitry”
they limited their analysis to studies that included an
untreated group (ie, study designs that underestimate the
Limbic cortex/hippocampus
placebo effect; table).
Mediodorsal
In trials of antidepressants, some 35% of patients
nucleus
24
receiving placebo show improvement. Some authors
(thalamus)
suggest that the overall effect is even higher. For example,
Kirsch and Sapirstein33 concluded from their meta-analysis
Nucleus
Pallidum (ventral)
of 19 trials of antidepressants that about 75% of the
accumbens
effectiveness of these drugs derives from the placebo effect.
Because such a strong placebo effect may result in a negative
Ventral
Hypothalamus
Amygdala
interaction, detection of the effect of an active treatment
tegmental area
may be very difficult in studies of depression, leading to
spurious rejection of benefit.24
Periaqueductal
There is evidence that depressed patients who respond to
grey
placebo antidepressants show a specific pattern of prefrontal
activation as detected by quantitative electroencephalography. This pattern is distinct from that seen in patients
Inhibition of pain transmission (spinal cord, thalamus)
who respond to active medication, and also different from
Inhibition of pain perception (limbic system/cortex)
the activation observed in patients who do not respond to
either placebo or active drug.34
11
2
Placebo effect and Parkinson’s disease
We have already pointed out that there is extensive clinical
evidence for a prominent placebo effect in Parkinson’s
88
Figure 2.The nucleus accumbens occupies a central position in the
dopamine-opioid connection of the reward circuitry. Apart from the
nucleus accumbens, other structures—including the ventral tegmental
area, amygdala, limbic cortex, and periaqueductal grey—could also have
a major role in the placebo effect.
THE LANCET Neurology Vol 1 June 2002
http://neurology.thelancet.com
For personal use. Only reproduce with permission from The Lancet Publishing Group.
The placebo effect in neurological disorders
effect of the active drug decreases.21,37 This evidence
strengthens the idea that the effects of placebos and active
drugs may not summate in a simple fashion. Nevertheless,
because the interaction is negative, conclusions reached
from an RCT in which the positive effect of an active
antiparkinsonian drug is found to be statistically significant
are still valid.
Whether the same applies to other disorders remains
unknown. If confirmed, such a negative interaction between
the placebo effect and the effect of the active drug may lead
to the paradoxical situation in which a researcher might
want to minimise the placebo effect in a particular patient
involved in an RCT, but maximise the placebo effect when
the same patient is seen in the usual clinical setting. Whereas
the aim in the first case would be to obtain “clean” estimates
of the effect of a new drug, the aim in the second situation
would be to maximise the clinical benefit, and perhaps lower
the risk of potential side-effects.
The evidence for a strong biochemical placebo effect in
Parkinson’s disease reinforces the view that RCTs are
necessary not only when testing new drugs, but also when
evaluating the potential efficacy of surgical interventions.39
Indeed, there is some indication that physical placebos may
be even more powerful than oral placebos.4,5,12,16,39 In a recent
study of surgery for Parkinson’s disease, the degree of
improvement at 18 months was the same after a sham
operation and after stereotaxic intrastriatal implantation of
fetal porcine ventral mesencephalic tissue.28 Another study
on the efficacy of dopaminergic-cell implants in Parkinson’s
disease, however, has not shown as prominent a placebo
effect.40 Naturally, ethical issues and consideration of the
risks and benefits inherent in the conduct of the study
will dictate whether or not a placebo treatment group
is feasible.39
By contrast with pain and depression, Parkinson’s
disease is a disorder in which the response to treatment can
be assessed directly by the examiner. This direct
measurability might allow a better evaluation of the placebo
effect. For example, a prominent placebo effect has been
reported in all domains of parkinsonian disability (although
there was a trend for a greater effect on bradykinesia and
rigidity than on tremor and gait/balance).27 We emphasise,
however, that clinical scales of motor function are also
subjective measurements. Indeed, patients may sometimes
be less prone to experience (and report) clinical changes
than clinicians to observe them.40
Although we propose that dopamine may be a common
biochemical substrate for the placebo effect, this idea does
not exclude the possibility that other neurotransmitters or
neuromodulators have a role, perhaps varying from one
disorder to another. The critical role of endogenous opioids
in placebo analgesia is one such example.
The placebo effect in other neurological disorders
Interestingly, several other neurological disorders thought
to be associated with dysfunction of dopaminergic
neurotransmission in either the nigrostriatal or the
mesocorticolimbic dopamine pathway (or both) may be
susceptible to the placebo effect. These disorders include
THE LANCET Neurology Vol 1 June 2002
Review
dystonia,41 tremors,42 tics/Tourette’s syndrome,43–45 tardive
akathisia,46,47 tardive dyskinesia,48,49 restless legs syndrome,50
obsessive-compulsive behaviour,45,51,52 and panic disorder.53
Definite conclusions about the mechanism of the placebo
effect are hard to reach from these studies, however,
because many of the disorders are heterogeneous in nature
(eg, dystonia) and in most of the conditions there is
uncertainty about whether the underlying pathophysiology
is due to an increased or decreased dopaminergic state or
even a combination of the two (eg, decreased activity in
one dopaminergic pathway and increased activity in
another). In addition, some of these results are based
on single observations or studies with insufficient
statistical power.
Placebos to treat drug addiction
Placebos might also have a role in substitution strategies for
the treatment of drug addiction,37 a disorder that seems to be
related to dopamine-related reward mechanisms.54–56 In fact,
there is emerging evidence to suggest that a variety of
behavioural disorders may be related to what has been called
the “reward deficiency syndrome”.57 Individuals with this
syndrome search for “unnatural rewards” to compensate
their intrinsic reward deficit. In other words, individuals
with the reward deficiency syndrome would be at risk for
abuse of drugs such as alcohol, nicotine, cocaine,
amphetamine, and heroin; would tend to show obsessivecompulsive behaviours, that might lead to abnormal
gambling, eating, and sex; and would be prone to practise
high-risk behaviours. For example, a substantial proportion
(15%) of drug abusers also show obsessive-compulsive
behaviour58,59 or have a history of attention-deficit/
hyperactivity disorder (12–32%).60–62
Tourette’s syndrome
Both obsessive-compulsive and attention-deficit/hyperactivity disorders are encountered in many patients with
Tourette’s syndrome.63 As indicated, the placebo effect may
be prominent in Tourette’s syndrome, a disorder in which
dopaminergic dysfunction is widely assumed to have an
important role, although whether there is a primary hyperdopaminergic or hypodopaminergic state (or some
combination of the two) is unclear. Thus, several
observations suggest that Tourette’s syndrome may be the
result of an overactive plasma-membrane dopaminetransporter system, resulting in excessive removal of
dopamine from the synaptic cleft. In particular, patients
with this syndrome have: low concentrations of the
dopamine metabolite homovanillic acid (HVA) in their
CSF;64 increased density of dopamine D2 receptors,65 which
may be a compensatory mechanism for synaptic dopamine
deficiency; and increased density of dopamine-transporter
sites.66,67 In addition, these patients benefit from treatment
with direct dopamine agonists.68 Indirect dopamine agonists
(eg, methylphenidate) are the treatment of first choice for
patients with Tourette’s syndrome and attentiondeficit/hyperactivity disorder.63 In these patients, tics may
actually improve on methylphenidate,69 although tic
worsening has also been reported.
http://neurology.thelancet.com
89
For personal use. Only reproduce with permission from The Lancet Publishing Group.
Review
The placebo effect in neurological disorders
grouped together under the same category disorders with
different causes, prognoses, and treatments. For example,
epilepsy is a heterogeneous syndrome. There is preclinical
evidence that dopamine neurotransmission may have a
crucial role in controlling some types of seizures.77,78 This
observation opens the possibility of obtaining beneficial
effects from placebo-induced dopamine release, although
release of other neuroactive substances, such as endogenous
benzodiazepines, might contribute to the placebo effect in
this situation.
Search strategy and selection criteria
Data for this review were identified by searches of PubMed
with the keyword “placebo effect” in combination with the
appropriate term for each neurological disorder. Articles and
book chapters were also identified through searches of the
extensive files of the authors; some references were provided
by anonymous reviewers. Abstracts from meetings were
included only when they were considered particularly relevant
to the topic. Only papers published in English were reviewed.
However, all these findings are also compatible with a
hyperdopaminergic state in Tourette’s syndrome. Indeed,
this may be the prevailing view among neurologists. For
example, the primary defect could be upregulation of
dopamine D2 receptors, which could lead to decreased
dopamine release and upregulation of dopamine transporter
sites as compensatory mechanisms. The combined action of
these effects could explain the low concentrations of HVA in
CSF in patients with Tourette’s syndrome. The response to
direct dopamine agonists in this context could be explained
by their action on presynaptic dopamine receptors, which
would decrease dopamine release even further.
Immune-mediated disorders
Immune-mediated neurological disorders such as multiple
sclerosis can also display a placebo effect, the potency of
which varies among studies.70,71 Ader72 observed placeboinduced changes in blood-cell counts in eight of ten patients
with multiple sclerosis included in a protocol of
cyclophosphamide therapy. Animal models of placeboinduced immunosupression73 support the notion that the
immune system can be modulated by brain processes, which
could have implications for the role of placebos in other
disorders associated with altered immunological
functioning.74 Neurotransmitters, dopamine in particular,75
have been implicated in neuroimmunomodulation. The
nigrostriatal and mesolimbic dopamine pathways, as well as
other central pathways (eg, tuberoinfundibular) and
peripheral dopamine systems may influence the immune
responses. Indeed, direct dopamine agonists seem to have a
beneficial effect in several autoimmune diseases, including
rheumatoid arthritis and systemic lupus erythematosus.76
The recent meta-analysis by Hrobjartsson and Gotzsche18
found no evidence for a placebo effect in several disorders
associated with a risk of vascular events, as well as dementia,
epilepsy, schizophrenia, and other disorders such as anxiety
and insomnia. Apart from the considerations mentioned
earlier about the optimum criteria for investigating the
placebo effect, an additional problem is that the authors
References
2
3
4
90
Altman DG. Practical statistics for medical research.
London: Chapman & Hall, 1991.
Fisher LD, van Belle G. Biostatistics: a methodology
for the health sciences. New York: John Wiley &
Sons, 1993.
Beecher HK. The powerful placebo. JAMA 1955;
159: 1602–06.
Shapiro AK, Shapiro E. The placebo: is it much ado
Converging lines of evidence indicate that the placebo effect
can be very powerful in neurological disorders such as pain,
depression, and Parkinson’s disease. We have shown that the
biochemical substrate of the placebo effect in Parkinson’s
disease is the release of dopamine in the striatum.21 The
placebo effect in other medical disorders may also be
mediated, at least in part, by dopamine release.
On the basis of previous studies and our own results, we
propose that the neural circuits involved in reward-related
mechanisms, and in particular the dopaminergic system,
have a critical role in the placebo effect. This idea suggests
that placebos could also be used in long-term substitution
strategies for the treatment of drug addiction.37 Although the
potential for placebo addiction exists,37 a reduction in the use
of illicit drugs would represent a major achievement in terms
of health and social safety.
Finally, the expectation theory of the placebo effect has
major implications for the future design of powerful placebo
studies. Studies aimed at investigating the power of placebos
should ensure that certain conditions are met in order to
optimise the expression of the placebo effect. We hope that
the points addressed here will assist in the design of better
studies in the future. The placebo effect can no longer be
thought of as a nuisance that we have to combat and
annihilate. Instead, we should learn to maximise the placebo
effect inherent in any active drug that we give to a patient,
and apply this knowledge to our clinical practice.
Authors’ contributions
All authors contributed equally to the manuscript.
Conflict of interest
Other disorders
1
Conclusions
5
6
We have no conflicts of interest.
Role of the funding source
The authors are supported by the Canadian Institutes of Health
Research, the National Parkinson Foundation (Miami, FL, USA), the
British Columbia Health Research Foundation (Canada; RF-F), the
Pacific Parkinson’s Research Institute (Vancouver, BC, Canada; RF-F),
and the Canada Research Chairs program (AJS). None of these funding
bodies had a role in the preparation of the manuscript or in the
decision to submit it for publication.
about nothing? In: Harrington A, ed. The placebo
effect: an interdisciplinary exploration. Cambridge,
Massachusetts: Harvard University Press, 1997:
12–36.
Brody H. Placebos and the philosophy of medicine:
clinical, conceptual, and ethical issues. Chicago: The
University of Chicago Press, 1980.
Wolf S. The pharmacology of placebos. Pharmacol
Rev 1959; 11: 689–704.
7
8
Morris DB. Placebo, pain, and belief: a biocultural
model. In: Harrington A, ed. The placebo effect: an
interdisciplinary exploration. Cambridge,
Massachusetts: Harvard University Press, 1997:
187–207.
Evans FJ. Expectancy, therapeutic instructions, and
the placebo response. In: White L, Tursky B,
Schwartz GE, eds. Placebo: theory, research, and
mechanisms. New York: Guilford Press, 1985:
215–28.
THE LANCET Neurology Vol 1 June 2002
http://neurology.thelancet.com
For personal use. Only reproduce with permission from The Lancet Publishing Group.
The placebo effect in neurological disorders
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Brody H. The doctor as therapeutic agent: a placebo
effect research agenda. In: Harrington A, ed. The
placebo effect: an interdisciplinary exploration.
Cambridge, Massachusetts: Harvard University
Press, 1997: 77–92.
Moerman DE. Cultural variations in the placebo
effect: ulcers, anxiety, and blood pressure.
Med Anthropol Q 2000; 14: 51–72.
de Craen AJ, Tijssen JG, de Gans J, Kleijnen J.
Placebo effect in the acute treatment of migraine:
subcutaneous placebos are better than oral placebos.
J Neurol 2000; 247: 183–88.
Kaptchuk TJ, Goldman P, Stone DA, Stason WB. Do
medical devices have enhanced placebo effects?
J Clin Epidemiol 2000; 53: 786–92.
Kirsch I. Specifying nonspecifics: psychological
mechanisms of placebo effects. In: Harrington A, ed.
The placebo effect: an interdisciplinary exploration.
Cambridge, Massachusetts: Harvard University
Press, 1997: 166–86.
Kaptchuk TJ. Powerful placebo: the dark side of the
randomised controlled trial. Lancet 1998; 351:
1722–25.
Fleiss JL. The design and analysis of clinical
experiments. New York: John Wiley & Sons, 1986.
Ernst E, Resch KL. Concept of true and perceived
placebo effects. BMJ 1995; 311: 551–53.
Kleijnen J, de Craen AJM, van Everdingen J, Krol L.
Placebo effect in double-blind clinical trials: a review
of interactions with medications. Lancet 1994; 344:
1347–49.
Hrobjartsson A, Gotzsche PC. Is the placebo
powerless? An analysis of clinical trials comparing
placebo with no treatment. N Engl J Med 2001; 344:
1594–602.
Levine JD, Gordon NC, Fields HL. The mechanism
of placebo analgesia. Lancet 1978; 2: 654–57.
Kaptchuk TJ. The double-blind, randomized,
placebo-controlled trial: gold standard or golden
calf? J Clin Epidemiol 2001; 54: 541–49.
de la Fuente-Fernández R, Ruth TJ, Sossi V,
Schulzer M, Calne DB, Stoessl AJ. Expectation and
dopamine release: mechanism of the placebo effect
in Parkinson’s disease. Science 2001; 293: 1164–66.
Rosenthal R. Designing, analyzing, interpreting, and
summarizing placebo studies. In: White L, Tursky B,
Schwartz GE, eds. Placebo: theory, research, and
mechanisms. New York: Guilford Press, 1985:
110–36.
Turner JA, Deyo RA, Loeser JD, von Korff M,
Fordyce WE. The importance of placebo effects in
pain treatment and research. JAMA 1994; 271:
1609–14.
Enserink M. Can the placebo be the cure? Science
1999; 284: 238–40.
Diamond SG, Markham CH, Treciokas LJ. Doubleblind trial of pergolide for Parkinson’s disease.
Neurology 1985; 35: 291–95.
Shetty N, Friedman JH, Kieburtz K, Marshall FJ,
Oakes D, and the Parkinson Study Group. The
placebo response in Parkinson’s disease. Clin
Neuropharmacol 1999; 22: 207–12.
Goetz CG, Leurgans S, Raman R, Stebbins GT.
Objective changes in motor function during placebo
treatment in Parkinson’s disease. Neurology 2000; 54:
710–14.
Watts RL, Freeman TB, Hauser RA, et al. A doubleblind, randomized, controlled, multicenter clinical
trial of the safety and efficacy of stereotaxic
intrastriatal implantation of fetal porcine ventral
mesencephalic tissue (Neurocelltm-PD) vs imitation
surgery in patients with Parkinson’s disease (PD).
Parkinsonism Relat Disord 2001; 7: S87.
Rajput AJ. Clinical features and natural history of
Parkinson’s disease (special consideration of aging).
In: Calne DB, ed. Neurodegenerative Diseases.
Philadelphia: WB Saunders, 1994: 555–71.
Cooper JR, Bloom FE, Roth RH. The biochemical
basis of neuropharmacology, 7th edn. New York:
Oxford University Press, 1996.
Petrovic P, Kalso E, Petersson KM, Ingvar M.
Placebo and Opioid Analgesia—Imaging a shared
neuronal network. Science 2002; 295: 1737–40.
Faymonville ME, Laureys S, Degueldre C, et al.
Neural mechanisms of antinociceptive effects of
hypnosis. Anesthesiology 2000; 92: 1257–67.
Kirsch I, Sapirstein G. Listening to prozac but
hearing placebo: a meta-analysis of antidepressant
THE LANCET Neurology Vol 1 June 2002
medication. Prevention & Treatment 1998; 1 (June
26; http://www.journals.apa.org/prevention/volume1/
pre0010002a.html)
34 Leuchter AF, Cook IA, Witte EA, Morgan M,
Abrams M. Changes in brain function of depressed
subjects during treatment with placebo.
Am J Psychiatry 2002; 159: 122–29.
35 Breier A, Su T-P, Saunders R, et al. Schizophrenia is
associated with elevated amphetamine-induced
synaptic dopamine concentrations: evidence from a
novel positron emission tomography method.
Proc Natl Acad Sci USA 1997; 94: 2569–74.
36 Schultz W. Predictive reward signal of dopamine
neurons. J Neurophysiol 1998; 80: 1–27.
37 de la Fuente-Fernández R, Stoessl AJ. The placebo
effect in Parkinson’s disease. Trends Neurosci 2002;
25: 299–303.
38 Zubieta J-K, Smith YR, Bueller JA, et al. Regional mu
opioid receptor regulation of sensory and affective
dimensions of pain. Science 2001; 293: 311–15.
39 Freeman TB, Vawter DE, Leaverton PE, et al. Use of
placebo surgery in controlled trials of a cellularbased therapy for Parkinson’s disease. N Engl J Med
1999; 341: 988–92.
40 Freed CR, Greene PE, Breeze RE, et al.
Transplantation of embryonic dopamine neurons
for severe Parkinson’s disease. N Engl J Med 2001;
344: 710–19.
41 van Hilten JJ, Hoff JI, Thang MC, van de Meerakker
MM, Voormolen JH, Delhaas EM. Clinimetric issues
of screening for responsiveness to intrathecal
baclofen in dystonia. J Neural Transm 1999; 106:
931–41.
42 Henderson JM, Ghika JA, Van Melle G, Haller E,
Einstein R. Botulinum toxin A in non-dystonic
tremors. Eur Neurol 1996; 36: 29–35.
43 Goetz CG, Tanner CM, Wilson RS, Carroll VS,
Como PG, Shannon KM. Clonidine and Gilles de la
Tourette’s syndrome: double-blind study using
objective rating methods. Ann Neurol 1987; 21:
307–10.
44 Leckman JF, Hardin MT, Riddle MA, Stevenson J,
Ort SI, Cohen DJ. Clonidine treatment of Gilles de la
Tourette’s syndrome. Arch Gen Psychiatry 1991; 48:
324–28.
45 Kurlan R, Como PG, Deeley C, McDermott M,
McDermott MP. A pilot controlled study of
fluoxetine for obsessive-compulsive symptoms in
children with Tourette’s syndrome. Clin
Neuropharmacol 1993; 16: 167–72.
46 Sachdev P, Loneragan C. Intravenous benztropine
and propanolol challenges in tardive akathisia.
Psychopharmacology (Berl) 1993; 113: 119–22.
47 Remington G, Fornazzari L, Sethna R. Placebo
response in tardive akathisia. Can J Psychiatry 1993;
38: 248–50.
48 Sommer BR, Cohen BM, Satlin A, Cole JO,
Jandorf L, Dorsey F. Changes in tardive dyskinesia
symptoms in elderly patients treated with
ganglioside GM1 or placebo. J Geriatr Psychiatry
Neurol 1994; 7: 234–37.
49 Chouinard G. Effects of risperidone in tardive
dyskinesia: an analysis of the Canadian multicenter
risperidone study. J Clin Psychopharmacol 1995; 15
(suppl 1): 36S–44S.
50 Trenkwalder C, Brandenburg U, Hundemer H-P,
et al. A randomized long-term placebo-controlled
multicenter trial of pergolide in the treatment of
RLS: the PEARLS-Study. Parkinsonism Relat Disord
2001; 7 (suppl): S129.
51 Berman I, Sapers BL, Chang HH, Losonczy MF,
Schmildler J, Green AI. Treatment of obsessivecompulsive symptoms in schizophrenic patients
with clomipramine. J Clin Psychopharmacol 1995;
15: 206–10.
52 Jenike MA, Baer L, Minichiello WE, Rauch SL,
Buttolph ML. Placebo-controlled trial of fluoxetine
and phenelzine for obsessive-compulsive disorder.
Am J Psychiatry 1997; 154: 1261–64.
53 Rosenberg NK, Mellergard M, Rosenberg R, Beck P,
Ottosson JO. Characteristics of panic disorder
patients responding to placebo. Acta Psychiatr Scand
Suppl 1991; 365: 33–38.
54 Fibiger HC, Phillips AG. Reward, motivation and
cognition: psychobiology of the mesotelencephalic
dopamine systems. In: Bloom FE, Geiger SR, eds.
Handbook of physiology. The nervous system. Vol.
4: Intrinsic regulatory systems of the brain. Bethesda,
MD: American Physiological Society, 1986: 647–75.
http://neurology.thelancet.com
Review
55 Wise RA, Rompre PP. Brain dopamine and reward.
Annu Rev Psychol 1989; 40: 191–225.
56 Robinson TE, Berridge KC. The neural basis of drug
craving: an incentive-sensitization theory of
addiction. Brain Res Rev 1993; 18: 247–91.
57 Comings DE, Blum K. Reward deficiency syndrome:
genetic aspects of behavioral disorders. Prog Brain
Res 2000; 126: 325–41.
58 Kranzler HR, Satel S, Apter A. Personality disorders
and associated features in cocaine-dependent
inpatients. Compr Psychiatry 1994; 35: 335–40.
59 Skinstad AH, Swain A. Comorbidity in a clinical
sample of substance abusers. Am J Drug Alcohol
Abuse 2001; 27: 45–64.
60 Levin FR, Evans SM, Kleber HD. Prevalence of adult
attention-deficit hyperactivity disorder among
cocaine abusers seeking treatment. Drug Alcohol
Depend 1998; 52: 15–25.
61 Clure C, Brady KT, Saladin ME, Johnson D, Waid R,
Rittenbury M. Attention-deficit/hyperactivity
disorder and substance use: symptom pattern and
drug choice. Am J Drug Alcohol Abuse 1999; 25:
441–48.
62 Schubiner H, Tzelepis A, Milberger S, et al.
Prevalence of attention-deficit/hyperactivity disorder
and conduct disorder among substance abusers.
J Clin Psychiatry 2000; 61: 244–51.
63 Tolosa E, Jankovic J. Tics and Tourette’s syndrome.
In: Jankovic J, Tolosa E, eds. Parkinson’s disease and
movement disorders, 3rd edn. Baltimore: Williams
& Wilkins, 1998: 491–12.
64 Singer HS, Butler IJ, Tune LE, Seifert WE Jr,
Coyle JT. Dopaminergic dysfunction in Tourette
syndrome. Ann Neurol 1982; 12: 361–66.
65 Wolf SS, Jones DW, Knable MB, et al. Tourette
syndrome: prediction of phenotypic variation in
monozygotic twins by caudate nucleus D2 receptor
binding. Science 1996; 273: 1225–27.
66 Singer HS, Hahn IH, Moran TH. Abnormal
dopamine uptake sites in postmortem striatum from
patients with Tourette’s syndrome. Ann Neurol 1991;
30: 558–62.
67 Malison RT, McDougle CJ, van Dyck CH, et al.
[123I]beta-CIT SPECT imaging of striatal dopamine
transporter binding in Tourette’s disorder.
Am J Psychiatry 1995; 152: 1359–61.
68 Gilbert DL, Sethuraman G, Sine L, Peters S,
Sallee FR. Tourette’s syndrome improvement with
pergolide in a randomized, double-blind, crossover
trial. Neurology 2000; 54: 1310–15.
69 Gadow KD, Nolan EE, Sverd J. Methylphenidate in
hyperactive boys with comorbid tic disorder: II,
short-term behavioral effects in school settings.
J Am Acad Child Adolesc Psychiatry 1992; 31: 462–71.
70 Noseworthy JH, Ebers GC, Vandervoort MK,
Farquhar RE, Yetisir E, Roberts R. The impact of
blinding on the results of a randomized, placebocontrolled multiple sclerosis clinical trial. Neurology
1994; 44: 16–20.
71 Weinshenker BG, O’Brien PC, Petterson TM, et al.
A randomized trial of plasma exchange in acute
central nervous system inflammatory demyelinating
disease. Ann Neurol 1999; 46: 878–86.
72 Ader R. Placebo: conversations at the disciplinary
borders. In: Harrington A, ed. The placebo effect: an
interdisciplinary exploration. Cambridge,
Massachusetts: Harvard University Press, 1997:
227–28.
73 Ader R, Cohen N. Behaviorally conditioned
immunosupression and murine systemic lupus
erythematosus. Science 1982; 215: 1534–36.
74 Anisman H, Sklar LS. Stress as a moderator variable
in neoplasia. In: White L, Tursky B, Schwartz GE,
eds. Placebo: theory, research, and mechanisms.
New York: Guilford Press, 1985: 351–94.
75 Basu S, Dasgupta PS. Dopamine, a neurotransmitter,
influences the immune system. J Neuroimmunol
2000; 102: 113–24.
76 McMurray RW. Bromocriptine in rheumatic and
autoimmune diseases. Semin Arthritis Rheum 2001;
31: 21–32.
77 Deransart C, Riban V, Le B, Marescaux C,
Depaulis A. Dopamine in the striatum modulates
seizures in a genetic model of absence epilepsy in the
rat. Neuroscience 2000; 100: 335–44.
78 Rubinstein M, Cepeda C, Hurst RS, et al. Dopamine
D4 receptor-deficient mice display cortical
hyperexcitability. J Neurosci 2001; 21: 3756–63.
91
For personal use. Only reproduce with permission from The Lancet Publishing Group.