Author + information
- aDivision of Cardiology, Emory University School of Medicine, Atlanta, Georgia
- bDivision of Medical Ethics, Weill Cornell Medical College, New York, New York
- ↵∗Address for correspondence:
Dr. Neal W. Dickert, Emory University School of Medicine, 1462 Clifton Road, #508, Atlanta, Georgia 30322.
Percutaneous coronary intervention (PCI) transformed modern cardiology. The evolution of our understanding of its role has been similarly remarkable. The COURAGE (Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation) trial, for example, dramatically shifted practice by establishing that the principal benefit of PCI in patients with stable coronary disease is symptom control as opposed to freedom from myocardial infarction or death (1). The recent ORBITA (Percutaneous Coronary Intervention in Stable Angina) trial further challenges the role of PCI in stable angina in suggesting that this long-observed benefit may derive from the placebo effect (2). ORBITA was the first randomized, double-blind, sham-controlled trial of PCI. It enrolled patients with angiographically obstructive single-vessel disease and Canadian Cardiovascular Society class 2 or 3 angina. After instituting aggressive medical management over 6 weeks, all patients were randomly assigned to receive PCI or sham intervention (angiogram and fractional flow reserve only). In the trial, PCI was not significantly better than the sham intervention with respect to the primary outcome of exercise tolerance or secondary measures of symptom relief.
This study has important historical precedent. Specifically, its results mirror the findings of 2 landmark sham-controlled trials of internal mammary artery (IMA) ligation for severe angina published in 1959 and 1960 (3). Although that treatment—which lacked physiological plausibility—was associated with dramatic improvement in patients’ symptoms and quality of life, it was demonstrated to be no better than blinded skin incisions without ligation. In the only other sham-controlled trial for stable angina, laser transmyocardial revascularization (TMR) was also shown to be no better than sham (4). Importantly, in all of these studies, both groups improved, which is consistent with a placebo effect. More recently, sham-controlled trials have played a pivotal role in the evolving technology of renal denervation for hypertension, and blinded trials clarified the benefit of cardiac resynchronization therapy in heart failure (5,6). Given this history, especially the IMA and TMR trials, why did it take so long to conduct a sham-controlled trial of PCI for angina? And, closely related, should ORBITA open the door for greater use of sham controls?
It is remarkable that PCI—introduced into clinical practice 30 years ago and estimated to be performed 500,000× annually at substantial cost—had never been subject to a placebo-controlled trial. After all, cardiology is committed to evidence and has pioneered large randomized trials. One reason, voiced by some commentators regarding ORBITA, is ethical concern about the risks of sham procedures, or of the acceptance of those risks by hospitals, institutional review boards, or patients (7). A second reason, in addition to financial incentives, is that cardiologists and patients have been convinced that PCI is beneficial. PCI makes physiological sense and is known to be effective in acute coronary syndromes. Moreover, belief in its efficacy is naturally reinforced by observing an open artery post-procedure and meaningful improvement in patients’ symptoms. Clinical experience, however, cannot discern whether the observed benefits associated with PCI derive from the intervention itself or from a placebo effect, and neither can randomized trials comparing PCI with medical therapy without blinding patients or clinicians to treatment assignment.
Given the potential value of sham-controlled intervention trials, it is important to address the ethical concerns specifically, which center around the risk of the sham itself (8). In ORBITA, the risks of the sham are principally those associated with the pre-randomization angiogram and fractional flow reserve assessment. The patients assigned to sham would not have otherwise had this catheterization, because their coronary anatomy had already been defined prior to entry in the study. In addition, sham control patients took 6 weeks of dual antiplatelet therapy after the sham. These risks are not minimal from a regulatory perspective. How are they justified?
The most important element of the justification is that this study was designed to answer a scientifically, clinically, and socially valuable question: does PCI improve symptoms through revascularization or by a placebo effect from psychological (and neurobiological) factors associated with receiving an invasive intervention presented as beneficial? There is, as described previously, reason to believe that a substantial placebo effect could exist, and this procedure is expensive, invasive, and carries risks. A randomized, double-blind, placebo-controlled study is the most rigorous method for answering this critically important question when the principal outcome is closely tied to symptoms or other subjective elements.
The second major justification for the risks of the sham in the ORBITA trial was its potential to benefit patients in the trial. It may seem counterintuitive to claim that the sham had a potential benefit when the obvious implication would be not to perform PCI in the future if the study showed no benefit from the procedure. After all, the conventional view of placebo interventions in clinical trials is that they are “nontherapeutic” research procedures. This view unfortunately ignores the substantial laboratory and clinical trial evidence for placebo effects, especially for relief of pain (9). Angina is known to be a placebo-responsive condition, as evidenced by the TMR and IMA trials. There was, thus, a reasonable hypothesis that a primary mechanism for the accepted benefit of the PCI procedure in routine clinical practice may be via placebo effect. We thus contend that the masked sham PCI offered trial participants a prospect of therapeutic benefit, which at least partially justifies the risks of the procedure (3). This is in no way an argument that placebo-PCI would be justified in clinical practice, only that the placebo effect is a relevant aspect of the risk-benefit assessment for patients in the study.
Two further points are essential regarding the ethical justification of sham trials. First, a sham intervention does not have to involve a prospect of benefit, although in the ORBITA trial it did. The scientific and societal benefits of sham studies alone can justify the risks of the sham, particularly if risks are low. There is no ethical requirement that all competent adult trial participants be offered a personally favorable risk-benefit ratio from participation in a study or from particular elements of the study. Nonclinical research angiograms, for example, are routinely incorporated into studies as part of follow-up assessments when they have no clinical value. A second, related point is that the risks of sham or blinding procedures will vary. In the FAME-2 (Fractional Flow Reserve–Guided PCI versus Medical Therapy in Stable Coronary Disease) trial, for example, a highly-influential trial of PCI in stable angina that was stopped early for the somewhat subjective endpoint of unstable angina (and subsequent revascularization), we have argued previously that blinding would have substantially improved the trial (10,11). In that study, patients were randomized while on the table at the time of the initial diagnostic angiogram. They could have been blinded to treatment assignment with essentially no added risks, because no additional procedures needed to be performed.
A final answer to why it took so long to finally mount a sham-controlled trial of PCI is regulatory in nature. Whereas the U.S. Food and Drug Administration (FDA) often requires placebo-controlled trials as a condition for licensing drugs to treat a specific disease, it has no jurisdiction over the use of invasive procedures unless it involves a novel medical device. This is the reason the FDA was able to insist on a sham-controlled trial of the novel renal denervation therapy for hypertension. Although this was a landmark decision on the FDA’s part, it is important to recognize that regulation of devices is, in this respect, less rigorous than it is for pharmaceuticals (12). The medical community, and payers, should insist on rigorous demonstration of benefit (using shams where appropriate) when considering expansion of use of devices and procedures.
The clinical implications of the ORBITA trial require further discussion and further study, likely with larger numbers of patients and longer-term endpoints, and potentially in patients with more severe disease. However, this trial is a wake-up call about potential placebo effects of procedural interventions and the importance of early use of sham controls, especially when procedures are invasive and involve substantial costs. These trials are not only ethically acceptable but also ethically advantageous and highly influential.
Both authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2018 American College of Cardiology Foundation
- Al-Lamee R.,
- Thompson D.,
- Dehbi H.M.,
- et al.
- Leon M.B.,
- Kornowski R.,
- Downey W.E.,
- et al.
- Kolata G.
- King S.B. 3rd.,
- Dickert N.W.,
- Miller F.G.