Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
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.