Sham Procedures in CV Research: What’s Sacrificed for Science?

Sham controls can be integral to CV device trials, providing a unique way to overcome the placebo effect, but they also present an ethical minefield of concerns that merit close consideration, a new editorial asserts.

“If performing a sham procedure is the right thing to do for science, why does it feel so wrong?” asks Alexander Postalian, MD (The Texas Heart Institute, Houston), in the paper, which was published online Monday in JACC: Cardiovascular Interventions.

Years ago, as a cardiology fellow, he first came face-to-face with the realities of what this facet of clinical trial design means in actual practice. That day in the cath lab, he encountered an elderly woman set to undergo the sham procedure, who had severe coronary artery atherosclerosis and ischemic cardiomyopathy that were deemed unsuitable for revascularization. Ultimately, she fared well after the pretend procedure—but, as Postalian posits in his piece, what if she hadn’t?

At the time, “it sort of hit me that, well, this feels weird,” Postalian told TCTMD, referring to the ethical questions involved. That feeling stuck with him, and was the motivation behind his new editorial.

Although his perspective has evolved over the years, with sham procedures carrying less of an emotional punch, “I still think it's important to not forget the humanity of the patients. . . . I mean, we do procedures all the time and we stop feeling the small emotions,” he said, noting how easily the people who are “patients” can slip into being viewed as “subjects.”

That’s not to say there’s no role for sham-controlled trials, Postalian stressed. “I think all interventionists agree that sham procedures are extremely important. They are essential to advance science and to prevent useless or even harmful procedures from becoming widespread. Otherwise I wouldn't be participating in these studies [as a researcher],” he commented. “But I think we should recognize the trust that they place in us as individuals that have more information regarding the nuance of their condition and the procedure they're about to undergo.”

We should recognize the trust that they place in us as individuals that have more information regarding the nuance of their condition and the procedure they're about to undergo. Alexander Postalian

Importantly, Postalian writes in his piece, use of a sham “allows us to challenge the dogma of established paradigms and potentially avoid a great deal of patient morbidity, and even mortality, by discontinuing nonbeneficial procedures that carry risk.” Prime examples in cardiovascular medicine include SYMPLYCITY HTN-3 and ORBITA, whose use of sham controls were not without controversy.

Neal W. Dickert, MD, PhD, who along with Chandan M. Devireddy, MD (both from Emory University, Atlanta, GA), co-authored an editorial that offers a counterargument to Postalian’s paper, noted that the suitability of sham-controlled trials is an issue that has “percolated” in cardiology over several decades.

ORBITA and SYMPLICITY HTN-3 then “raised this issue to prominence, and I think people have to think more about the need for sham arms now than they might have if you go back 10 years ago,” Dickert said in an interview with TCTMD. He cited the FAME 2 trial, comparing optimal medical therapy with or without FFR-guided PCI in patients known to have at least one flow-limiting lesion, as an example of when a sham control was not used, but perhaps should have been.

Dickert and Devireddy, in their viewpoint, elaborate on their stance that sham controls are a necessary element of cardiovascular research. But to do such research right requires thought, they stress: a clear understanding of any sham intervention’s possible consequences, ethical standards that recognize what level of scientific benefit and social value can justify that particular risk exposure, and acknowledgement of the harms that can occur when unproven interventions see clinical use.

Perhaps where the two editorials most disagree is whether shams constitute deception.

“In its simplest deconstructed interpretation,” Postalian states, “a sham procedure involves intentionally harming an individual under the false pretense of possibly improving their health, all for the potential benefit of the general population.”

Dickert and Devireddy, though, say that this characterization is “incorrect and unhelpful.”

“There is also no ‘false pretense’ in ethically conducted sham research that is supported by a robust informed consent process,” they write. “It is certainly the case that informed consent plays a critically important role, and there is a need to study how best to conduct consent processes in sham studies in order to promote understanding and avoid therapeutic misconceptions or misestimations. However, there is no ‘false pretense’ in a well-designed and well-conducted sham-controlled study.”

Postalian, to TCTMD, allowed that his wording might be dramatic, but reiterated his point. “Intentionally harming does not mean malice or vicious intent, and false doesn't mean lying. But [with] the theatrics involved in sham procedures, you're trying to make the patient believe you're doing something [and] most certainly attempting to deceive the patient,” he said.

Ethical Elements

Where the two papers do overlap is in the shared message that, if a sham-based trial design is to be pursued, it’s necessary for trialists to proceed cautiously.

Informed consent is crucial in situations like these where study participants are agreeing to go through some degree of physical intervention beyond merely taking a medication. Every single person goes through that consent process during trials, though that’s no guarantee the message gets through, Postalian cautioned.

“The problem is, what are their internal perception and their internal hopes? Because we want to believe that we achieved perfect informed consent: that the patient understands the minute details of both their condition and the planned procedure, and they are participating in the sham trial from a position of selflessness to the betterment of others,” he said. “But we should probably assume the much more likely scenario, which is that they have limited information about the nuances of their condition and plan procedure. And the only reason they are agreeing to participate is with the hopes of feeling better or living longer.”

Postalian added: “Once you're directly involved in talking to these patients, you realize that most of the time it's more likely the latter scenario.”

[The] consent process to a sham control trial does not involve deception. It doesn't involve hiding from people that you're going to be randomized into one of two arms. Neal W. Dickert

Dickert noted that while the sham procedure itself does mislead people about what they’re actually receiving, the “consent process to a sham control trial does not involve deception. It doesn't involve hiding from people that you're going to be randomized into one of two arms, one of which is going to be a sort of procedural placebo.”

How patients perceive sham-controlled trials is an area ripe for research, he added. Still, “it's not wrong for a patient to sign up for a sham-controlled trial understanding they may have a sham and hoping for benefit. Those two things, they don’t cancel each other out. . . . I don't think that the hope for benefit and the understanding of the study necessarily are at cross purposes.”

Beyond informed consent, Dickert called out two main elements that should underpin sham-controlled trials. “One really important issue from a design perspective is achieving the sham effect with as little risk as possible,” he said. “The second piece is I think a closely connected concept: assuring that any elements of studies that are done primarily for research purposes that impose a level of risk, we need to be sure that the scientific value of the information produced by those components is robust.”

For operators doing the procedures, there also may be relatively simple ways to reduce the risk of harm to patients, Postalian’s paper suggests. It could be possible to lessen the invasiveness of the intervention while preserving the placebo effect, such as by making a venous puncture and using the smallest-allowable sheath size, or maybe even just injecting a local anesthetic. Lastly, he says it might be possible to come up with statistical methods that could predict the placebo effect, as a means to improve the external validity of study results and in some cases eliminate the need for a sham-control approach.

Dickert and Devireddy take issue with that final suggestion, calling it “profoundly unscientific,” adding that it grossly overestimates the knowledge we have of likely placebo (or nocebo) effects in specific contexts. It is also unrealistically optimistic about the power of statistical adjustment,” they comment.

Postalian, while agreeing that this would be challenging to pull off, said that it’s not impossible: “We should not become shy of exploring advancing the field of statistics and improving our understanding of the placebo effect and helping quantify it and predict it.”

Despite the many points and counterpoints, Postalian believes the two papers shouldn’t be portrayed as “dueling views,” as they overlap more than they diverge. “I think we're all on the same page, to be honest,” he said.

The key takeaway is that sham-based controls must keep the patient in mind, concluded Postalian. “We know that the patient having the sham procedure stands to gain no benefit—only risk. So it's our duty to not only make the risks ‘low,’ but we have to decrease it to the absolute minimum possible.”