Portal de Periódicos da CAPES

Placebo Controls in Clinical Trials of New Therapies for Osteoporosis

2003; Oxford University Press; Volume: 18; Issue: 6 Linguagem: Inglês

10.1359/jbmr.2003.18.6.1154

1523-4681

Robert J. Levine,

Pain Management and Placebo Effect

THE USE OF placebos as controls in research designed to assess the efficacy of therapeutic or preventive agents is highly controversial. This paper is concerned with only one part of this debate: the ethical justification of placebo controls in the evaluation of therapies for diseases or conditions for which there exists a therapy known to be at least partially effective. One such disease is osteoporosis. After reviewing the current status of the controversies over the use of placebo controls in general, this paper turns to the specific case of clinical trials of new therapies designed to arrest the progression of osteoporosis. The review of the ethical justification of placebo controls entails a consideration of the historical anti-placebo stance of the Declaration of Helsinki as well as the closely related position held by some commentators that use of placebo as controls rather than known effective therapy is a violation of the physician's ethical duty (fiduciary responsibility) to provide only the best-known therapy for the patient. The Declaration of Helsinki has recently been revised extensively. This revision was accomplished in response to two major criticisms1: first, that the document was logically flawed as are all documents that rely on the spurious distinction between therapeutic and nontherapeutic research; second, it was alleged that its position on the ethical justification of placebos was excessively restrictive, equivocal, and susceptible to differing interpretations. Let us begin with a consideration of Helsinki's position on placebo controls as reflected in the fifth edition of the Declaration of Helsinki (1996). Then we will appraise the revisions in this position embodied in the most recent sixth edition (2000); these will be called Helsinki V and VI, respectively.2 This appraisal will yield the conclusions that the position on placebo was not ambiguous in the 1996 edition and that it changed very little, if at all, in the 2000 version. Finally, we will consider the "clarification" of the position on placebo controls issued by the World Medical Association (WMA) in 2001. In any medical study, every patient—including those of a control group, if any—should be assured of the best proven diagnostic and therapeutic method. This does not exclude the use of inert placebo in studies where no proven diagnostic or therapeutic method exists. The implications of this article extend far beyond the use of placebo controls in clinical trials. This article, if strictly applied, would rule out the development of all new therapies for conditions for which there are already existing "proven" therapies. One cannot evaluate a new therapy unless you withhold those that have already been demonstrated safe and effective for the same indication. Strict application of this standard would have prevented the evaluation of the effectiveness of cimetidine and other H2 receptor antagonists for the treatment of peptic ulcer because the withholding of belladonna and its derivatives would have been considered an unethical withholding of the "best proven therapeutic method." Similarly, the development of new and improved antihypertensive drugs would have ceased with the establishment of the ganglionic blockers. Article II.3 also forbids placebo controls in clinical trials in which there is virtually no risk from withholding proven therapy. Consider research in the field of analgesics and antihistamines. No experienced person would ever employ an active control in the evaluation of a new analgesic. Article II.3 also rules out the use of placebo controls in clinical trials in which there is a very remote possibility of a serious adverse consequence of withholding the active drug, such as trials of new antihypertensives and of new oral hypoglycemic agents. Insisting on active controls in these areas would introduce major inefficiencies with virtually no compensating benefit; the amount of injury to research subjects that would be prevented by requiring active controls is so small that it can be and generally is considered negligible. Placebo-controlled trials of analgesics, antihypertensives, and oral hypoglycemics are conducted commonly, and the results are published in reputable, peer-reviewed medical journals. Parenthetically, it is worth noticing that such publication is a violation of Helsinki; Article I.8 (Helsinki V) and Article 27 (Helsinki VI) hold that "Reports of experimentation not in accordance with the principles laid down in this Declaration should not be accepted for publication." The most controversial interpretation of Article II.3 is that it requires the provision of the best proven therapeutic method that is available in the industrialized countries even when conducting research in countries in which such therapy is not available. This interpretation provoked the most acrimonious debate in the field of research ethics since the 1970s. The debate was begun with the publication in The New England Journal of Medicine of an article that denounced as unethical the clinical trials that were being carried out in several countries to evaluate the effectiveness of the short-duration regimen of AZT in preventing perinatal transmission of HIV infection.3 This debate, which is beyond the scope of this discussion, is not entirely resolved.1, 4-6 The benefits, risks, burdens, and effectiveness of a new method should be tested against those of the best current prophylactic, diagnostic, and therapeutic methods. This does not exclude the use of placebo, or no treatment, in studies where no proven prophylactic, diagnostic, or therapeutic method exists. The only improvement over Article II.3 is the removal of the proscription of the development of all new therapies for conditions for which there are already existing "proven" therapies (supra). And even this salutary effect is not entirely clear; it depends completely on the interpretation of the new Article 28 (infra). Helsinki's absolute proscription remains intact for placebo controls in clinical trials designed to evaluate therapies for diseases or conditions for which there already exists a therapy known to be at least partially effective. "The Declaration of Helsinki is one of the WMA's most well known position statements and acknowledged as the cornerstone of research ethics. Its current guideline on the ethical use of placebo-controlled trials has caused some confusion in the research world and for this reason the WMA Council decided to publish a note of clarification on the interpretation of this guideline. This extraordinary step was necessitated by the cancellation of the 53rd WMA General Assembly, which was due to take place during October 2001. The Assembly is the only body with the authority to adopt formal changes to the Declaration of Helsinki. The full text of the note of clarification reads: The WMA is concerned that paragraph 29 of the revised Declaration of Helsinki (October 2000) has led to diverse interpretations and possible confusion. It hereby affirms its position that extreme care must be taken in making use of a placebo-controlled trial and that in general this methodology should only be used in the absence of existing proven therapy. However, a placebo-controlled trial may be ethically acceptable, even if proven therapy is available, under the following circumstances: where for compelling and scientifically sound methodological reasons its use is necessary to determine the efficacy or safety of a prophylactic, diagnostic, or therapeutic method, or where a prophylactic, diagnostic, or therapeutic method is being investigated for a minor condition and the patients who receive placebo will not be subject to any additional risk of serious or irreversible harm. All other provisions of the Declaration of Helsinki must be adhered to, especially the need for appropriate ethical and scientific review. At the same meeting, the WMA decided to appoint a panel of advisers, representative of all stakeholders in research, to assist the WMA in its continuing review of the Declaration of Helsinki." The second of the two italic sentences in the statement quoted above seems eminently reasonable; it removes the unreasonable ban on placebo controls that had been imposed by the "unclarified" Declaration when deprivation of the best proven therapeutic method will not expose subjects to any additional risk of serious or irreversible harm. The first of the two underlined sentences is certainly an improvement over the "unclarified" Article 29. It affords an opportunity to evaluate the ethic justification of placebo controls on a case by case basis in that most difficult category of clinical trials in which deprivation of the best proven therapeutic method could result in serious adverse consequences7; however, many commentators, myself included, would have preferred some guidance as to what might be considered a "compelling" reason. The version issued in 2002 of the CIOMS International Ethical Guidelines for Biomedical Research Involving Human Subjects seems much more promising in this regard.8 Several commentators on the ethics of clinical trials have argued that the choice of a control group should be dictated by the physician's ethical obligation to provide for each patient only the best-known therapeutic method. This obligation is variously known as the "duty to care,"9, 10 the duty to provide 'the good of personal care,'11, 12 and the fiduciary obligation of undivided loyalty to the interests of the patient.10, 11 If the clinician knows, or has good reason to believe, that a new therapy (A) is better than another therapy (B), he cannot participate in a comparative trial of therapy A versus therapy B. Ethically, the clinician is obligated to give therapy A to each new patient with a need for one of these therapies. If the physician (or his peers) has genuine doubt as to which therapy is better, he should give each patient an equal chance to receive one or the other therapy. The physician must fully recognize that the new therapy might be worse than the old. Each new patient must have a fair chance of receiving either the new, and hopefully better, therapy or the limited benefits of the old therapy. Implicit in this position is a rejection of the possibility that one could justify a placebo control in clinical trials designed to evaluate therapies for diseases or conditions for which there already exists a therapy known to be at least partially effective. Perhaps traditional physician ethics, with its highly individualistic commitment to patient welfare, needs to be modified. A more socially oriented ethic might permit RCTs to proceed with a statistically adequate sample of patient-subjects…. If the traditional ethical rules governing physician-patient interaction are to be changed, then all parties should be made aware of this fact. The physician-ethicist, Weijer, has written an excellent and comprehensive review of the major ethical considerations in the justification of placebo controls in clinical trials.10 For Weijer, the central consideration in the ethical justification of clinical trials is the concept introduced in 1987 by Benjamin Freedman, "clinical equipoise"9; "clinical equipoise" is a term used to describe a state of knowledge in the expert clinical community with regard to the relative merits of two (or more) therapies for a given condition. If the expert clinical community is genuinely uncertain as to whether therapy A is superior or inferior to therapy B for the treatment of a given condition, considering both risks and benefits, then a state of clinical equipoise exists. Clinical equipoise exists, although some members of the expert clinical community earnestly believe that one of the therapies is superior to the other. Justification of a particular clinical trial necessarily requires that there can be no third therapy C that is known to be superior to A and B that is being withheld from trial subjects. The underlying grounding for the concept of clinical equipoise is the "duty to care" or the fiduciary responsibility of the physician to the patient. As Weijer and Freedman each envision "clinical equipoise," it serves to enforce adherence to the duty to care in the design and conduct of clinical trials. Using this analytic tool, Weijer reaches conclusions about the conditions in which placebo controls can be justified that are nearly identical with mine (infra). In particular, he agrees that placebo controls are justified in certain circumstances in which "effective treatment exists but is not available because of cost or short supply." It is not customary to insist on a state of clinical equipoise in placebo-controlled trials unless the purpose of the active agent being evaluated is to mitigate that component of a disease process that leads to disabling or lethal complications. Thus, it is not uncommon to see placebo-controlled trials of new analgesics, anxiolytics, and antihypertensives that are not justified by a state of clinical equipoise. To avoid Weijer's challenge, I should have added two points: First, I was referring to studies that had been reviewed and approved by Institutional Review Boards (IRBs) and that this meant that a large number of committees that are charged with the responsibility of evaluating the ethical propriety of research had found these studies ethically justified. Second, I think there should be a threshold standard for invoking the requirement for the clinical equipoise justification. Below a certain level of risk, the probability of doing any lasting damage to the patient-subject is so small that special justifications such as "clinical equipoise" are unnecessary. This proposition is closely related to the law's de minimis doctrine; de minimis non curat lex or, the law does not concern itself about trifles or insignificant matters. Weijer and some others ask: can it be said that the fiduciary duty to the well being of the patient exists only when the physician is doing things that increase the patient's likelihood of sustaining a nontrivial injury? And I reply, of course not. However, we already recognize the authority of the physician to conduct research involving patients when there is no possibility of benefit to the individual patient. This recognition is explicit in U.S. federal regulations for the protection of human research subjects even when the subjects are incapable of informed consent if the permission of a responsible relative or the legal guardian is granted. The conduct of such research is clearly not justified by the duty to care. We find it ethically acceptable to allow the physician to perform nontherapeutic procedures or interventions to serve the interests of research when the goals are of sufficient importance and the risks are reasonable in relation to the expected benefits. To be consistent, we must equally find it ethically acceptable to allow the use of placebo controls even when there is a therapy other than the one being evaluated that is known to be effective when the goals are of sufficient importance and the risks are reasonable in relation to the expected benefits. The concept of fiduciary requires undivided loyalty to the health interests of the patient. If the physician has or even appears to have any conflicting interests, these must be disclosed. The patient, then, is enabled to make a choice of whether or not to become a subject with full awareness of the potential for divided loyalty. I do not mean to claim that informed consent is the answer to all such problems. It has long been known that many patients tend to think that anything proposed by a physician either is or could be intended by the physician to benefit the patient.15 This is the phenomenon to which Appelbaum et al. gave the name "therapeutic misconception."16, 17 This does not mean that informed consent is not possible; rather, it means that one should be especially careful when negotiating informed consent to complex activities such as controlled clinical trials in which there are both therapeutic and nontherapeutic components. Many commentators have recommended that the inherent conflict between the aspirations of the medical practitioner and those of the medical researcher might best be managed by separating these two roles. One individual could serve in the role of treating physician while another could be the researcher. Most of these recommendations have centered on the problem of informed consent. I have generally resisted such proposals for reasons elaborated elsewhere.18 However, now it seems to me that the potential for confusion might, in many cases, be sufficient to make such separation worthwhile. In most cases, formal clinical trials are not conducted by the patients' primary care physicians. Rather, patients are referred to specialists who are conducting clinical trials. This is almost invariably the case in developing countries where the research setting is obviously very different from the typical health care setting, particularly when the clinical trial is being conducted with sponsorship from a developed country. In technologically developed countries, there is a common scenario, which is particularly problematic. Patients with certain chronic diseases (e.g., cancer, depression) are referred to a medical center with an expectation that there they will receive expert medical advice, and perhaps, treatment. Once there, they are invited by the specialists to become subjects in controlled clinical trials. This unexpected encounter may easily lead either to the therapeutic misconception or to a feeling of intimidation; either of these can tend to invalidate the process of informed consent. I have recently proposed that certain clinical trials should be conducted in settings that are physically removed from the patient care setting by investigators who have not previously had a therapeutic relationship with any of the patient-subjects.19 The investigators should make it very clear to the subjects that their principal occupation is to conduct clinical research. Such arrangements should be considered for all placebo-controlled trials designed to evaluate new therapies for diseases or conditions for which there are other therapies known to be at least partially effective. When withholding of the known effective therapy could result in a nontrivial adverse consequence, there should be a prima facie obligation to establish such a distinct clinical research setting. A prima facie obligation means that persons must act accordingly unless there are important ethical reasons to do otherwise. Some participants in the contemporary debate about the ethics of placebo controls have voiced several variations of the idea of role separation. Among them are some who have opposed the physician's participation in placebo-controlled trials on the grounds that this would be a violation of the fiduciary duty to care. Some of them have stated that a physician in the distinct role of clinical trialist would be sufficiently free of the customary duties of the treating physician that it would not be unethical for him or her to conduct placebo-controlled trials. I find no need to enter the argument as to whether such conditions release the physicians from their fiduciary duties because I do not think that the fiduciary duties preclude participation as a researcher in placebo-controlled trials. The WMA's decision to classify medical research as "experiments in new diagnostic and therapeutic methods" and "experiments undertaken to serve other purposes than simply to cure an individual" resulted in it's adoption of the illogical distinction between therapeutic and nontherapeutic research as its basic organizing principle. This faulty line of thinking also gave rise to its adoption of the "best therapeutic method" standard as articulated in Article II.3, the Article that proscribed placebo controls except where there is no known effective therapy.19 In Helsinki VI, the language of therapeutic and non-therapeutic research has been removed; unfortunately, the conceptual construct has not been completely eliminated. This is discussed elsewhere.19 When the new therapy is being evaluated for the treatment of a disease or condition for which there is no existing therapy known to be at least partially effective. This should be understood as including clinical trials having as an inclusion criterion patients who have tried known existing therapies without success. It should further be understood to include patients who are aware of existing therapies and have rejected them for reasons other than a wish to enroll in a clinical trial. (e.g., Jehovah's Witnesses have been enrolled in clinical trials of artificial blood substitutes after having rejected transfusions on religious grounds.) When the new therapy is being evaluated for relief of symptoms and there are provisions in the protocol for allowing patient-subjects to withdraw from the study at any time. In such studies, the prospective subjects should be informed that if their reason for withdrawal is a desire to receive known effective symptomatic relief, this will be provided promptly. When the new therapy is designed to treat a manifestation of disease that, if untreated, could eventually lead to death or nontrivial disability, and there are existing therapies that are at least partially effective in arresting or delaying the progression to death or disability, the use of placebo controls should be limited. There must be a demonstration that under the conditions of the trial withholding of the known effective therapy would be very unlikely to result in a serious adverse consequence. For example, studies of new antihypertensive agents employ reliable surrogate end-points, recruit subjects with "mild" hypertension who are very unlikely to have any serious adverse consequences even if untreated and unsupervised, are closely monitored, and of relatively short duration. Under such conditions the probability of a serious adverse consequence is extremely small. When the new therapy is intended to be an inexpensive alternative to expensive therapies that are considered "the best proven therapeutic method" in technologically developed countries and the research is to be carried out in a developing (resource poor) country with the assistance of sponsors or investigators from one or more of the wealthy countries, the clinical trial should be responsive to the health needs and priorities of the host country and the product being evaluated should meet the "reasonable availability" and "highest attainable and sustainable" standards. (In multinational research there are other standards that must be met by the investigators and sponsors; these are beyond the scope of this discussion.) In categories 3 and 4, there should be good reasons to believe that the new therapy to be evaluated could be superior to existing and available therapies for at least some members of the patient population from which the subjects are to be recruited. When withholding of the known effective therapies could result in a nontrivial adverse consequence, there should be a prima facie obligation to establish the clinical trial in a distinct clinical research setting. Such a setting should be physically removed from the patient care setting and the investigators should include no health care professionals who have previously had a therapeutic relationship with any of the patient-subjects. The investigators should make it very clear to the subjects that their principal occupation is to conduct clinical research. In considering the ethical justification of placebo controls in trials of new drugs for the treatment of osteoporosis, it is essential to have an accurate and thorough account of the relevant facts about this disease and its treatment; for this purpose I rely on the account of these facts set forth in the recent report on this topic of the working group sponsored by the American Society for Bone and Mineral Research (ASBMR).20 The clinical trials we are considering now are of therapies designed to reduce the incidence of fractures in people with osteoporosis. There are several types of therapies that are known to be partially effective in preventing such fractures. A crucial question, then, is whether withholding these known partially effective therapies for the purpose of conducting a placebo-controlled trial is likely to expose research subjects to a greater incidence of disability or death. Unfortunately, the answer seems to be yes. It has been estimated that the withholding of existing therapies for osteoporosis carries a risk of 6–80 preventable vertebral and nonvertebral fractures per 1000 patient-years, depending on bone mineral density (BMD), age, gender, and the existence of one or more prevalent fractures.20 These fractures, even those that are not diagnosed clinically, are associated with significant morbidity and a higher risk of death.20 Based on the foregoing considerations, it seems appropriate to classify placebo-controlled clinical trials in this field as being in Category 3. Justification of such trials requires a demonstration that under the conditions of the trial withholding of the known effective therapy would be very unlikely to result in a serious adverse consequence. One method commonly used to accomplish the goal of reducing the likelihood of serious adverse consequence is to employ a surrogate end-point(s). Some clinical trialists have proposed that BMD could serve as a surrogate end-point in the field of osteoporosis research; unfortunately, there is no consensus that BMD is a valid surrogate end-point.20 Thus, for the time being it will be necessary to do clinical trials in which the primary outcome measure is fracture. There are those who argue that one could reduce the exposure to disabling injury by limiting the subject population to patients at low risk for fractures (e.g., patients without prevalent fractures) or by limiting the duration of the clinical trial. These are not satisfactory solutions for this problem. For a placebo-controlled clinical trial to yield a statistically valid result, there must be an excess of at least 24–30 preventable fractures in the placebo arm.20 Moreover, limitation of the trial's duration would limit the generalizeability of the results. One would not know at the end of the trial whether the demonstrated effectiveness was only a transient effect. The most satisfactory resolution of the issues is afforded by the use of the "add-on" placebo-controlled design. In an add-on design, all subjects receive the standard therapy, which ordinarily includes the "best proven therapeutic method." In addition, the patients are assigned by randomization to receive either the new therapy or a placebo. In the case of osteoporosis for which there are several known partially effective therapies, the one that is most commonly used as the "standard therapy" in clinical trials of new therapies is the combination of calcium and vitamin D.20 This is a combination that has been demonstrated effective in placebo-controlled clinical trials. There is no other therapy that is known to be superior to "optimal" doses of calcium and vitamin D. In some clinical trials, the combination of an investigational drug with vitamin D and calcium has been demonstrated more effective than a combination of placebo, vitamin D, and calcium. In such trials, however, the dose of calcium and vitamin D has been substantially less than the "optimal" level.20 Vitamin D and calcium offer another advantage in the design of add-on trials in that they operate by a different mechanism of action than any of the other therapies that are now being developed; one cannot conduct a satisfactory add-on trial if the standard therapy has the same mechanism of action as the investigational therapy that is to be evaluated. An add-on design in which patients in each of the two treatment arms would receive an optimal dose of calcium and vitamin D would be relatively easy to justify ethically. Most importantly, no patient will be deprived of any therapy know to be superior to what all subjects in the clinical trial will receive. I must concede that this may be only a transient resolution of the ethical problems in the design of clinical trials for evaluation of new treatments for osteoporosis. As also noted in the report of the ASBMR Working Group,20 as soon as one add-on trial establishes the therapeutic superiority of an investigational therapy plus optimal doses of calcium and vitamin D over placebo plus optimal doses of calcium and vitamin D, the new best proven therapeutic method will be the combination that was demonstrated to be superior. Therefore, it would then be very difficult to justify the conduct of another add-on trial of another investigational drug in which all patients will receive the optimal dose of calcium and vitamin D. Clinicians who are highly experienced in the treatment of patients state that there are many patients with osteoporosis who refuse to take medications designed to prevent fractures because they find the side effects unacceptable; the limited benefits of these medications are not considered by these patients sufficient incentive to experience these side effects. As noted earlier, the involvement of such patients in placebo-controlled trials is ethically unproblematic; it is as if there were no known effective therapy for their condition. It is essential, however, to ensure that these patients have truly rejected the known therapy for reasons that are entirely unrelated to the conduct of the clinical trial. It is necessary to scrupulously avoid exploitation of the "therapeutic misconception" as a motivation to enroll in clinical trials under such circumstances. Careful consideration should be given to carrying out such clinical trials in dedicated research units devoted exclusively to the conduct of clinical research (supra). A final category to be considered is clinical trials in which all subjects are patients for whom all conventional therapies have failed to produce a satisfactory effect, so-called "treatment refractory" patients. The conduct of placebo-controlled clinical trials involving such patients may be justified ethically as if there were no known effective therapy for their condition. However, there are at least two serious limitations to this proposition. First, it is possible that treatment refractory patients are biologically different from other patients having the same disease; the results obtained in studies carried out in this group may, therefore, not be generalizable to the entire population of patients with osteoporosis. Second, in osteoporosis, it may be very difficult or impossible to identify which patients are refractory to therapy. The occurrence of fractures while on therapy is not necessarily good evidence of treatment failure; perhaps the patient would have had even more fractures in the absence of therapy. This work was supported in part by Grant PO1 MH/DA 56 826–01A1 from the National Institute of Mental Health and the National Institute on Drug Abuse and by Grant 1 P30 MH 62294–01 A1 from the National Institute of Mental Health.