Against the placebo effect : A personal point of view

Transcription

1 Complementary Therapies in Medicine (2013) 21, Available online at jou rnal h om epa ge: Against the placebo effect : A personal point of view Daniel E. Moerman William E Stirton Professor Emeritus of Anthropology, University of Michigan-Dearborn, 6515 Cherry Hill Road, Ypsilanti, MI 48198, United States KEYWORDS Meaning; Meaning response; Placebo; Inert treatment Summary The author reviews 10 of his favorite studies which are said to be about the placebo effect, but which, instead, show the significance of meaning in a medical context. Placebos, he argues, are inert substances which can t do anything. Yet it s clear that after the administration of such drugs, things do happen. The one (and maybe only) clear thing here is that whatever happens is not due to the placebo (that is what inert means). But placebos can be of various colors and forms which can convey compelling meaning to patients. They often represent medical treatment in compelling ways; they can be metonymic representations of the entire medical experience (a metonym is a representation where a part of something comes to represent it all, as in counting noses, where the nose represents the whole person, or a White House statement where the White House represents the Executive Branch of the US Government; here, the pill represents the whole medical experience). More precisely, they can be metonymic simulacra (a simulacrum is a sort of artificial object, like a statue rather than a man, or a placebo rather than an aspirin). Such objects are well known for their powerful abilities to contain and convey meaning; for example, a European cathedral ordinarily is constructed of thousands of metonymic simulacra, from the rose window to the altar. In this context, a placebo can repeatedly remind the patient of the medical encounter, its shadings and comforts. Placebos can convey the physicians innermost feelings about medication and treatment; and the clinician can by her simple presence enhance the effectiveness of a medical procedure (and a clinician is hardly a placebo, hardly inert). Inert placebos can help us see the human dimensions of medical treatment; but calling these things placebo effects dramatically distorts our understanding of such treatments, by focusing on the inert, and avoiding the meaningful. Think meaning response, not placebo effect Elsevier Ltd. All rights reserved. I thought this would be easy. I would pick out my 10 favorite studies, the ones I ve learned the most from over the years, and go thru them from 10 to 1. This turns out to have been Tel.: ; fax: address: dmoerman@umich.edu more difficult than I had imagined it would be. But nonetheless, interesting, and, I will argue that in most of these, the results usually make more sense if we try to determine how a meaningful interaction occurred, rather than trying to understand the effectiveness of... nothing. I will argue that there is never nothing going on here. Here s a good example: /$ see front matter 2013 Elsevier Ltd. All rights reserved.

2 126 D.E. Moerman Number 10 1 In an important study, 835 women who reported that they regularly treated headaches with over the counter analgesics were randomly placed in 4 groups: one group received unlabeled placebo, one received placebo marked with a widely advertised brand name, one of the most popular...analgesics in the United Kingdom widely available for many years and supported by extensive advertising, one received unbranded aspirin, and one received branded aspirin. They noted the amount of headache pain relief an hour after taking the pills. Results: First, aspirin was more effective than placebo. But brand name aspirin was more effective than generic aspirin, and brand name placebo was more effective than generic placebo. In particular, 55% of headaches reported by branded placebo users improved after an hour (rated 2, 3 or 4 on the scale) while only 45% of 410 headaches were reported to be that much better by unbranded placebo users ( 2 = 6.76, p <.01). Aspirin relieves headaches. But so does the knowledge that the pills you are taking are good ones, which you learned on TV. The difference here is to be attributed not to the placebo (which is, after all, inert) but to the brand name which clearly is not, enhancing the effect of both placebo and aspirin. Note that saying that this is Smith s Aspirin is not a lie if, indeed, it is Smith s aspirin. Both aspirin and placebo work better when they have a highly advertised brand name on them. That s NOT a placebo; that s meaning, something added to the tablets with WORDS. Number 9 2 Rick Gracely has described a phased experiment in which dental patients were told they would receive either placebo (which might reduce the pain of third-molar extraction, or might do nothing), naloxone (which might increase their pain, or do nothing), or the synthetic narcotic analgesic fentanyl (which might reduce their pain, or do nothing). Subjects were all recruited from the same patient stream, with consistent selection criteria by the same staff. In the first phase of the study, clinicians (but not patients) were told that because of administrative problems with the study protocol, fentanyl was not yet a possibility, yielding the PN ( Placebo Naloxone ) group; it is worth noting that fentanyl is well known in medical circles as a very powerful drug, much more potent than morphine. In the second phase, clinicians were told that, now patients might indeed receive fentanyl, yielding the PNF (Placebo Naloxone Fentanyl) group. Placebo treated patients during the first phase of the study received no relief from it, and, after an hour, their pain reports increased significantly. In the second phase of the study, placebo treated patients experienced significant pain reduction from their inert treatments. The only apparent difference between the two groups was that the clinicians knew that no one in the first group would get fentanyl while the patients in the second group might (although no one reported on here actually did; they all received only placebo). It is not at all clear how physicians elicited these effects from their patients in a double blind trial. But they did; the clinicians were clearly more impressed by fentanyl than were the patients. This study clearly shows how physician knowledge of the context in which placebos are administered can dramatically change the outcome. Number 8 3 In a landmark study in 1978, Levine and colleagues showed that pain relief brought on by prescribing a placebo could be reversed by administration of an opiate antagonist, naloxone or Narcan. The clear implication was that somehow, the brain produced endogenous opiates which led to the pain relief which was extinguished by the naloxone. In this study, students were enrolled who had impacted third molars. Following third molar extraction, patients were told (twice) that they might receive morphine, placebo, or naloxone, an opiate antagonist. Two hours following the initial anesthesia patients were told they would receive either morphine, placebo, or naloxone: 9 responded to the placebo and 14 didn t. At three hours (180 min) all these individuals were given naloxone as a second treatment. It had no appreciable effect on the nonresponders, but definitely eliminated the pain relief in the placebo responders. This was not a perfect experiment; a lot went on which I haven t described, and the paper was very controversial. But, 18 years later, Fabrizio Benedetti said of this paper it marked the date that the biology of placebo was born. 4 It is now generally recognized that this is the first study to show convincingly that inert treatment could stimulate the production of endogenous opiates in the brain. In a personal communication about this study, Howard Fields told me The first time we did this and did not have morphine as a possibility, there was no placebo effect. Once we truly blinded it, so that nobody really knew what they were getting, we started seeing robust effects from saline infusions. As in the previous study by Gracely, only when clinicians knew that patients might get morphine did patients have significant meaning responses. Number 7 4 This study by Fabrizio Benedetti was largely designed as a replication of the previous one by Levine, Gordon and Fields. In this study, subjects induced pain by squeezing on a hand exerciser with a tourniquet on the upper arm creating intense pain. When pain reports reach 7 on a scale of 10, an open injection of saline presented as a helpful pain reliever in about 6 or 8 words is given to the members of one group (see line with squares in Fig. 1); the outcome is compared to another group which receives a hidden injection of saline the same injection, but with no words in the other group (diamonds). That s the only difference between the two groups. Yet the open saline group shows a persistent decline in pain reports while the hidden infusion group shows a continued rise in pain. Let me qualify this: Does this show us that placebos have effects? No, because both groups

3 Against the placebo effect hidden saline open saline hidden naloxone hidden proglumide Figure 1 Experimental pain was reduced by open saline injection (presented as a helpful pain reducer; squares), but was not reduced by hidden saline injection (diamonds). A second injection of hidden naloxone 15 min later reduced the analgesia (triangles); this is a clear replication of the Levine, Gordon and Field study (Number 7). A second injection of hidden proglumide (an opiate enhancer; ridged circles) increases the analgesic effect of open saline. Redrawn from Ref. 4. got placebos. The difference between the two groups was words, language, meaningful utterances. Another group, given open saline, is, after 15 min, given an injection of hidden naloxone (triangles); the pain returns; this is the replication of Levine, Gordon and Fields. Then Benedetti adds another trick: after 15 min, he gives a hidden injection of proglumide, an opiate enhancer, and the pain drops even more (ridged circles). Benedetti manipulates placebo like a magician. Number 6 5,6 It is also important to note that these matters, where meaning has an influence on health and even mortality, can occur well outside the ordinary bounds of the clinic. Dr. P.D. Phillips and colleagues have shown that, in the presence of a broad range of diseases in Chinese Americans in California, those who are understood by Chinese traditions of astrology to be particularly susceptible to these conditions by virtue of the year of their birth die significantly earlier than those with the same conditions born in other years. Here are three examples from 6 or 8 which Phillips described: Chinese born in earth years, that is, years ending with 8 or 9 like 1958 or 1969 and consequently deemed by Chinese medical theory to be especially susceptible to diseases involving lumps, nodules, or tumors and who have lymphatic cancer, die, on average, 4 years sooner than Chinese with lymphatic cancer born in other years. Those with lung diseases born in metal years, years ending in 0 or 1 in Chinese theory, the lung is the organ of metal die on average 5 years younger (roughly 7% of length of life!) than those born in other years. There were no such differences found in a similar examination of the mortality of thousands of non-chinese Californians. These are very compelling examples of meaning responses. In another study, Phillips showed that Chinese-Americans and Japanese-Americans were more likely to die on the 4th day of the month than any other because 4 is an unlucky number; In Chinese, the word for four is (approximately) sì (with a falling tone) while the word for death is sĭ (with a falling-rising tone). In Fig. 2 you can see the enhancement of death due to chronic heart disease in Chinese and Japanese Californians from on the 4th day of the month. The vertical line indicates the 95% confidence interval. The lower graph represents non-asian Californians. If 13 is an unlucky number for Californians in general, it s not unlucky enough to increase the mortality rate. It is worth noting that these meanings of metal and the lung, or of earth and lumps, or of deathly fours are not notions concocted by individual patients or therapists; they are icons of a sort which permeate the language and culture of, in this case, immigrant Chinese or Japanese people and/or their American born children, to some degree or other. These relationships have nothing to do with having Asian genes, but with having Asian ways of living, thinking, behaving and being which can have significant effect on mortality. Number newly admitted neurotic outpatients at the Johns Hopkins psychology clinic were invited to participate in a study. They presented an array of symptoms; all were pretty unhappy people. They were told that during their workup, they were going to be prescribed sugar pills; that is, pills with no medicine in them at all. They were told such pills had helped many people in the past, and that the doctor thought that it would help them. They were told to take 3 a day at mealtimes. Fourteen of the patients returned in a week; on an array of measures, physician and patient scores, 13 were markedly better than a week earlier (one woman was worse; her husband had attempted suicide during the week). Placebos can help, can be deeply meaningful, even if you know that they are inert. Number Ted Kaptchuk did a replication of this trial 45 years later with 80 patients with severe irritable bowel syndrome. Patients were randomized to open label placebo tablets; they were presented as placebo pills made of an inert substance, like sugar pills, that have been shown in clinical studies to produce significant improvement in IBS symptoms through mind-body self-healing processes. The control groups received the same treatment as the placebo group but didn t get the placebo pills. The placebo group was better at midpoint and endpoint on all measures. Placebos administered without deception may be an effective treatment for IBS. Number 4 9 Irving Kirsch and his colleagues took the unprecedented step of making a Freedom of Information Act (FOIA) request of the FDA for the studies used to approve the use of SSRIs

4 128 D.E. Moerman Figure 2 Upper graph shows mortality by day of month from chronic heart disease for Chinese and Japanese Californians hospitalized patients. Lower graph shows the same mortality for white Californians. Reprinted with permission from Ref. 6. for depression. They analyzed data on 6 drugs from 38 studies with a total of 6944 patients randomized to drug or placebo treatment. Table 1 shows the results of those studies. Mean improvement with drug treatment was a drop in the Hamilton scale Table 1 Mean improvement on Hamilton score for 6 SSRIs approved for use by the FDA. Overall average improvement for Drug groups 10.01; for control groups Drug Drug group improvement Fluoxetine Paroxitene Sertraline Venlafaxine Nefazodone Citalopram Placebo group improvement score of 10 points; mean improvement with placebo treatment was a drop of 8 points. Nearly 80% of the improvement from the drug was replicated by the placebo treatment, and the difference between drug and placebo was about 2 points on the Hamilton Scale. So: Placebos can improve a LOT of depression. Number 3 10 Walsh and colleagues reviewed 75 published trials of various antidepressants: tricyclics, and SSRIs compared with placebo. The results of his study show that the effectiveness of drug treatment for depression has trended up substantially between 1981 and 2000, so that the proportion of patients responding to tricyclic antidepressants and to SSRIs had increased from about 40% to about 55%. Over the same period, the proportion of patients responding to placebo increased from about 20% to about 35%. The proportion responding was strongly correlated with the year of

5 Against the placebo effect 129 publication of the study for both drug and placebo treatment. The authors conclude that Some factor or factors associated with the level of placebo response must therefore have changed significantly during this period. Unfortunately, we were not able to identify these factors. However, the matter doesn t seem too complicated to me. Over the past generation, there has been a clear shift in consciousness among doctors, patients, friends, and, generally, everyone, to the effect that depression can be treated with drugs. This was simply not the case (or at least not broadly shared) 20 or 25 years ago. As recently as 1970, for example, Goodman and Gilman s Pharmacological Basis of Therapeutics, one of the standard reference sources, was clearly more enthusiastic about electro-convulsive therapy (ECT) than it was about treatment with imipramine or amitriptyline, which were said never to be more effective than ECT. 11 Today, while we practically never hear of ECT, we all know that drugs are effective for depression; we read it in the newspapers, in the scientific journals; we see it on TV dramas, and, in the US at least, we see it in drug company advertisements everywhere, both in professional media and on TV commercials, blogs, Twitter and Facebook. Antidepressant drugs are available in the drugstore, and, in the form of St. John s Wort, at the drug section of your local supermarket. As we change our views of the effectiveness of drugs, their effectiveness changes, as do their placebo mimics in trials. Meanings change and so do meaning responses. Placebos stay the same, always inert. Number 2 12,13 Brain imaging has had a significant effect in placebo studies, making clearer to people just what is going on; I m not convinced that the imaging studies showed us a whole lot we didn t know before, but I believe they converted a lot of people previously skeptical. Such images are indeed very compelling: I will consider only two of many that have been published, mostly since about Parkinson s disease has long been known by clinicians to be susceptible to influence by inert treatments. Imaging studies by a group from British Columbia have shown a neurological basis for this common clinical observation. Using PET scanning, the authors showed substantial increase in occupancy of D2 receptors with dopamine in the striatum after an injection of saline solution to a Parkinson s patient presented as his standard medication; the increased dopamine crowds out the radioactive dye. 12 In a somewhat more complex study, regional glucose metabolism in PET scans of fluoxetine (Prozac in the US) has been shown to overlap the metabolic pattern of placebo in depressed patients. The active regions in fluoxetine responders overlap the area where activity was evident in placebo responders. 13 Although the clinical response of drug and placebo patients was very similar in this study, drug response in brain activity was somewhat more general than placebo response. This may help to account for why it is that, while placebo treatment of depression is often very nearly as effective as is treatment with SSRIs, there is often substantially less evidence of unwanted side effects with placebo. 9 These studies Figure 3 When the clinician is present for the injection, patients report less pain. Note in particular that there are no placebos anywhere in this study, hence no placebo effects.. Reprinted with permission from Ref. 14. underscore the vital reality of meaningful treatment in serious illnesses. Number 1 14 Finally, Benedetti and his group have reported on a clinical experiment where surgery patients were treated with four different drugs appropriate to their conditions; however, half the patients received their drugs openly, with an injection by a clinician, while half received equivalent doses of the same drugs by hidden infusion through an intravenous line. One of the four groups was given tramadol. Patients receiving the medication openly, who were told they were about to receive it (the lower line in Fig. 3), reported more pain relief than those who received equivalent amounts of drugs secretly (upper line). Pain researcher Don Price, in an accompanying editorial, described this study as assessing placebo effects without placebo groups. 15 As much as I respect Don Price, this is an unfortunate use of language. There were no placebos here; both groups got tramadol. So obviously, there weren t any placebo effects. What differentiated the separate groups in this study were human interaction and words. Benedetti has replicated his open/hidden drug experiment in three other areas: diazepam in anxiety state, stimulation of the subthalmic nucleus in Parkinson s patients, and administration of beta-blocker (propranalol) or muscarinic antagonists (atropine) in healthy volunteers. In all these cases, when the treatment was given openly, it was more effective than when given secretly. 16 In his discussion of these cases, Benedetti and his colleagues write this: It is probably wrong to call placebo effect the difference between open and hidden treatments, since no placebos are given. Meaning response is perhaps

6 130 D.E. Moerman more appropriate, 17,18 in order to make it clear that the crucial factor is not so much the inert treatment per se but rather the meaning around the medical treatment... Therefore, it might be time to limit the use of the term placebo effect to those situations in which inert (dummy) medical treatments are given. However, it is worth noting that even if a placebo is given, there is no such thing as a placebo effect, since this term deflects our gaze from what is really important (the meaning and the meaning-induced expectations) and aims it at what is not (the inert pills and, in general, the inert medical treatments). 16 Conclusions I, of course, agree fully with Benedetti. A placebo, unless made by an incompetent pharmacist, is inert. That means it doesn t do anything. But we often find that things do happen to people after placebos are administered. The one thing we can know for sure is that these effects are NOT due to the placebo. But as these long time favorite studies of mine (plus others) show, the meanings of drugs or other treatments to patients, clinicians, families, friends, community, are supremely important here. It is long past time to give up on a flawed notion, the placebo effect or the placebo response. People don t respond to placebos. They respond to what placebos, drugs, clinicians, and others mean and when there are no placebos in the study, they respond to the person who brings it to them. People respond to what we know, think, and feel... People respond to what we are told, believe and know... People respond to their various cultural backgrounds... They respond to language, to caring, to culture, to community, to history. In a word, they respond to meaningful phenomena. Conflict of interest statement: none declared. Acknowledgements Special thanks to Dr. Prof. Robert Jütte and the Robert Bosch Stiftung for supporting the fascinating conference at Villa La Collina, at Lake Como, in May 2012; thanks to all the participants who vigorously discussed the ideas presented in this paper. Also special thanks to Irving Kirsch for discussion and support; and to Howard Fields for his reminiscences of his participation in one of the landmark studies considered here. For Jason Gold, my beekeeper. References 1. Branthwaite A, Cooper P. Analgesic effects of branding in treatment of headaches. British Medical Journal (Clinical Research Ed) 1981;282: Gracely RH, Dubner R, Deeter WR, Wolskee PJ. Clinicians expectations influence placebo analgesia. Lancet 1985;1: Levine JD, Gordon NC, Fields HL. The mechanism of placebo analgesia. Lancet 1978;2: Benedetti F. The opposite effects of the opiate antagonist naloxone and the cholecystokinin antagonist proglumide on placebo analgesia. Pain 1996;64: Phillips DP, Ruth TE, Wagner LM. Psychology and survival. Lancet 1993;342: Phillips DP, Liu GC, Kwok K, Jarvinen JR, Zhang W, Abramson IS. The Hound of the Baskervilles effect: natural experiment on the influence of psychological stress on timing of death. British Medical Journal 2001;323: Park LC, Covi L. Nonblind placebo trial. Archives of General Psychiatry 1965;12: Kaptchuk TJFE, Kelley JM, Sanchez MN, Kokkotou E, Singer JP, Kowalczykowski M, et al. Placebos without deception: a randomized controlled trial in irritable bowel syndrome. PLoS One 2010;5:e Kirsch I, Moore TJ, Scorboria A, Nicholls SS. The Emperor s new drugs: an analysis of antidepressant mediction data submitted to the U.S. Food and Drug Administration. Prevention and Treatment 2002;5. volume5/pre a.html 10. Walsh BT, Seidman SN, Sysko R, Gould M. Placebo response in studies of major depression: variable, substantial, and growing. Journal of the American Medical Association 2002;287: Goodman LS, Gilman A. The pharmacological basis of therapeutics. 4th ed. New York: The Macmillian Company; de La Fuente-Fernandez 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: Leuchter AF, Cook IA, Witte EA, Morgan M, Abrams M. Changes in brain function of depressed subjects during treatment with placebo. American Journal of Psychiatry 2002;159: Amanzio M, Pollo A, Maggi G, Benedetti F. Response variability to analgesics: a role for non-specific activation of endogenous opioids. Pain 2001;90: Price DD. Assessing placebo effects without placebo groups: an untapped possibility? Pain 2001;90: Benedetti F, Maggi G, Lopiano L, Lanotte M, Rainero I, Vighetti S, et al. Open versus hidden medical treatments: the patient s knowledge about a therapy affects the therapy outcome. Prevention and Treatment 2003;6:1a. 17. Moerman DE. Meaning, medicine and the placebo effect. Cambridge: Cambridge University Press; Moerman DE, Jonas WB. Deconstructing the placebo effect and finding the meaning response. Annals of Internal Medicine 2002;136:471 6.