| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Division of Pulmonary, Allergy, and Critical Care Medicine, 2 Center for Bioethics, 3 Center for Clinical Epidemiology and Biostatistics, and 4 Leonard Davis Institute of Health Economics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; 5 Wall Center for Pulmonary Vascular Disease, Division of Pulmonary and Critical Care Medicine, Stanford University, Palo Alto, California; 6 Gilead Sciences, Inc., Foster City, California; and 7 Department of Medicine, College of Physicians and Surgeons, and 8 Department of Epidemiology, Joseph L. Mailman School of Public Health, Columbia University, New York, New York
Correspondence and requests for reprints should be addressed to Scott D. Halpern, M.D., Ph.D., M.Bioethics, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania School of Medicine, 711 Blockley Hall, 423 Guardian Drive, Philadelphia, PA 19104-6021. E-mail: scott.halpern{at}uphs.upenn.edu
ABSTRACT
Randomized clinical trials (RCTs) conducted during the past decade have shown that several therapies produce improvements in surrogate endpoints for patients with pulmonary arterial hypertension (PAH). Whether these therapies also influence clinical outcomes remains uncertain. These changes in the PAH landscape have raised several complex ethical issues regarding the conduct of RCTs in PAH, but to date, these issues have not been fully explored. In this article, we consider patients' potential motives for enrolling in PAH RCTs and identify those that are ethically acceptable. Second, we consider the efficiency of PAH RCTs—the ratio of the value of the information gained from RCTs to the risks and costs of obtaining it—and how this efficiency quotient contributes to a trial's ethics by enabling the fulfillment of patients' motives for participating. Third, we discuss the ethics of PAH RCTs using placebo control subjects versus those using active-treatment control subjects. Finally, we consider the ethical issues surrounding the roles of physician-investigators in PAH RCTs. We conclude with several recommendations, including (1) that investigators seek to ensure that patients enrolling in RCTs do so primarily to fulfill altruistic motives, (2) that research be conducted to determine the long-term risks associated with brief periods of withholding PAH therapies before further placebo-controlled trials without background therapies are conducted in PAH, and (3) that incentives for investigators to enroll more patients in PAH RCTs, such as enrollment-based authorship, be eliminated.
Key Words: ethics • pulmonary hypertension • randomized trials • placebo-controlled trials • conflict of interest
Pulmonary arterial hypertension (PAH) comprises a group of disorders characterized by progressive increases in pulmonary vascular resistance leading to right ventricular failure, exercise limitation, and death. The last decade has witnessed the transformation of PAH from a disease with one available therapy to a disease with six drugs now approved by the U.S. Food and Drug Administration on the basis of results from randomized clinical trials (RCTs). However, the conduct of RCTs in PAH poses several ethical challenges, which have only grown as our therapeutic options have expanded.
In this article, we consider patients' potential motives for enrolling in PAH RCTs and identify those that are ethically acceptable. Second, we consider the efficiency of PAH RCTs—the ratio of the value of the information gained from RCTs to the risks and costs of obtaining it (1)—and how this efficiency quotient contributes to a trial's ethics by enabling the fulfillment of patients' motives for participating. Third, we discuss the ethics of PAH RCTs using placebo control subjects versus those using active-treatment control subjects. Finally, we consider the ethical issues surrounding the roles of physician-investigators in PAH RCTs.
WHY DO PATIENTS WITH PAH ENROLL IN RCTs?
Patients' motivations to enroll in RCTs are complex. Enrollment decisions may be influenced by specific trial characteristics, how and by whom recruitment is conducted, and patients' individual predilections for research participation (1). Table 1 lists many factors that may influence enrollment decisions in a variety of clinical settings.
|
Although required to ensure valid results, these features mean that patients in RCTs must temporarily forgo opportunities for individualized clinical care. Such sacrifices may be acceptable if participants are truly indifferent between the treatments they might receive within the study or through routine clinical care, or if they are willing to make small sacrifices in individualized treatment so as to contribute to science. However, forgoing opportunities for individualized care so as to receive an investigational drug will rarely be in patients' best interests. Thus, investigators have a duty to quell, rather than promote, patients' desires to receive novel therapies as a motivation for enrollment (3).
The distinctions between research and clinical care also mean that patients should not enroll in RCTs because they cannot obtain approved therapies outside the trial. Although patients lacking health care insurance may be tempted to enroll in studies (and investigators may be tempted to recruit them) so as to receive treatment and monitoring, studying such patients may compromise participant autonomy. Ideally, all research participants would have unfettered access to available treatments outside of the trial so that they may make autonomous enrollment decisions. Unfortunately, many PAH therapies are expensive and not routinely covered by many insurance plans. As a result, many patients lack access to these therapies and may therefore enroll in studies providing such access without fully considering the risks and benefits of participating.
In addition, targeted enrollment of patients without sufficient health care coverage is inherently unjust because it results in disenfranchised patients bearing the risks and burdens of research that could ultimately benefit all patients with PAH (4). Such selective enrollment would be unethical because it subjugates the interests of certain patient subgroups (typically the most vulnerable patients) for the well-being of all future patients. Nonetheless, a clinician operating in a climate of inequitable access to care might be justified in suggesting trial enrollment for a specific patient with no other opportunities for treatment.
Patients with PAH may also be motivated to participate in trials to please their physicians, who may simultaneously serve as RCT investigators. When such physician-investigators suggest enrollment for their patients, patients' trust in their physicians' recommendations may represent a powerful stimulus to consent. Patients may even feel pressured to heed their physicians' recommendations to enroll in RCTs so as to avoid disappointing their providers and (seemingly) impeding their future care. The fact that patients with PAH have limited opportunities to "vote with their feet" and seek an alternate PAH physician nearby only heightens such pressure.
We have observed firsthand how these problematic motives drive recruitment in PAH RCTs. Empirically documenting their frequency may help define the extent of the problem. However, investigators have the ultimate responsibility for ensuring that patients do not enroll in trials due to real or perceived pressures to do so. Awareness of these unacceptable reasons for patient participation in RCTs may help investigators discourage such errant motivations, making for more readily justifiable trials.
The main ethically acceptable motivation for patients to participate in RCTs is altruism (i.e., contributing to the care of future patients) (5). The frequency with which altruism influences patients' enrollment decisions varies among diseases (6, 7) and is unknown in PAH. Research is therefore needed to determine the degree to which patients with PAH who enroll in RCTs do so out of ethically appropriate desires, or due to more problematic motives.
THE EFFICIENCY OF RCTs
To fulfill patients' altruistic motives for participating in trials without subjecting them to undue or unjust burdens, an RCT must have an acceptable ratio of research value to research risk. Several factors contribute to the numerator and denominator of this efficiency quotient.
Research Value
The primary goal of an RCT is to answer a clinically relevant scientific question (1, 4, 8). Investigators must therefore pursue knowledge that would be important to both patients and clinicians. However, this deceivingly simple requirement depends on the available knowledge base for a disease, and thus changes over time. For example, whether a new drug improves clinical outcomes in an absolute sense (i.e., compared with placebo) may be important in PAH in 2008 because few agents, of uncertain effectiveness, are presently available. By contrast, the comparative effectiveness of a new agent against a standard regimen would be much more clinically relevant if there existed multiple PAH therapies with well-established track records, as in systemic hypertension (9). This means that the value of an identical research question posed at two distinct points in time would likely be different, necessitating the continued reevaluation of the questions asked in RCTs as the field develops.
To be valuable, trial results must also be clinically applicable. The applicability of current PAH RCTs to individual patients could be greatly improved by using analyses that focus on the proportions of patients who achieve a clinically important change in an endpoint (e.g., a specific incremental improvement in six-minute-walk distance), rather than comparing summary measures (e.g., median changes in walk distance) between groups. Summary effect estimates of continuous variables may serve to broaden the market of eligible patients whom drug companies can target, but such analyses limit the ability of bedside physicians to determine whether such drugs may help individual patients in meaningful ways (10, 11). The extent to which the surrogate endpoints used in a trial correspond to clinically important outcomes also contributes to the trial's value (see Ventetuolo and colleagues in this symposium, pp. 617–622) (35).
In addition to being clinically important and applicable, the research question must be addressed in a valid fashion (8, 12). Thus, RCTs that are underpowered, biased, nonapplicable to routine care, or otherwise flawed (and hence unable to improve the care of future patients) are ethically suspect because they prevent the fulfillment of patients' altruistic motives for enrolling (13). For similar reasons, it is unethical for investigators to expose patients to the risks and burdens of research and then fail to freely disseminate the results, because withholding results will prevent any possible impact of the research on patient care (4, 14).
Research Risk
The denominator of the efficiency equation of an RCT includes the several types of risks and burdens to which research participants are exposed. The myriad procedures that study participants undergo cannot be onerous out of proportion to either the protection they provide for patients (in the case of therapeutic or diagnostic procedures) or the incremental knowledge they produce (in the case of nontherapeutic procedures) (15). For example, a baseline right-heart catheterization may be justified to ensure that patients enrolled in an RCT meet diagnostic criteria for PAH and are those patients for whom the therapy is believed to work. This requires no external justification based on the knowledge to be gained because the procedure is intended to promote the safety and well-being of the trial participants by ensuring that those who might be harmed by the therapy (e.g., patients with pulmonary venous hypertension) are appropriately excluded. Serial monitoring of liver function tests to monitor for hepatotoxicity from the investigational drug is also intended to promote participant safety. By contrast, serial blood draws to measure plasma brain natriuretic peptide and repeated right-heart catheterizations are nontherapeutic procedures. As such, their associated burdens must be justified by the importance of the knowledge to be gained through their use.
In addition to the risks and burdens of study procedures, there are risks associated with forgoing treatments outside of the study. These fall into three categories: risks of disease progression, which may be reversible or irreversible; risks of exacerbation of symptoms; and risks of hard clinical outcomes, including hospitalization, transplantation, or death, which may occur in the short term (i.e., during the RCT) or in the long term (i.e., some time after the randomized treatment phase of the RCT). These long-term effects are relevant to the ethics of the RCT if they are attributable to some feature of the trial design, and may be particularly relevant to the ethics of placebo-controlled trials (PCTs).
In summary, ethical RCTs in PAH must have acceptable research-efficiency quotients. The acceptable balance of research value to research risk may vary among evaluators. For example, what seems to be appropriate to the sponsor or institutional review board may not be acceptable to a potential research participant with PAH (and vice versa). Continued successful execution of ethically sound RCTs in PAH will therefore require consideration of the research quotient from viewpoints of the multiple parties involved.
ETHICS OF PCTs
A PCT often represents the most definitive method for evaluating a new intervention. However, when such trials are conducted without background therapies, they pose unique ethical challenges.
It is generally agreed that investigators may not assign patients to receive an intervention (e.g., placebo) known to produce inferior and consequential results compared with available alternatives (16, 17). "Consequential results" include not only permanent declines in health but also substantial worsening of symptoms, even if these are fully reversible (18). This does not necessarily prohibit the use of placebo without background therapy when known effective therapies exist as long as the temporary assignment to placebo has no consequential impact on patients in either the short or long terms (3).
Although some consider PCTs without background therapy unethical in the setting of known effective therapies because they deprive patients in the control group from receiving the best-proven diagnostic or therapeutic modalities (19, 20), recent authors support the use of placebos in certain situations even when known effective therapies exist (3, 21). Indeed, recognizing the statistical efficiency and readily interpretable results (i.e., research value) of PCTs, the Declaration of Helsinki presently allows placebo use if "the patients who receive placebo will not be subject to any additional risk of serious or irreversible harm" (22).
Is this condition met in PAH? Two recent meta-analyses of PAH PCTs (23, 24) help elucidate the short-term risks of withholding active treatment for 12 to 18 weeks. These studies suggest that short-term withholding of active therapy does not significantly increase patients' risk for short-term mortality (23, 24). It should be noted, however, that the power of these meta-analyses remained limited to identify small but potentially important mortality differences. Furthermore, these studies showed that patients assigned to placebo experienced greater deterioration in functional capacity and hemodynamics and more frequent hospitalizations and use of rescue therapies compared with those assigned to active therapies (23, 24). Importantly, however, there currently exist insufficient data to determine whether these short-term declines in functional capacity were reversible after the subsequent introduction of therapy.
These meta-analyses also fail to address the question of whether brief periods of withholding therapy result in longer term risks that might not be captured during a 3- to 4-month trial. Thus, although there is no clear evidence that brief periods of withholding therapy result in irreversible disease progression or increased mortality, investigation of the long-term outcomes of patients who received placebo in previous PCTs may be warranted before conducting future PCTs in untreated patients with PAH.
The research risks associated with PCTs in untreated patients with PAH must also be considered in light of the potential for alternative RCT designs. The primary alternatives to PCTs are active-controlled trials (ACTs), in which investigational drugs are compared with available therapies. Such designs are attractive because they ensure that all patients receive some therapy, and they ask clinically useful questions regarding how new therapies compare with existing options (25, 26).
However, there are several problems with using ACTs in PAH. First, active-controlled equivalency, noninferiority, and superiority trials all require more patients to achieve adequate statistical power than do superiority PCTs. This may not be feasible for rare diseases like PAH.
Second, to differentiate therapies that produce clinically meaningful benefits from those that do not (21), ACTs must compare an investigational agent with a gold-standard therapy. Unfortunately, no such standard presently exists in PAH.
Finally, active-controlled equivalency or noninferiority trials are highly susceptible to bias attributable to imprecise research methods. Any errors in trial design or conduct tend to bias results toward findings of similarity, making the interpretation of trial results challenging (27). Thus, although PCTs pose ethical challenges in PAH, the alternatives also have significant weaknesses, possibly limiting their use in PAH.
THE ROLES OF PHYSICIAN-INVESTIGATORS
Ideally, the roles of physician and investigator would be separated entirely in PAH RCTs. When physicians enroll their own patients into studies in which they are personally invested (academically, scientifically, or financially), conflicts arise between serving their patients' best interests and these other inevitable interests (3). For example, if an enrolled patient were experiencing severe headaches, a clinician might recommend withdrawal from the study to see if the headaches improved by eliminating the study drug. An investigator, on the other hand, might treat the headache pharmacologically in an effort to retain the patient in the trial and avoid the potential biases associated with drop-out.
Such conflicts are unavoidable for physician-investigators conducting RCTs. Unfortunately, there may not be enough PAH clinicians to enable the full separation of investigators and physicians in PAH RCTs. If research participants' well-being cannot be monitored by clinicians without connections to the trial, physician-investigators should accept that such conflicts are inevitable, strive to identify them, and work to minimize their impact.
Although most physicians recognize that overt payments for enrolling patients in RCTs represent an unwarranted conflict of interest, per-patient reimbursements for trial-related expenses also may be influential. Even more insidious are the conflicts that arise when a physician's career advancement is based on his or her enrollment success or the success of the trial. Such influences are common in PAH RCTs, in which authorship on study manuscripts, consulting arrangements, advisory board memberships, and anointment as key opinion leaders are often based on patient enrollment and trial success. No matter how well intending the physician, such incentives will inevitably shift the balance of the physician's loyalties in an ethically unfavorable direction.
Therefore, one approach would be to determine authorship and other personal benefits that investigators may reap from conducting RCTs before and independently from trial enrollment. As in other types of scientific investigation, rewards of authorship and career advancement should be based on investigators' expertise and their intellectual contributions to the research. Hopefully, investigators will still be motivated to conduct high-quality research, including the enrollment of sufficient numbers of patients overall. However, delinking personal benefits from one's own enrollment success should help minimize the conflicts investigators face when recruiting individual patients.
CONCLUSIONS
There is much work to be done to promote the future science and ethics of PAH RCTs. First, we should learn why patients with PAH choose to enroll in RCTs and determine the influence of therapeutic misconceptions and other errant motives. Second, trial analyses should produce applicable effect estimates of valid endpoints to better serve the needs of physicians and patients (10, 11). The ongoing use of nonvalidated endpoints limits trials' value and may increase risk if clinical worsening among participants is not detected promptly.
Third, the long-term consequences of short-term withholding of treatment in the context of PCTs must be examined. The short-term risks of forgoing treatments for 3 to 4 months do not appear to be sufficiently large or irreversible that they require the use of active controls in premarketing trials, but the long-term risks must be shown to be similarly innocuous before future patients are subjected to the potential risks of placebo without background therapy. If the long-term risks are shown to be acceptable, there would be no clear reason to prohibit PCTs at this time. However, the advent of PAH therapies that clearly improve health for most patients would ultimately prevent the ethical conduct of PCTs in PAH without background therapy.
Finally, immediate changes are needed to offset the conflicts that physician-investigators face when considering whether to enroll their patients. To maintain incentives such as enrollment-based authorship despite the inability to separate the roles of physicians and investigators limits the most important safeguard against harm for research participants, the benevolent and watchful eye of an independent clinician. The conduct of RCTs in PAH should strive to meet the highest scientific and ethical standards, thereby ensuring both the fulfillment of investigators' goals and the protection of research participants' rights and welfare.
FOOTNOTES
Support for this conference, including travel for each of the authors, was provided by unrestricted educational grants from Actelion Pharmaceuticals, Lung Rx, Pfizer, Gilead Sciences, and United Therapeutics.
Conflict of Interest Statement: S.D.H. has received Actelion's "Entelligence" Research Award for Young Investigators. This research grant of $75,000 would help support his research on predictions of response to endothelin receptor antagonists. R.D. is an employee of Gilead Sciences and received $8,000 from Actelion in 2007, and $3,000 from Encysive in 2007 and $6,000 in 2006, and $1,500 from Gilead in 2007. S.M.K. has received consulting fees, advisory board fees, speaking fees, unrestricted educational grants, and/or research funding from Pfizer, Actelion, Encysive, Gilead, United Therapeutics, INO Therapeutics, and Lung Rx.
(Received in original form February 19, 2008; accepted in final form April 17, 2008)
REFERENCES
Related articles in Proceedings of the American Thoracic Society:
This article has been cited by other articles:
|
|
 |
|
  M. Gomberg-Maitland Traditional and Alternative Designs for Pulmonary Arterial Hypertension Trials Proceedings of the ATS, July 15, 2008; 5(5): 610 - 616. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
