J7ournal ofMedical Ethics, 1996; 22: 140-146 J Med Ethics: first published as 10.1136/jme.22.3.140 on 1 June 1996. Downloaded from

Placebos in clinical practice and research

P P De Deyn and R D'Hooge Department ofNeurology, Middelheim General Hospital and Laboratory of Neurochemistry and Behaviour, University ofAntwerp, Antwerp, Belgium

Abstract consequences of the administration of placebos The main current application ofplacebo is in clinical became known as "the placebo effect", comprising research. The term placebo effect refers to diverse non- the total ofunexplained consequences of administer- specific, desired or non-desired effects ofsubstances or ing placebos as well as active treatments. Due to, procedures and interactions between patient and amongst other things, the unpredictability ofplacebo therapist. Unpredictability of the placebo effect effects in individuals and patient groups, placebos necessitates placebo-controlled designs for most trials. came to play an irreplaceable role as "inactive" Therapeutic and diagnostic use ofplacebo is ethically controls in randomised clinical trials (RCTs). acceptable only in few well-defined cases. While The first rather isolated placebo-controlled "therapeutic" application ofplacebo almost invariably clinical trial took place in 1916 and was conducted implies deception, this is not the casefor its use in by Macht who compared the analgesic action of research. Conflicts may exist between the therapist's morphine with that of physiological saline.2 It wascopyright. Hippocratic and scientific obligations. The authors only in the 1940s and 1950s that the large-scale provide examples in neuropsychiatry, illustrating that use of placebos in clinical research emerged, simul- objective scientific data and well-considered guidelines taneously with scientific knowledge pertaining to the may solve the ethical dilemma. Placebo control might placebo effect. The unabated need for placebo- even be considered an ethical obligation but some controlled clinical trials is illustrated by several anec- provisos should be kept in mind: (a) no adequate dotes in the development of presumedly therapeutic therapyfor the disease should exist and/or (presumed) procedures. Surgical ligation of the internal http://jme.bmj.com/ active therapy should have serious side-effects; (b) mammary artery, once proclaimed to be efficacious placebo treatment should not last too long; (c) placebo for treatment of angina pectoris, might serve as an treatment should not inflict unacceptable risks, and (d) example. In the 1 950s, considerable relief of the experimental subject should be adequately informed symptoms was reported for patients with angina and informed consent given. pectoris subjected to bilateral ligation of the internal mammary artery. In the early '50s, Italian clinicians Battezzati and colleagues3 were the first to apply this Placebo-controlled randomised clinical technique, and the Reader's Digest published an on September 24, 2021 by guest. Protected trials enthusiastic report.4 A year later promising results Shapiro' defined placebo as "any therapeutic proce- were also reported by American researchers who dure (or that component of any therapeutic proce- based their enthusiasm upon data from non-con- dure) which is given deliberately to have an effect, or trolled trials.5-7 It only took two double-blind unknowingly has an effect on a patient, symptom, placebo-controlled studies, involving 35 subjects, to syndrome, or disease, but which is objectively disprove the presumed efficacy of this putative without specific activity for the condition being treatment. Fourteen subjects were placebo-treated treated". Since the earliest days of medical science, (sham-operated) and 21 underwent a ligation of the and certainly since the advent of chemotherapy, internal mammary arteries under local anaesthesia.89 placebo effects have been considered more often a The placebo procedure consisted of parasternal skin nuisance than a therapeutic tool. Both inert and incisions without ligation, while the "active" treat- active therapies were observed to produce effects ment consisted of parasternal skin incisions followed beyond their predicted physiological properties. by ligation of the internal mammary artery. It was These rather surprising observations on the clinical demonstrated that internal mammary artery ligation did not increase cardiac muscle perfusion and had no effect on the pathophysiology of coronary artery Key words disease. Although deception was obvious in these Medical ethics; placebo; research; epilepsy; Alzheimer's two placebo-controlled trials (patients were not disease; deontology. informed about the possibility of sham operation), PP De Deyn and R D'Hooge 141 J Med Ethics: first published as 10.1136/jme.22.3.140 on 1 June 1996. Downloaded from the double-blind evaluation of the procedure's effect Covi'4 found that a number of patients even refused in 35 patients prevented wide-spread introduction of to believe they had received placebo. Alternatively, this non-efficacious surgical procedure. in discussing placebo use with patients, one might Demonstration of safety and effectiveness of a consider another ethically acceptable or preferable drug is a legal requirement for marketing drugs in description of placebo: "Placebo is a medicine many countries. In the USA, evidence submitted to which does not act directly through known bodily the Food and Drug Administration (FDA) to meet mechanisms but which may work through the this requirement has to include results of "adequate mind". and well-controlled investigations" capable of distin- Although many authors'5 vividly defend con- guishing "the effect of a drug from other influences, sequentialist or utilitarian strategies and deceptive use such as spontaneous change in the course of the of placebo, others point to the important unethical disease, placebo effect, or biased observation". and non-scientific use of placebo in so-called thera- According to FDA regulation this usually implies peutic or diagnostic contexts. Goodwin et al16 standard clinical trial features such as (inactive) surveyed the knowledge ofplacebo action in 60 house control groups, randomised assignment to treatment, officers and 39 registered nurses, and their patterns of and blinded outcome assessment. Difference in placebo use. Most respondents underestimated or outcome among patients, concurrently randomised were unaware of the power ofplacebo effects, and the to therapy, should not be due to personal beliefs, following unfortunate patterns of placebo use secondary interests, and/or prejudices of patients, emerged: (a) to prove the patient "wrong" (ie, to investigators or sponsors. FDA official Leber'" expose a patient thought to be exaggerating, imagin- stresses the necessity of RCTs. Even though ing or faking symptoms); (b) in disliked or "undeserv- Leber states that "[t]here is no alternative to the ing" patients (for example, alcoholic, psychotic or randomised controlled design - no ifs, ands, or manipulative and demanding patients); (c) in situa- buts!", he certainly does not take the RCT device as tions where standard treatments either fail or make infallible. In any case, open studies, case series and patients worse, (d) as a group activity to act out staff studies relying on external or historical controls do annoyance towards "difficult" patients. copyright. not allow sound conclusions. Historical controls are Deception in the therapeutic use of placebo might highly unreliable, especially in regard to diseases be ethically acceptable in few well-defined cases, for with poorly known or highly variable natural courses example, when the physician is dealing with subjects such as insomnia, anxiety, depression and pain syn- with a history of substance abuse or when subjects dromes." '-3 As Leberl' argues, it is often impossible have to be withdrawn from certain addictive agents. to show improvement to be unrelated to pharmaco- Indeed, in these cases, giving placebo without logical effects of the administered agents without the obtaining consent of the patient, will almost invari- http://jme.bmj.com/ inclusion of a placebo arm. ably contribute to the patient's well-being, and in addition does not imply that one withholds a possibly beneficial medical treatment. Other thera- Deceptive use ofplacebo peutic usage of placebo in non-informed patients is The two major ethical problems in the use of regrettable, ethically unacceptable and illustrative of placebo involve deception when used either thera- ignorance and prejudice. It needs to be stressed, peutically or in research without adequate patient however, that malingerers or drug addicts are not on September 24, 2021 by guest. Protected information and consent, and the potential conflict relieved more efficiently by placebo. Quite on the between the Hippocratic and scientific obligations of contrary, most studies suggest that such patients are the therapist-researcher. Even though it remains less likely placebo responders. Also, complaining or inadmissible to use placebo for therapeutic or diag- "manipulative" patients are no more likely to nostic purposes, most therapists do occasionally respond to placebo than patients who are well liked apply placebo in a deceptive but rather benevolent by the hospital staff.'7 way. Use of placebo for therapeutic and diagnostic In the scientific application of placebo, however, use is inadmissible since the patient may be withheld informed consent should be mandatory, implying from an active treatment, may be denied his or her that thorough information with regard to the ratio- right to self-determination and finally, because there nale, the design, the eventual standard therapeutic is no scientific ground for the application of placebo procedures for the given disease, the randomisation in diagnosis. Deception is certainly unacceptable procedure as well as the chance of being randomised when applying placebo as an inactive control in to placebo should be communicated and discussed. clinical research. Moreover, informed consent procedures require Park and Covil4 demonstrated that deceipt may subjects to be informed about all risks and benefits not be necessary in therapeutic application. They associated with the trial and their right to withdraw showed that in neurotic patients presenting signs of at any time. Some authors disagree with this and anxiety, placebo could be used openly as a therapeu- oppose compulsory informed consent. Brewin'8 tic agent. Virtually all patients accepted the placebo argues that "too much information may be as bad as and improved over a one-week treatment. Park and too little". This author also suggests that some 142 Placebos in clinical practice and research J Med Ethics: first published as 10.1136/jme.22.3.140 on 1 June 1996. Downloaded from investigators, once in possession of written informed patients in placebo-controlled trials of their "right to consent, might become less concerned than they receive" treatment of acknowledged efficacy merely should about their ethical responsibilities. Still other in order to verify whether or not other active treat- authors perceive an incompatibility between ments exist. Withholding treatment of proven informed consent and placebo control. Levine'9 efficacy clearly violates the therapist's commitment warns us about subjects who are thoroughly informed to individual patient welfare. However, the conflict of the expected therapeutic and adverse effects of between therapist and researcher more often is emo- pharmacologically active agents in placebo-con- tionally, rather than scientifically, based. Patients trolled trials. In several clinical trials, subjects can be allocated to standard therapy control groups, unblinded the study by correctly guessing their treat- or to placebo groups when no standard therapy ment assignment. Levine'9 proposes a modification exists, without violation of Hippocratic commit- in consent procedures that would eliminate or at ment; developing new treatments of an already (par- least reduce this possibility without defeating the tially) treatable disease need not necessarily be ethical purpose of informed consent. Levine dangerous to the subject. Some therapists stress the suggests adding irrelevant side-effects in the pragmatic viewpoint of the RCT: only the applica- provided information that are of the same order of tion of an accepted treatment as control arm can importance as the actual expected side-effects. This answer the question whether the new treatment is consistent with the ethical purposes of informed improves on a standard method. The use ofplacebo- consent in that it entails disclosure of material controlled RCT designs is the only way to minimise risks. the number of ineffective drugs and therapeutic pro- Another poignant issue concerns the patient's cedures. The conviction that ignorance may cause ability to understand placebo-controlled trials. Do patient harm may not only be used as a justification the patients really know what placebo means, do for research but in addition renders research an they realise that they have a certain chance of receiv- ethical obligation. The important ethical difficulties ing placebo and why they will be on it? Stanley20 associated with the widespread application of a

reviewed the literature and concluded that irrespec- new treatment without a trial, and consequentlycopyright. tive of their condition and whether or not they were potentially without specific effects but with varying psychiatric patients, patients were fairly able to degrees of side-effects, is far greater than those understand the risks and benefits of a proposed associated with the trial itself. The optimal and treatment and the purpose of a particular treatment therefore often placebo-controlled and ethically or procedure. founded RCT meets the duties of benefiting society and increasing knowledge without jeopardising the well-being of the experimental subjects. http://jme.bmj.com/ Conflict between Hippocratic and There are many examples of marketed com- scientific obligations pounds without scientifically demonstrated efficacy The therapist-researcher is subject to two funda- and, in addition, several companies market so-called mental ethical obligations, a Hippocratic one and a alternative medicines without needing to meet scientific one. The therapist's Hippocratic obligation rigorous drug regulatory requirements for proof of obliges him/her to apply all existing knowledge for the efficacy. Recently, Pope2' correctly stated that forty best possible treatment of individual patients. On the different compounds for treatment of Alzheimer's other hand, in agreement with the researcher's scien- disease are available in Europe without any proven on September 24, 2021 by guest. Protected tific obligations it is unethical to produce unsound sci- efficacy. The Ginkgo biloba case is an example entific data. Thus, it is the duty of the researcher to where rigorous scientific standards of efficacy are not acquire new knowledge so that future patients might met.2' 25 In 1988, 52224 million prescriptions for benefit, and to communicate accurately this know- Ginkgo biloba extract involved a cost of DM 370 ledge to the scientific community in order to con- million to former West German health insurance. In tribute to the collective benefit. However, these two 1992, sales were comparable with total costs of obligations may on certain occasions be in contradic- approximately DM 368 million. The drug, licensed tion to the execution and design ofplacebo-controlled under the 1986 drug law which does not require RCTs. According to certain people, randomisation by scientific evidence from controlled trials but merely itself means that patients are no longer treated purely positive therapeutic "experiences", is promoted for for their own good but are used at best for the benefit treatment of disorders as diverse as peripheral of future sufferers from their condition, at worst arterial disease, memory impairment, vertigo and merely to satisfy scientific curiosity, and that they risk impotence. Kimbel,25 writes that "[t]he obvious dis- being treated inappropriately. It is thought that some crepancy that Ginkgo biloba extracts are among the kind of sacrifice is being demanded of them and that most prescribed medicines in France and Germany they should either be given a full explanation or else but not licensed in Anglo-American and not randomised. Scandinavian countries suggests widely differing The primary question is whether it can be standards of acceptance. Would it not be advisable ethically justified to deprive a certain percentage of to base any therapeutic conclusions on the criteria of P P De Deyn and R D'Hooge 143 J Med Ethics: first published as 10.1136/jme.22.3.140 on 1 June 1996. Downloaded from leading regulatory agencies - ie, evaluation of all possibilities for escape in case of deterioration of the published and unpublished studies and use of estab- symptomatology and/or increased risk for complica- lished standards for clinically relevant efficacy?" tions such as suicide.

SCHIZOPHRENIA Acceptability ofplacebo-controlled Several authors have questioned the ethical randomised clinical trials acceptability of short-term placebo treatment of One definitely needs to be fully informed about chronic schizophrenics.29 Even though double-blind current biomedical knowledge with regard to the placebo-controlled trials provide the most reliable disease at hand in order to be able to make ethical evaluation of antipsychotic agents, some clinicians decisions correctly. Some examples pertaining to the fear that subjects exposed to placebo or inactive new development of new agents against depression, agents might suffer lasting harm when the trial leads schizophrenia, epilepsy and dementia will illustrate to relapse. Sixty-six per cent of patients given this necessity. placebo and only eight per cent of those given fluphenzine decanoate relapsed during a trial that DEPRESSION showed no differences between groups - neither in With the advent of a variety ofpotent anti-depressive any clinical or social variable measured at the end of agents, one might be tempted to conclude that seven years follow-up, nor in the number of relapses placebo-controlled trials should be considered ethi- after the trial.29 Placebo-controlled clinical trials of cally unacceptable. However, we believe that the dif- antipsychotic agents can still be acceptable on condi- ference in improvement rates between drug-treated tion that escape treatment is provided by applying and placebo-treated non-endogenous patients is not well thought-through exit criteria within reasonable big enough to make placebo control unethical. time after relapse. Endogenous depressions have placebo response rates around 30 per cent, while neurotic or reactive EPILEPSY

depressions have rates approaching 70 per cent.26 The first guidelines of the International League copyright. The more severe depressions (Hamilton Depression Against Epilepsy on the development of anti-epileptic score above 20) have placebo response between 30 agents state that new agents cannot be introduced and 40 per cent, whereas those with less severe unless it has been proved that patients with illness (Hamilton score below 14) have placebo intractable epilepsy become seizure-free or experi- response rates greater than 50 per cent. Brown et al26 ence an appreciable (for example, 25 or 50 per cent) found no difference between baseline clinical and reduction in seizure frequency.30 The placebo-

demographic characteristics of placebo responders controlled add-on design is the appropriate clinical http://jme.bmj.com/ and non-placebo responders in a large group of trial to identify such agents. In this type of clinical patients with major depression, illustrating the trial, the patient's baseline medication is maintained limited predictability of placebo response. Quitkin but in addition, either placebo or the new investiga- et al27 analysed the heterogeneity of placebo tional drug is introduced. However, research on new responses in 144 depressive patients (DSM-III anti-epileptic drugs with similar efficacy as drugs criteria) randomly assigned to placebo medication in already marketed but with fewer or no side-effects four double-blind drug trials. Half of should also find a place. The placebo-controlled

antidepressant on September 24, 2021 by guest. Protected the patients were rated as improved for at least one add-on design is obviously not the best approach in week: 23 per cent with abrupt improvement; 27 per this latter case. Equally active anti-epileptic agents cent with gradual improvement. Gradual improve- with fewer side-effects might be missed since side- ment was observed later in the course of treatment, effects are difficult to assess in add-on designs due to more resembled drug response, and was more likely drug interactions and difficulties in analysing indi- to persist than abrupt improvement. The authors vidual drug actions. Determination of the activity of attributed gradual improvement to spontaneous a compound may also be impossible in add-on remission. designs when the agent tested influences the blood Although placebo use may not always be advisable levels of concomitant agents.3' Add-on trials may (for example, in depressed patients at significant risk overestimate the toxicity of the test compound, of suicide, with psychotic features, or with severe especially if toxicity of the new compound is similar functional impairment), Brown's 28 suggestion to and additive to that of concomitant drugs. Finally, apply placebo in an open manner in mild to placebo-controlled add-on trials are often expected moderate depression might hold some merit. He to yield an unlikely 50 per cent reduction in subject proposes to inform the patients about the fact that seizure frequency, whereas Schmidt32 found only 15 their condition tends to respond to placebo (he per cent reduction with co-administration of a defines placebo adequately), and that in an attempt marketed drug. to treat them, they will receive placebo for a period Since epileptic seizures can result in serious dis- of six weeks after which the need for other treatment comfort or lasting disability, the sometimes more will be evaluated. The authors provide, of course, appropriate placebo-controlled, single-drug designs 144 Placebos in clinical practice and research J Med Ethics: first published as 10.1136/jme.22.3.140 on 1 June 1996. Downloaded from

face adverse ethical judgment. (Of course, certain with sufficient seizure control but side-effects can be seizure types such as primary generalised tonic- enrolled in this type of trial. Although the exact per- clonic seizures are imminently more dangerous centage of seizure-free patients suffering side-effects than, for example, simple focal seizures without sec- is unknown, one could still find it unethical to run ondary generalisation; also, patients who only the risk of losing seizure control only in the hope present seizures during sleep are less likely to that an adverse effect would disappear. Therefore, develop complications). We argue that placebo inclusion criteria eliminating patients with severe application in clinical epilepsy research can be con- epilepsy and escape criteria are again mandatory. sidered ethically acceptable under certain condi- Treatment failure, the primary efficacy variable in tions and in the following trial designs: (a) these trials, is defined by specific exit criteria placebo-controlled add-on designs in intractable relating to either seizure frequency or seizure epilepsy; (b) developmental drug monotherapy severity. versus placebo in pre-surgical video-EEG or video- Placebo-controlled designs could be considered in invasive neuroelectrophysiological monitoring (in de novo patients presenting with a first-ever seizure, order to be able optimally to observe epileptic provided that development of epilepsy is not too seizures in this patient population, the baseline imminent as it might be in patients with epileptiform medication is withdrawn in the pre-surgical work- electroencephalographic elements or in those with up; following this short observation period, the predisposing structural lesions. As early as 1881, efficacy of the new developmental drug can be Gowers noted that seizures apparently beget compared against that of placebo); (c) active control seizures, an observation that remains controversial designs with attenuated form of the active control up to the present.36 According to some authors but (in this design, an active control group is used but not to others, one of the most decisive factors in the administered dosage of the classical anti- long-term seizure remission might be the number of epileptic agent is too low to be really effective); pretreatment seizures.3738 This consideration also (d) placebo-controlled design in de novo patients holds for non-de novo patients in whom additional presenting with a first-ever epileptic seizure. seizures could change the prognosis of the disease. copyright. We suggest that monotherapy designs b and c Only a trial with a placebo arm and an active anti- should always be preceded by more than one epileptic arm would be reliable in helping to resolve placebo-controlled add-on design indicating the issue of the risk of epileptogenesis after a first- probable anti-epileptic efficacy of the test com- ever epileptic seizure. pound. Monotherapy trials were proposed to overcome the deadlock of no-difference outcome,33 ALZHEIMER'S DISEASE

and have actually been used in the recent develop- With recent FDA approval oftacrine as a drug against http://jme.bmj.com/ ment of the anti-epileptic agents vigabatrin and fel- Alzheimer's disease, one might question the accept- bamate. The acceptability ofthe designs rests upon a ability of placebo-controlled clinical trials. For several variety of factors such as type, frequency and onset reasons we argue that placebo-controlled clinical time of seizures, investigational circumstances (for trials in dementia are not only acceptable but ethically example, pre-surgery investigation of patients), and scientifically preferable: (a) cholinesterase preset safeguards such as escape criteria, duration of inhibitors such as tacrine cannot nearly be called treatment, and (tentative) anti-epileptic efficacy of standard therapy as only 23 per cent of Alzheimer on September 24, 2021 by guest. Protected the test drug. Preset escape criteria are mandatory in patients benefit from tacrine treatment39; the clinical design (b) trials. Bourgeois et al 34 used as primary relevance of the therapeutic efficacy, at best only efficacy parameter the time required to reach either a symptomatic and very limited, could even be certain number of seizures or a fixed number of doubted; (b) cholinesterase inhibitors are not devoid seizure-free days, whichever first. After a predeter- of side-effects since some 30 per cent of patients mined number of seizures or, for example, one develop transaminitis, a reversible increase in liver (secondarily) generalised seizure, patients returned transaminases, and some 15 per cent develop gas- to the preregistration treatment schedule. This trointestinal complaints39 40; (c) placebo treatment protocol was considered justified in the case of felba- does not inflict unacceptable risk on the patient suf- mate because of its promising clinical profile and fering from this slowly progressing neurodegenerative strong interaction with other antiepileptics. In favour disorder; (d) investigating new anti-dementic agents of a placebo-controlled design instead of an active- in cholinesterase inhibitor-controlled designs could controlled design was the lack of sensitivity of the lead to false positive findings (type II errors) when latter.35 applying equipotency or no-difference outcome In design (c), the test compound is compared criteria; (e) cholinesterase inhibitor-controlled against attenuated active control. The active control designs would require larger patient populations in (a marketed anti-epileptic) can be administered in order to reach sound conclusions: (f) cholinesterase different doses (the lowest dose is not considered inhibitor treatment as an active control arm would efficacious and such doses should be seen as definitely unblind the study because of the high pseudoplacebos) or at one fixed low dose. Patients frequency of transaminitis. P P De Deyn and R D'Hooge 145 J Med Ethics: first published as 10.1136/jme.22.3.140 on 1 June 1996. Downloaded from

Conclusion 10 Leber P. Is there an alternative to the randomized con- trolled trial? Psychopharmacology Bulletin 1991; 27: 3-8. Since it is inadmissible to perform ill-designed 11 Lasagna L. Testimony given at Hearings before the clinical trials and to market compounds or employ Subcommittee on Antitrust and Monopoly of the treatments without specific effect (efficacy not Committee on the Judiciary, United States Senate, exceeding that of placebo) and/or with serious side- 87th Congress, 1st Session S. Res 52 on S. 1961; 1552: effects, properly controlled RCTs form the only Part 1: 282-3. scientifically valid tool. Nature of the disease 12 Leber PD. The implicit assumptions of active control process, duration of the study period, therapeutic- trials (a critical examination). Controlled Clinical Trials toxic ratio of the agent tested, availability and appro- 1983; 14: 133. priateness of alternative therapy, and many other 13 Leber PD. Hazards of inference: the active control investigation. Epilepsia 1989; 30: S1: 57-63. considerations all play a role in clinical trial design. 14 Park LC, Covi L. Nonblind placebo trial. An explo- Even though the placebo-controlled RCT remains ration of neurotic patients' responses to placebo when the gold standard in therapy development, the need its inert content is disclosed. Archives of General for and acceptability of placebo control has to be Psychiatry 1965; 12: 336-45. evaluated case by case, considering and reconciling 15 Elander G. Ethical conflicts in placebo treatment. both scientific and ethical issues. Often, placebo J7ournal ofAdvanced Nursing 199 1; 16: 947-5 1. control might even be considered an ethical obliga- 16 Goodwin JS, Goodwin JM, Vogel AV. Knowledge and tion but some provisos should be kept in mind: (a) use of placebos by house officers and nurses. Annals of no adequate therapy for the disease should exist Internal Medicine 1979; 91: 106-10. 17 Lasagna L, Masteller F, von Felsinger JM, Beecher and/or the (presumed) active therapy should have HK. A study of the placebo response. American Journal serious side-effects; (b) placebo treatment should ofMedicine 1954; 16: 770-9. not last too long; (c) placebo treatment should not 18 Brewin TB. Consent to randomised treatment. Lancet inflict unacceptable risks on the patients, and (d) the 1982; ii: 919-21. experimental subject should be adequately informed 19 Levine RJ. The apparent incompatibility between and informed consent given. informed consent and placebo-controlled clinical trials.

Clinical Pharmacology and Therapy 1987; 42: 247-9. copyright. Peter P De Deyn, MD, PhD, MMPR, is Professor of 20 Stanley B. Informed consent in treatment and research. Neurology in the Department of Neurology at the In: Weiner JB, Hess AK, eds. Handbook of Forensic Psychology. New York: John Wiley & Sons, 1987: University of Antwerp. Rudi D'Hooge, MA, MSc, 63-85. PhD, is Research Fellow in the same department. 21 Pope K. FDA chief answers critics. Academic Focus Address for correspondence: Universiteitsplein 1, 2610 1992; 28: 28a-d. Welregk-Antwerp, Belgium. 22 Schonhofer PS, Schulte-Sasse H, Manhold C, Werner B. Sind Extracte aus den Blattern des http://jme.bmj.com/ Ginkgobaumes bei peripheren Durchblutungs- und References Hirnleistungsstorungen im Alter wirksam? Beurteilung 1 Shapiro AK. Factors contributing to the placebo effect. von Tebonin und Rokan. Internal Praxis 1989; 29: Their implications for psychotherapy. American Journal 585-601. ofPsychotherapy 1961; 18: 73-88. 23 Schonhofer PS. Ginkgo biloba extracts. Lancet 1990; 1: 2 Delay J, Pichot P. Medizinische Psychologie. Stuttgart: 788. Thieme, 1973. 24 Kleijnen J, Knipschild P. Ginkgo biloba. Lancet 1992;

3 Battezzati M, Tagliaferro A, de Marchi G. La legatura 340: 1136-9. on September 24, 2021 by guest. Protected delle due arterie mammarie interne nei disturbi di vas- 25 Kimbel KH. Ginkgo biloba. Lancet 1992; 340: 1474. colarizzazione del miocardio: nota preventiva relativa 26 Brown WA. Predictors of placebo response in depres- de primi dati sperimentali e clinici. Minerva Medica sion. Psychopharmacology Bulletin 1988; 24: 14-7. 1995; 46: 1178-88. 27 Quitkin FM, Rabkin JG, Stewart JW, et al. 4 Ratcliff JD. New surgery for ailing hearts. Reader's Heterogeneity of clinical response during placebo treat- Digest 1957; 71: 70-3. rnent. American Jrournal ofPsychiatry 1991; 148: 193-6. 5 Brill IC, Rosenbaum WM, Rosenbaum EE, Flanery JR. 28 Brown WA. Placebo as a treatment for depression. Internal mammary artery ligation. Northwest Medicine Neuropsychopharmacology 1994; 10: 265-9. 1958; 57: 483-6. 29 Curson DA, Hirsch SR, Platt SD, et al. Does short term 6 Harken DE. Clinical conference: long-term manage- placebo treatment of chronic schizophrenia produce ment of patients with coronary artery disease. long term harm? British Medical J7ournal 1986; 293: Circulation 1958; 17: 945-52. 726-8. 7 Kitchell JR, Glover RP, Kyle RH. Bilateral internal 30 Meinardi H. Clinical evaluation of antiepileptic drugs: mammary artery ligation for angina pectoris. American Guidelines of the International League Against J7ournal ofCardiology 1958; 1: 46-50. Epilepsy. In: De Deyn PP, D'Hooge R, Schafer A, eds. 8 Cobb LA, Thomas GI, Dillard DH, et al. An evaluation The ethics ofanimal and human experimentation. London: of internal-mammary-artery ligation by a double-blind John Libbey, 1994: 279-85. technique. New England Journal ofMedicine 1959; 260: 31 Theodore WH, Rauberras RF, Porter RJ. Felbamate: a 1115-8. clinical trial for complex partial seizures. Epilepsia 1991; 9 Diamond EG, Kittle CF, Crockett JE. Evaluation of 32: 392-7. internal mammary artery ligation and sham procedure 32 Schmidt D. Two antiepileptic drugs for intractable in angina pectoris. Circulation 1958; 18: 712-3. epilepsy with complex partial seizures. Jrournal of 146 Placebos in clinical practice and research J Med Ethics: first published as 10.1136/jme.22.3.140 on 1 June 1996. Downloaded from

Neurology, Neurosurgery and Psychiatry 1982; 45: 37 Beghi L, Tognoni G. Prognosis of epilepsy in newly 1119-24. referred patients: a multicentre prospective study. 33 Gram L, Schmidt D. Innovative designs of controlled Epilepsia 1988; 29: 236-43. clinical trials in epilepsy. Epilepsia 1993; 34 S7: 1-6. 38 Reynolds EH. The influence of antiepileptic drugs on 34 Bourgeois B, Leppik JF, Sackellares JC. Felbamate: a the natural history of epilepsy. Epilepsy Research 1991; double-blind controlled trial in patients undergoing S3: 15-20. presurgical evaluation of partial seizures. Neurology 39 Knapp MJ, Knopman DS, Solomon PR, et al. A 30- 1993; 43: 693-6. week randomised controlled trial of high-dose tacrine 35 Pledger GW, Kramer LD. Clinical trials of investiga- in patients with Alzheimer's disease. J7ournal of the tional antiepileptic drugs: monotherapy designs. American Medical Association 1994; 271: 985-91. Epilepsia 1991; 32: 716-21. 40 Watkins PB, Zimmerman HJ, Knapp MJ, et al. 36 Elwes RDC, Johnson AL, Reynolds EM. The course of Hepatotoxic effects oftacrine administration in patients untreatable epilepsy. British Medical Journal 1988; 297: with Alzheimer's disease. Jtournal of the American 948-50. Medical Association 1994; 271: 992-8.

News and notes First International Conference on DNA Sampling

This first international conference on DNA sampling The conference is being organised by the Research will be held in Montreal, Quebec, Canada from Center in Public Law (CRDP), Faculty of Law, September 6-8 this year. Universite de Montreal, in collaboration with the Alberta, Quebec The conference will provide a forum for interdiscipli- Health Law Institute, University of copyright. nary discussion on: DNA sampling and banking; Network of Applied Genetic Medicine, Quebec Health patenting and commercialisation; legal status of human Research Fund. genetic material and information; models of consent For information contact: Ms Samaa Elibyari, Tel: and confidentiality; policy and ethical concerns; and (514) 343-2142, Fax: (514) 343-7508, e-mail: genetic epidemiology and diversity. genet(crdp.droit.umontreal.ca. http://jme.bmj.com/ on September 24, 2021 by guest. Protected