Canadian Journal of Cardiology 33 (2017) 1535e1542
Review
The Placebo Effect in Cardiology: Understanding and
Using It
Robert Sheldon, MD, PhD,a and Morwenna Opie-Moran, MA, PhDb
a
Libin Cardiovascular Institute of Alberta, University of Calgary, Calgary, Alberta, Canada
b
King’s College Hospital NHS Foundation Trust, London, United Kingdom
ABSTRACT
The placebo effect is the clinical benefit caused by interaction with a
caregiver and health care system in the absence of a biologically active
intervention and has been used successfully for millennia. The placebo
response results from the interaction of psychosocial mechanisms, hu-
man relationships, and preconceptions functioning in specific neuroan-
atomic locations with known genes and neurotransmitters. It occurs with
or without the administration of an inactive substance to deliberately
deceive patients. Our purpose is to review the history, benefits, and
mechanisms of the placebo effect. The placebo response results from
classic conditioning and positive expectations about outcome expressed
by the caregiver. The outcomes are usually symptoms such as pain
rather than biological outcomes such as death, and the powerful pla-
cebo may account for more than half the effect of treatment in many
situations. The placebo effect results from activation of opioid, canna-
binoid, and dopaminergic pathways involved in reward, expectancy,
conditioning, and pain modulation. Eleven specific anatomic features in
the brain identified by positron emission tomography and magnetic
The time has come to re-evaluate the role of placebo in car-
diovascular medicine. It is now clear that the placebo effect is
a complex interaction of psychosocial mechanisms, human
relationships, and preconceptions, operating through specific
neuroanatomic locations, neurotransmitters, and alleles of
specific genes. Most efforts in clinical trials have focused on
minimizing the impact on study power of this mix of patients,
practitioners, practices, beliefs, treatment characteristics, and
interactional styles. However, these factors may be central to
patient improvement, and the placebo can be used for clinical
benefit.
The Provenance of Placebo
The placebo effect may be defined as a clinical benefit
elicited by interaction with a caregiver or health care system in
the presence or absence of a biologically inactive intervention.
Received for publication September 20, 2017. Accepted September 28, 2017.
Corresponding author: Dr Robert Sheldon, University of Calgary, 3280
Hospital Dr NW, Calgary, Alberta T2N 4Z6, Canada. Tel.: þ1-403-220-
8191; fax: þ1-403-210-9350.
E-mail: sheldon@ucalgary.ca
See page 1541 for disclosure information.
https://doi.org/10.1016/j.cjca.2017.09.017
0828-282X/Ó 2017 Canadian Cardiovascular Society. Published by Elsevier Inc. All rights reserved.

RESUM 
E
Par effet placebo, on entend le bienfait clinique de coulant de l’inter-
action avec un professionnel de la sante  et le système de soins de
sante en l’absence d’une intervention biologiquement active, phe no-
mène ayant e  te
 utilise
 avec succès depuis des mille naires. La re
ponse
placebo re sulte de l’interaction de me canismes psychosociaux, de
relations humaines et de pre conceptions fonctionnant dans des zones
neuro-anatomiques pre cises sur la base de gènes et de neuro-
transmetteurs connus. Elle survient avec ou sans l’administration
d’une substance inactive visant à de libe
re
ment tromper le patient.
Notre objectif est de revoir l’histoire, les bienfaits et les me canismes
de l’effet placebo. La re ponse placebo re sulte d’un conditionnement
classique et d’une attente positive quant aux re sultats exprimes par le
fournisseur de soins. Les re sultats touchent habituellement des
symptômes, comme la douleur, plutôt que des issues biologiques,
comme la mort, et la puissance du placebo peut compter pour plus de
la moitie de l’effet du traitement dans de nombreuses situations.
L’effet placebo re sulte de l’activation de voies opioïdes, cannabinoïdes
It has been used for millennia (Table 1), but the term “pla-
cebo” is less than a thousand years old.2 In medieval Europe,
funerals included chanting prayers for the dead. One of these,
“placebo domino in regione vivorum” meant “I will please the
Lord in the land of the living.” After the funeral, the
mourning family gave gifts to the chanting mourners. Later, as
plagues swept through Europe, the land of the living became
less populated, and professional mourners arose. They sang
the chant for gifts, without actually mourning, and became
known as placebosd“falsely pleasing.”
In the 16th century, priests exorcised illness with holy
water and by wielding fragments of the ostensible true cross.2
Many individuals with a variety of neurologic diseases, and
probably some with conversion syndromes, were apparently
cured. The difficulty was to explain nonresponders. This
caused political problems for Henry IV of France, who struck
a commission to resolve the issue.2 Single-blind studies
compared false fragments and false holy water to true icons
and holy water. Some responded to the false interventions,
and these were deemed to be placebo responders. Michel de
Montaigne in 1580 suggested that at times physicians pro-
vided “false promises. and their fraudulent concoctions,”
and that medicine’s efficacy might in part arise in the patient’s
imagination.3
1536
resonance imaging are involved. Polymorphisms in the structural
genes for catecholamine O-methyltransferase and fatty acid amide
oxidase significantly influence the placebo response. The placebo ef-
fect may be important in symptom suppression in angina, paroxysmal
atrial fibrillation, and congestive heart failure. In the absence of
deliberate deception, there are no ethical issues and given its potency,
the time has come to consider how best to use the placebo in clinical
practice.
In the 1780s, Mesmer proposed that an invisible substance
in objects could permit them to both cause and cure disease.
In 1784, Louis XVI commissioned the liberal polymath
Benjamin Franklin and the father of chemistry Antoine Lav-
oisier to prove or disprove whether the force existed.4 They
used a single-blind, two-step, cross-over design and could find
no evidence for the existence of the substance. By the
beginning of the 19th century, the concept of the role of the
placebo was acknowledged, the need for blinding with placebo
controls accepted, and the structure of crossover studies
developed.
The placebo now is largely understood as an inactive
substance deliberately administered with deception to a pa-
tient with the intent of inducing a beneficial response, usually
in controlled trials. However there are 3 important questions.
Does the placebo response necessarily involve a biomedical
intervention? What are the effective components of how the
placebo is administered? What are the human components
that permit or foster a successful placebo response?
We will examine the nature of the placebo, its compo-
nents, and its role in cardiovascular research and care. It can
be a potent tool in treating pain and mental health disorders as
well as inflammatory and stress-associated lifestyle illnesses. It
is time to revisit the placebo as treatment armed in the light of
new mechanistic insights.
Effects and Settings of Placebo Use
The definition of a placebo as an inactive substance
administered with deliberate deception distracts from its
mechanisms (Table 2) and potential for clinical benefit. We
prefer to define the placebo response as a clinical benefit
caused by interaction with a caregiver or health care system in
the presence or absence of a biologically inactive intervention.
The placebo effect mainly involves how individuals subjec-
tively perceive, register, and appraise symptoms.3 This in-
cludes subjective interpretation of the significance of disease
and symptoms. Reducing these negative interpretations may
relieve anxiety, depression, and subjective symptoms.
Conversely, there is no evidence that a placebo reduces
mortality or alters the pathophysiological mechanisms of the
disease. For example, placebo improves asthma symptoms
with no measureable changes in airway reactivity,5 and
placebos reduce cancer symptoms and symptomatic side ef-
fects of treatments but do not inhibit tumor growth.
Canadian Journal of Cardiology
Volume 33 2017
et dopaminergiques implique es dans la re
compense, l’expectative, le
conditionnement et la modulation de la douleur. La tomographie par
mission de positrons et l’imagerie par re
e sonance magne tique ont
permis de de celer la participation de onze composantes anatomiques
cises de l’ence
pre phale à ce phe nomène. Des polymorphismes
observes dans les gènes de structure codant pour la cate cholamine-O-
methyltransferase et l’oxydase de l’acide amine  influent de façon
importante sur la re ponse placebo. L’effet placebo peut jouer un rôle
important dans la suppression des symptômes de l’angine, de la
fibrillation auriculaire paroxystique et de l’insuffisance cardiaque
congestive. En l’absence de tromperie de libe
ree, il n’y a aucun pro-
blème e thique et, en raison de son pouvoir, le temps est venu d’e tudier
comment tirer le meilleur du placebo dans la pratique clinique.
Improvement in vasovagal syncope
Studies of vasovagal syncope have a wide range of blinding,
controls, and populations, which permit us to explore these
variables. In these studies, at least 99% of patients fainted in
the 1-2 years before study entry.6 Examples include the first
International Study of Syncope of Uncertain Etiology (ISSUE
1), which was a prospective observational study in which
participants received an implantable loop recorder, and a few
received a specific intervention.7 Most participants did not
faint in up to 15 months of follow-up. Improvement occurred
without deception or offer of clinical gain, but all patients had
been seen by specialists. Similar results were reported in the
Physical Counterpressure Manoeuvres Trial (PC-Trial) of
physical manoeuvres.8 These reduced the likelihood of a
syncope recurrence compared with no specific treatment. The
control arm consisted of lifestyle and dietary suggestions of
dubious biomedical value, yet it also showed reduced syncope
recurrences. General advice by specialists about marginally
useful lifestyle changes appears to reduce syncope recurrence.
Open-label, randomized, pacemaker studies report similar
effects. In the Vasovagal Syncope International Study (VASIS)
study,9 all patients in the control arm had fainted in the
preceding year, with a median frequency of 3 faints per year,
yet over an extended follow-up, only 61% had a recurrence of
syncope. In a meta-analysis of observational and randomized
studies, we found that all studies reported large reductions in
the probability of syncope in the control arms compared with
the period before enrollment.6 There were no significant
differences based on the degree of blinding whether the
studies were observational cohorts or randomized controlled
studies, whether the data were collected prospectively or
retrospectively, and whether the active interventions were
devices, drugs, or lifestyle aspects. In short, all populations had
large proportions of patients who improved. The common
factor was interaction of the patient with a specialist physician.
Invasive vs noninvasive placebo
One syncope trial inadvertently addressed whether an
invasive placebo was more efficacious than a medication.
Ammirati et al.10 performed a prospective open-label
controlled trial that randomized patients to medical
(b-blocker) or invasive (permanent cardiac pacing) in-
terventions. Both were later shown elsewhere to have little or
no true benefit. Both groups improved, and the patients with
Sheldon and Opie-Moran
Placebo: Mechanisms and Use
Table 1. Recent western history of the placebo
Dateline Paradigms and changes
Bronze age Europe, Middle East, Numerous descriptions of rites and
Asia potions unlikely to be biologically
ineffective
Middle ages, Europe Emergence of professional mourners
known as placebos, from Placebo
Dominae (I will please the Lord)
Renaissance, France Controlled trials of holy water and
fragments of the p in exorcisms
Enlightenment, France Franklin-Lavoisier placebo-controlled
trials of mesmerism
1931 Randomized placebo-controlled trial of
sanocrysin in tuberculosis1
1931-present Dominance of the paradigm of the
placebo effect as deliberate deception
with an inactive substance or
procedure
21st century Re-emergence of the paradigm of the
placebo effect as a neurobiological
response to human interactions
pacemakers had significantly fewer outcomes than the patients
who received b-blockers. In retrospect, this was a randomized
trial of invasive vs noninvasive placebos, and patients with the
invasive placebo had a better outcome. These observations
have driven our interests reflected in this reviewdthat in
treating vasovagal syncope, the placebo effect may be the
signal and not the noise.
Biomedical components
Placebo research in orthopaedics speaks to the importance
of the perceived potency of intervention. Arthroscopic partial
meniscectomy is frequently used to reduce symptoms from a
meniscal tear. Sihvonen et al.11 performed a randomized
controlled study of arthroscopic meniscectomy compared with
sham surgery. After 1 year, no significant difference in
symptoms existed between the 2 groups, suggesting that much
of symptom suppression was caused by nonbiomedical factors.
A clinical trial of back surgery for osteoporotic fractures found
similar results.12 In patients with hypertrophic obstructive
cardiomyopathy, permanent cardiac pacing initially appeared
to reduce symptoms, but in a double-blind randomized trial,
there was no difference in symptoms between active and
inactive pacing.13 Jonas et al.14 concluded from a systematic
review of randomized sham controlled trials that the
nonspecific effects of surgery composed about 66% of the
beneficial effect, and often the benefits of true surgery are not
significant. Similar effects are seen with tablets. The size of the
placebo effect depends on size, colour, and labelling. In 1
study, sham migraine tablets labeled as active were as effective
as active tablets that were labeled as sham.15
Effect of observer blinding
In a meta-analysis by Sud et al.,16 the likelihood of a
syncopal recurrence was reduced in patients with active pacing
vs patients with no implanted pacemaker, and this reduction
was not seen in double-blinded randomized trials. Strikingly,
in single-blinded studies in which the patients, but not the
physicians, were blinded, there was still a large reduction in
the likelihood of syncope, ie, physicians knowing the treat-
ment allocations in some way altered outcome, perhaps by
1537
ascertainment bias, involuntary communications to the pa-
tient, or the expectancy effect.
Active engagement by care provider
Taken together, these findings suggest a powerful benefi-
cial force embedded in the physician-patient relationship in-
dependent of whether a biomedical placebo of any sort is
administered.17,18 Is there a gradation within this effect?
Kelley et al.19 reported an ingenious dose-ranging study of the
impact of involvement of health care providers on the placebo
effect. Patients with irritable bowel syndrome were enrolled in
a parent study of an acute placebo-controlled infusion of a
medication to reduce symptoms. The parent study of the
biological infusion produced negative results. However in the
placebo study, patients were allocated to 1 of 3 groups: to a
waitlist, to receive educational material in a sparse and neutral
infusion room, or to a 45-minute scripted encounter with an
investigator designed to mimic a sympathetic and engaged
physician. Both written information and engagement were
significantly better than waiting, and human interactions were
markedly more effective than simple written information.
Similar results were reported with gastroesophageal reflux
disease.20 The placebo effect is real, is independent of inten-
tional deception, is a potent tool for symptom suppression,
and is embedded in human interaction.
Statistical Considerations
Syndromes with recurring events commonly have pro-
longed asymptomatic periods, and these are often ascribed to
regression to the mean, a statistical term. Recurring events can
be described with models that assume a random series of
events: either a 1-time event such as death or a random
recurrence of events in each patient over time. These occur
randomly in time and some events occur sooner and some
later, leading to the appearance of long and short asymp-
tomatic intervals and periods of more and less frequent
symptoms. The apparent improvement in symptom frequency
might simply be regression from higher-frequency periods to
the mean frequency.
Testing whether the asymptomatic intervals are part of a
random process can be performed with a simple Poisson
distribution,21 also known as a monoexponential decay curve.
These fit many classic medical survival curves. The Poisson
Table 2. Key issues in placebo research and practice
Issue Key point
The placebo: a thing or a process? Conventionally, the placebo is defined
as an inactive biomedical
intervention. How accurate is this?
Magnitude of the placebo response How much of symptomatic relief
results from the placebo and not the
active intervention?
Regression to the mean Is the placebo response simply a
statistical artifact?
Pills or procedures Is the biomedical placebo a substance
or an invasive procedure, or both?
Blinding Who needs to be blinded? Patients?
Investigators? Both?
The human touch Does the placebo response depend on
the human touch, or will
information alone suffice?
1538
Table 3. Psychology: maximizing the placebo benefit
Psychological effect Mechanism Practice
Conditioning Links present clues to Stress how new
past benefits treatment builds on
previous similar
successes
Expectancy Anticipates positive Express clear optimistic
outcomes warm hope
Treatment variables Improves conditioning Reassuring expert
and expectancy environment; be
aware that
invasiveness may
heighten placebo
benefit
Physician behaviour Human touch Be warm, engaged,
improves expectancy curious about
patient story,
empathetic,
optimistic; plan a
follow-up and
reassure
distribution is often used to test the randomness of events
with low recurrence rates. It was used to determine whether
the number of false convictions in early 19th century France
truly did cluster,22 whether the likelihood of getting kicked by
a horse in the Prussian army was clustered,23 and whether
World War II flying bombs were targeting specific areas in
London.24 All were shown to be randomly occurring events.
The use of Poisson modelling shows promise for estimating
patient-specific changes in time if they are random events. They
permit calculation of patient-specific rates and address a critical
issue in clinical trials of paroxysmal events. The Poisson distri-
bution permits the calculation of a mean frequency in data sets
bounded on either side by an event for patients who withdraw
prematurely and does not require a full observation period.25 The
Poisson distribution enables the prediction of the likelihood of
recurrences and also the statistical significance of periods without
events. This directly tests whether asymptomatic periods are truly
simply regression to the mean.
The casual invocation of “regression to the mean” can
deflect the investigator from what is truly happening. If events
or states do not fluctuate randomly, this cannot be ascribed to
regression to the mean and must reflect a response to a sto-
chastic influence. Indeed the placebo response is a combina-
tion of factors interacting in the patient’s psyche.
Psychology and Expectancy
The placebo effect is complex and highlights the inadequacy
of the conceptual duality that separates mind and body.26-29 A
seminal review in 2001 concluded that the placebo effect did not
exist,30 but it included only studies in which biologically inactive
control substances were administered. This neglected the potent
psychologically mediated mechanisms in the patient-doctor
relationship. The placebo effect most likely results from psy-
chological interactions (Table 3) between pre-existing experiences
and behaviours, the health care setting, and the interaction be-
tween patient and physician.31 These can include conditioning,26
expectancy,29 and physician behaviour.27,28,30
Consider this example. A distressed patient, just moments
after being administered multiple brightly colored analgesic
tablets is told that they are a “higher dose of those highly
Canadian Journal of Cardiology
Volume 33 2017
potent medications you responded well to yesterday” by a
kindly and sympathetic senior consultant in a comfortable
private room with framed degrees. He feels rapid relief. This
scenario includes several potent psychological placebo pro-
cesses. These include expectancy,29 (calming cognition
anticipating and contributing to pain reduction), condition-
ing,26 (rapid unconscious pre-emptive biochemical changes
associated with previous benefit), treatment variables
(including the reassuring room and impressive pills), and
physician behaviour.27,28,30 Key to the latter seems to be a
warm, friendly, reassuring, and competent style.19
Expectancy
Expectation anticipates an event to better cope with it and
can include hope, anticipation of a positive outcome, relief,
and anxiety reduction.32 Clear positive messages cause positive
expectations, and these lead to analgesia. Kirsch first hy-
pothesized that placebo effects are produced by a self-fulfilling
effect called “expectancy.”29 If a person believes that he or she
will feel different, they actually will feel different. Brain im-
aging and in situ single nerve recordings demonstrate that
similar brain regions are activated by real or imagined change
(see further on). Patients who expect to feel better have subject
response expectancy, whereas subjects whose physicians expect
them to feel better have observer response expectancy. Both
are independently important, and placebo effects are seen
when a patient does not anticipate benefit but the clinician
does.
Conditioning
Pavlovian preconditioning, such as with a beneficial
response to a biologically effective medication, can enhance
the subsequent placebo response.31 These controlled placebo
effects, observable in biochemical changes as well as in self-
reporting, occur in a wide range of settings, including anal-
gesia, Parkinson disease, response to chemotherapy, and sports
medicine.31 Similar effects on physiology are elicited by
experienced meditators as well as when social isolation is
reduced for improved health.33,34 The patient brings factors to
the interaction, including optimism, pessimism, depression,
anxiety, and anticipatory anxiety. These psychological states
can be altered by caregiver interaction and by symptoms
causing anxiety and depression that in turn worsen symptoms.
These findings underline the importance of properly con-
ducted randomized clinical trials in assessing biomedical in-
terventions and also illuminate the powerful potential of
understanding the mechanism of the beneficial effect of a
placebo.
Neurobiology of the Placebo Effect
The neurobiology of the placebo effect has been mapped in
some detail.31 Most studies have addressed pain and pain
relief. Although many of the neurotransmitters and brain
centres have clear relevance to pain, some are more tangential.
To date, 3 neurotransmitter systems have been identified:
opioids, dopamine, and cannabinoids. The opioid and
cannabinoid pathways are active in expectancy and condi-
tioning; the dopamine pathways are active in expectancy,
motivation and reward; and all 3 pathways are active in pain
Sheldon and Opie-Moran
Placebo: Mechanisms and Use
Table 4. Key genetic examples
Pathway Gene symbol Protein
39,42
Dopamine COMT Catechol-O-methyltransferase
Endocannabinoid FAAH43 Fatty acid amide hydrolase
Opioids OPRM141 Mu opioid receptor
Examples of 3 genes with corresponding biological activities and placebo effectiveness. More complete summaries are in the Genetics of the Placebo Effect
section. Other alleles and genes with evidence of matching activities and effectiveness exist and are not covered here.41
modulation. Other neurotransmitters, such as cholecysto-
kinin, serotonin, nitric oxide, and oxytocin, might also be
involved.
Placebo analgesia releases endogenous opioids and can be
blocked by naloxone.35 A second nonopioid analgesic
component can be blocked by rimonabant, the cannabinoid
CB1 receptor antagonist.36 Dopamine is involved in reward,
motivation, and expectation of reward. Pain relief may itself
be a reward, and a placebo promises the reward of pain re-
lief,31 Expectancy triggers the dopamine pathway, which then
activates the downstream analgesic response. Interestingly
much of the understanding of the involvement of dopamine
in the placebo effect comes not from pain studies but from
studies on Parkinson disease. Some of these included single
neuron recordings in conscious patients using deep brain
stimulators.
The neuroanatomy of the placebo effect has been studied
with blood oxygen leveledependent magnetic resonance im-
aging and positron emission tomography.31,37 A number of
distinct areas of the brain are involved, and these have func-
tions that are plausibly involved. For example, the opioid
pathway involves, among others, the dorsolateral prefrontal
cortex, which modulates brain attentiveness to constant pain.
Similarly the anterior insula is involved in pain transmission
and modulation of internal emotional states, the peri-
aqueductal grey area modulates the experience of pain, and the
rostral anterior cingulate cortex integrates cognition and
emotions. The dopamine pathways are active in areas such as
the nucleus accumbens and the dorsal striatum.
Genetics of the Placebo Effect
The role of neurotransmitters on the placebo response
suggest that there may be genetic influences as well
(Table 4).38,39 Not everyone is susceptible to the placebo ef-
fect,40 raising the possibility of a genetic basis. The placebo
response in patients with social anxiety is linked to attenuated
amygdala excitability,38 which itself is associated with specific
alleles encoding tyrosine hydroxylase and the serotonin re-
uptake transporter. The persuasive evidence for genetic cor-
relates of the placebo effect led Hall et al.41 to coin the term
“placebome” in an elegant and accessible review.
Several model systems provide insights into the neurobi-
ological mechanisms of the placebo response. Hall et al.39
reported the associations of specific alleles of catecholamine
O-methyltransferase on symptoms in patients with irritable
bowel syndrome in their landmark placebo dosing experi-
ment. Catecholamine O-methyltransferase (COMT) in-
activates catecholamines, including dopamine. The met/met
variant is severalfold less active, leading to higher dopamine
levels. The placebo response is higher in individuals with met/
met than in those with val/met and val/val variants. The
1539
Function Allelic activity Placebo effect
Degrades dopamine Met/met least active Met/met most effective
Degrades fatty acids Pro-pro least active Pro-pro most effective
Binds opioids Asn/asn most active Asn/asn most effective
polymorphism might also predict the nocebo response.42
Patients were given cyclosporine or placebo and warned
about the side effects of cyclosporine. Participants receiving
placebo who were homozygous for val/val had severalfold
more side effects than did those with val/met or met/met.
These allele-specific effects are seen in a number of other
diseases and in inactivation of certain drugs.41 These findings
may have practical implications if we become able to target
treatment to patients depending on their COMT
polymorphism.
Several other genes are implicated. Fatty acid amide hy-
drolase (FAAH) polymorphisms may modulate the placebo
response to pain. FAAH inactivates cannabinoids, which bind
to the CB1 receptor to suppress pain. Patients with homo-
zygous Pro129/Pro129 are more likely to respond to placebo
than are those with Thre129 alleles, but whether this is
involved in a placebo pathway or pain pathway is unclear.43
Statistically significant associations between the placebo ef-
fect and genes involved in serotonin signalling and opioid
signalling have also been reported.41 The neurobiology of
placebo establishes its biological legitimacy, and the magni-
tude of the effect may have clinical implications.
Placebo in Cardiovascular Medicine
The perils and promise of placebo are well known in car-
diology. Myocardial laser revascularization showed a prom-
ising ability to reduce angina in open-label studies, but
subsequent sham-controlled studies produced negative re-
sults.44 Similarly open-label studies of pacing for vasovagal
syncope were encouraging, but sham-controlled randomized
trials produced negative results.6 More recently, renal artery
denervation, after much initial promise, failed to be effective
in a properly controlled study.45
These and similar studies suggest that a placebo might
reduce subjective symptoms and possibly autonomic re-
sponses. Hypertension, congestive heart failure, and parox-
ysmal atrial fibrillation all cause considerable symptom
burden. Patel et al.46 in a meta-analysis of blood pressure
changes in the placebo arms of hypertension clinical trials
noted that systolic blood pressure fell a highly significant mean
of 5.9-8.8 mm Hg. The causes are conjectural, but placebo
effects are possible. Paroxysmal atrial fibrillation is treated
increasingly with intracardiac ablation techniques. However,
there is more reduction in symptoms than in documented
atrial fibrillation,47 most recurrences are asymptomatic,47 there
is similar impact on quality of life in biologically successful vs
unsuccessful ablations,48 and there is no evidence that the
procedures prevent stroke or reduce mortality. This might
result in part from damage to autonomic ganglia or possibly
to a placebo effect. Given the numerous factors associated
with the development of atrial myopathy,49 an alternative
1540
investment in research might target lifestyle modification and
development of effective placebo interventions.
Heart failure syndromes are attractive arenas for placebo
research. Sohaib et al.50 systematically reviewed symptom
improvement in controlled randomized clinical trials of car-
diac resynchronization therapy (CRT). The mean responder
rate was 51% with CRT and 35% in the control arms,
indicating that about 70% of the apparent benefit did not
result from the biological intervention. Roughly similar re-
sponses in control arms were reported for b-blockers,
angiotensin-converting enzyme inhibitors, and aldosterone
antagonists.50
Yue et al.51 reported a provocative meta-analysis of the
effects of compliance with taking a placebo on mortality in 8
randomized clinical trials of cardiovascular drugs. Patients
with good compliance with taking a placebo had significantly
lower mortality than did those with poor compliance, as well
as lower mortality than patients with poor compliance with
taking true medications. There are several explanations, but at
the least it illustrates the difficulty of interpreting randomized
clinical trial results.
Finally, vasovagal syncope has a large and documented
placebo effect and occurs in many patients in clusters,25 and
placebo research appears promising.6 The role of placebo in
relieving anxiety, and the responses of the autonomic nervous
system to anxiety and stress, hint at the promise of placebo in
cardiovascular care.
The Practice of Using Placebo
The placebo has 2 faces. The placebo effect is usually
viewed as a difficulty to be overcome with randomized
controlled trials, and if anything, the negative results of the
sham-controlled SYMPLICITY trial of renal denervation
strengthen the arguments for tightly crafted studies.45,52
However these studies usually focus on whether a biological
intervention causes a true biological effect and not whether a
treatment strategy improves outcomes compared with base-
line. Pragmatic clinical trials are 1 way of testing strategies.53
The converse might also be true: Are we missing an
important therapeutic tool, 1 that despite our unstructured
approach is already quite effective? Understandably, most
clinicians are unwilling to use placebos that require active
concealment or deception. Clinicians respect patient auton-
omy and informed consent and believe that harnessing the
power of placebo may require deliberate deception.
Counterarguments to placebo are important.54 First, pre-
scribing a deceptive placebo prevents true informed consent.
Second, for placebos to be ethical they should be effective; this
would incorporate placebo effects into evidence-based medi-
cine. A meta-analysis of 114 randomized trials that compared
biomedical placebo to no treatment suggests that these in-
terventions may not be effective.30 However, this may over-
look aspects of nominal “no treatment,” including expert
opinion, reassurance and so ondthe conditions for positive
expectancy. Third, deliberately deceptive placebo administra-
tion runs the risk or rupturing physician-patient trust, con-
travenes the idea of patient dignity, and may lead to patient
unhappiness.
All these counterarguments become irrelevant if human
interactions and expectancy, and not deception, are the true
Canadian Journal of Cardiology
Volume 33 2017
intervention. Margo recently suggested that the physician is a
placebo,28 and Lichtenberg et al.55 argued that the “placebo is
deception only for those who would reduce treatment to a
purely biomedical pursuit.” Physicians intervene in caring for
their patients in many nonbiomedical ways, including au-
thority and insight, encouragement, reassurance, optimism,
offers of help, and resolving uncertainty.
Harnessing the placebo in practice
Patients do not need to be deceived intentionally for the
placebo effect; indeed, deliberate deception is probably irrel-
evant. Bishop et al.56 recently redefined the placebo effect as
“the physiological and/or psychological changes that result
from the meaning derived by a person in a health care
setting.” An earlier systematic review by Di Blasi et al.32
concluded that physicians “who adopt a warm, friendly, and
reassuring manner are more effective than those who keep
consultations formal and do not offer reassurance.” Subse-
quently Bishop et al.56 reviewed 169 publications dealing with
pain relief, developed a taxonomy of types of presumed
effective placebos, and validated the taxonomy with experts in
the field. They identified 30 interventions thought likely to be
effective and grouped them into 5 themes. These included
patient characteristics and beliefs, practitioner characteristics
and beliefs, the health care setting, treatment characteristics,
and patient-practitioner interactions.
Components of effective placebos
Patient characteristics include selecting patients who have
not yet failed to respond to the general type of intervention,
creating positive expectancy with a positive message, and
tailoring the intervention to the patient’s sociocultural
context. Positive practitioner characteristics include their own
positive expectations and physician status manifested by ti-
tles, reputation, and displayed certificates. The health care
setting refers to characteristics such as planned follow-up,
having patients participate in research, and having patients
self-monitor their symptoms with practitioner reassessment.
Effective treatment characteristics include withdrawing inef-
fective interventions, highlighting side effects as evidence that
the treatment is effective, and maximizing treatment char-
acteristics such as high doses, invasiveness, specific colours,
pill size, and office rituals. Finally, effective interactions with
the patient include paying focused attention, obtaining a
detailed history, performing tests, and eliciting patient
preferences.
Virtually nothing is taught in medical schools or post-
graduate training about placebo responses and the effect of
physician attitudes and behaviours, yet this seems critical for
effective therapy. Once we deconstruct the true meaning of
placebo from the deceptive use of inactive biological in-
terventions, its true potential becomes evident, and at the
centre of this is physician style. Patients respond better when
physicians are optimistic and positive about recommenda-
tions. The glass should always be half full and not half empty.
The “physician as healer” brings something that “doctor by
internet” cannot. We ourselves can be forces for good in
addition to whatever biomedical treatments and benefits are
available.
Sheldon and Opie-Moran
Placebo: Mechanisms and Use
Disclosures
The authors have no conflicts of interest to disclose.
References
1. Amberson JB, McMahon BT, Pinner M. A clinical trial of sanocrysin in
pulmonary tuberculosis. AmJ Tuberculosis 1931;24:401-35.
2. Kaptchuk TJ, Kerr CE, Zanger A. Placebo controls, exorcisms, and the
devil. Lancet 2009;374:1234-5.
3. Kaptchuk TJ, Miller FG. Placebo effects in medicine. N Engl J Med
2015;373:8-9.
4. Herr HW. Franklin, Lavoisier, and Mesmer: origin of the controlled
clinical trial. Urol Oncol 2005;23:346-51.
5. Wechsler ME, Kelley JM, Boyd IO, et al. Active albuterol or placebo,
sham acupuncture, or no intervention in asthma. N Engl J Med
2011;365:119-26.
6. Pournazari P, Sahota I, Sheldon R. High remission rates in vasovagal
syncope. JACC Clin Electrophysiol 2017;3:384-92.
7. Moya A, Brignole M, Menozzi C, et al. Mechanism of syncope in pa-
tients with isolated syncope and in patients with tilt-positive syncope.
Circulation 2001;104:1261-7.
8. van Dijk N, Quartieri F, Blanc JJ, et al. Effectiveness of physical coun-
terpressure maneuvers in preventing vasovagal syncope: the Physical
Counterpressure Manoeuvres Trial (PC-Trial). J Am Coll Cardiol
2006;48:1652-7.
9. Sutton R, Brignole M, Menozzi C, et al. Dual-chamber pacing in the
treatment of neurally mediated tilt-positive cardioinhibitory syncope:
pacemaker versus no therapy: a multicenter randomized study. The
Vasovagal Syncope International Study (VASIS) Investigators. Circula-
tion 2000;102:294-9.
10. Ammirati F, Colivicchi F, Santini M. Permanent cardiac pacing versus
medical treatment for the prevention of recurrent vasovagal syncope: a
multicenter, randomized, controlled trial. Circulation 2001;104:52-7.
11. Sihvonen R, Paavola M, Malmivaara A, et al. Arthroscopic partial
meniscectomy versus sham surgery for a degenerative meniscal tear.
N Engl J Med 2013;369:2515-24.
12. Kallmes DF, Comstock BA, Heagerty PJ, et al. A randomized trial of
vertebroplasty for osteoporotic spinal fractures. N Engl J Med 2009;361:
569-79.
13. Linde C, Gadler F, Kappenberger L, Ryden L. Placebo effect of pace-
maker implantation in obstructive hypertrophic cardiomyopathy. PIC
Study Group. Pacing in cardiomyopathy. Am J Cardiol 1999;83:903-7.
14. Jonas WB, Crawford C, Colloca L, et al. To what extent are surgery and
invasive procedures effective beyond a placebo response? A systematic
review with meta-analysis of randomised, sham controlled trials. BMJ
Open 2015;5:e009655.
15. Kam-Hansen S, Jakubowski M, Kelley JM, et al. Altered placebo and
drug labeling changes the outcome of episodic migraine attacks. Sci
Transl Med 2014;6:218ra5.
16. Sud S, Massel D, Klein GJ, et al. The expectation effect and cardiac
pacing for refractory vasovagal syncope. Am J Med 2007;120:54-62.
17. Olshansky B. Placebo and nocebo in cardiovascular health: implications
for health care, research, and the doctor-patient relationship. J Am Coll
Cardiol 2007;49:415-21.
18. Thomas KB. General practice consultations: is there any point in being
positive? Br Med J (Clin Res Ed) 1987;294:1200-2.
1541
19. Kelley JM, Lembo AJ, Ablon JS, et al. Patient and practitioner influences
on the placebo effect in irritable bowel syndrome. Psychosom Med
2009;71:789-97.
20. Dossett ML, Mu L, Davis RB, et al. Patient-provider interactions affect
symptoms in gastroesophageal reflux disease: a pilot randomized, double-
blind, placebo-controlled trial. PLoS One 2015;10:e0136855.
21. Wikipedia. Poisson distribution. Available at: https://en.wikipedia.org/
wiki/Poisson_distribution. Accessed September 1, 2017.
22. Poisson SD. Recherches sur la Probabilite des jugements en matière
criminelle et en matière civile, precedees des règles generales du calcul des
probabilities. Paris, France: Bachelier, 1837.
23. von Bortkiewicz L. Das Gesetz der kleinen Zahlen [The law of small
numbers]. Leipzig, Germany: B.G. Teubner, 1898.
24. Clarke RD. An application of the Poisson distribution. J Inst Actuaries
1946;72:48.
25. Sahota IS, Pournazari P, Sheldon RS. Clusters, gaps, and randomness:
vasovagal syncope recurrence patterns. J Am Coll Cardiol Clin Electro-
phys 2017;3:1046-53.
26. Voudouris NJ, Peck CL, Coleman G. The role of conditioning and
verbal expectancy in the placebo response. Pain 1990;43:121-8.
27. Miller FG, Rosenstein DL. The nature and power of the placebo effect.
J Clin Epidemiol 2006;59:331-5.
28. Margo CE. The placebo effect. Surv Ophthalmol 1999;44:31-44.
29. Kirsch I. Response expectancy as a determinant of experience and
behavior. Am Psychol 1985;40:1189-202.
30. 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.
31. Jubb J, Bensing JM. The sweetest pill to swallow: how patient neuro-
biology can be harnessed to maximise placebo effects. Neurosci Biobehav
Rev 2013;37:2709-20.
32. Di Blasi Z, Harkness E, Ernst E, Georgiou A, Kleijnen J. Influence of
context effects on health outcomes: a systematic review. Lancet
2001;357:757-62.
33. Nyklícek I, Mommersteeg PM, Van Beugen S, Ramakers C, Van
Boxtel GJ. Mindfulness-based stress reduction and physiological activity
during acute stress: a randomized controlled trial. Health Psychol
2013;32:1110-3.
34. Holt-Lunstad J, Smith TB, Baker M, Harris T, Stephenson D. Loneli-
ness and social isolation as risk factors for mortality: a meta-analytic re-
view. Perspect Psychol Sci 2015;10:227-37.
35. Levine JD, Gordon NC, Fields HL. The mechanism of placebo analgesia.
Lancet 1978;2:654-7.
36. Benedetti F, Amanzio M, Rosato R, Blanchard C. Nonopioid placebo
analgesia is mediated by CB1 cannabinoid receptors. Nat Med 2011;17:
1228-30.
37. Lidstone SC, Stoessl AJ. Understanding the placebo effect: contributions
from neuroimaging. Mol Imaging Biol 2007;9:176-85.
38. Furmark T, Appel L, Henningsson S, et al. A link between serotonin-
related gene polymorphisms, amygdala activity, and placebo-induced
relief from social anxiety. J Neurosci 2008;28:13066-74.
39. Hall KT, Lembo AJ, Kirsch I, et al. Catechol-O-methyltransferase
val158met polymorphism predicts placebo effect in irritable bowel syn-
drome. PLoS One 2012;7:e48135.
1542
40. Levine JD, Gordon NC, Bornstein JC, Fields HL. Role of pain in pla-
cebo analgesia. Proc Natl Acad Sci U S A 1979;76:3528-31.
41. Hall KT, Loscalzo J, Kaptchuk TJ. Genetics and the placebo effect: the
placebome. Trends Mol Med 2015;21:285-94.
42. Wendt L, Albring A, Benson S, et al. Catechol-o-methyltransferase
val158met polymorphism is associated with somatosensory amplification
and nocebo responses. PLoS One 2014;9:e107665.
43. Pecina M, Martinez-Jauand M, Hodgkinson C, et al. FAAH selectively
influences placebo effects. Mol Psychiatry 2014;19:385.
44. Saririan M, Eisenberg MJ. Myocardial laser revascularization for the
treatment of end-stage coronary artery disease. J Am Coll Cardiol
2003;41:173-83.
45. Bhatt DL, Kandzari DE, O’Neill WW, et al. A controlled trial of renal
denervation for resistant hypertension. N Engl J Med 2014;370:
1393-401.
46. Patel HC, Hayward C, Ozdemir BA, et al. Magnitude of blood pressure
reduction in the placebo arms of modern hypertension trials: implications
for trials of renal denervation. Hypertension 2015;65:401-6.
47. Verma A, Champagne J, Sapp J, et al. Discerning the incidence of
symptomatic and asymptomatic episodes of atrial fibrillation before and
after catheter ablation (DISCERN AF): a prospective, multicenter study.
JAMA Intern Med 2013;173:149-56.
Canadian Journal of Cardiology
Volume 33 2017
48. Wokhlu A, Monahan KH, Hodge DO, et al. Long-term quality of life
after ablation of atrial fibrillation the impact of recurrence, symptom
relief, and placebo effect. J Am Coll Cardiol 2010;55:2308-16.
49. Goette A, Kalman JM, Aguinaga L, et al. EHRA/HRS/APHRS/SOL-
AECE expert consensus on atrial cardiomyopathies: definition, charac-
terization, and clinical implication. Heart Rhythm 2017;14:e3-40.
50. Sohaib SM, Chen Z, Whinnett ZI, et al. Meta-analysis of symptomatic
response attributable to the pacing component of cardiac resynchroni-
zation therapy. Eur J Heart Fail 2013;15:1419-28.
51. Yue Z, Cai C, Ai-Fang Y, et al. The effect of placebo adherence on
reducing cardiovascular mortality: a meta-analysis. Clin Res Cardiol
2014;103:229-35.
52. Redberg RF. Sham controls in medical device trials. N Engl J Med
2014;371:892-3.
53. Thorpe KE, Zwarenstein M, Oxman AD, et al. A pragmatic-explanatory
continuum indicator summary (PRECIS): a tool to help trial designers.
J Clin Epidemiol 2009;62:464-75.
54. Asai A, Kadooka Y. Reexamination of the ethics of placebo use in clinical
practice. Bioethics 2013;27:186-93.
55. Lichtenberg P, Heresco-Levy U, Nitzan U. The ethics of the placebo in
clinical practice. J Med Ethics 2004;30:551-4.
56. Bishop FL, Coghlan B, Geraghty AW, et al. What techniques might be
used to harness placebo effects in non-malignant pain? A literature review
and survey to develop a taxonomy. BMJ Open 2017;7:e015516.