–Prominent interventional cardiologist says clinical trials are slowing progress
In their eagerness to embrace a glorious future of ever more spectacular technology-based advances, interventional cardiology— a subspecialty never exactly known for caution, patience, or self reflection— is poised to reject science, evidence-based medicine, and randomized controlled trials.
“Generating and publishing evidence is a tedious job,” writes Bernhard Meier (University Hospital of Bern, Switzerland), one of the top interventional cardiologists in the world, in the European Heart Journal. “Waiting for the results of randomized trials may preclude patients from an apparently good thing while the trials are ongoing.”
The problem, it should be unnecessary to point out, is that what often seems like “an apparently good thing” turns out, after careful scrutiny with good evidence, to be not good at all, and perhaps even harmful. And, ahem, expensive. Just because you are an interventional cardiologist doesn’t mean you are infallible or can see into the future.
Medicine has always existed somewhat uncomfortably in the dual worlds of science and belief. Beliefs derive from tradition or good ideas. Interventional cardiology is a field built on good ideas— balloon angioplasty, stents, and TAVR, to name a few. But we tend to forget that the field is also littered with the corpses of good ideas that didn’t work and even caused harm— just think of laser angioplasty, atherectomy, or radiation to prevent restenosis. Further, it’s important to recognize that even the good ideas that work out aren’t panaceas— they need to be carefully studied to find out precisely which patients derive benefit, and this is not an easy question to answer by any means. The best stent in the world can only bring harm to someone who doesn’t need it.
Science is based on doubt. You shouldn’t believe anything— especially your own good idea— unless it’s been tested by a scientific experiment and, preferably, replicated.
Now I confess that I don’t fully understand Meier’s piece. The prose is, to put it kindly, difficult to decipher and the logic convoluted. I’m not even convinced that Meier is being entirely serious all the time. But there are several moments of clarity that raise real concerns and, because they likely reflect the thinking of some significant portion of interventional cardiologists and many other physicians as well— really need to be addressed.
For Meier, evidence-based medicine is a relic of the past, though he acknowledges that it is “an uncontestedly laudable product of the last century.” He continues:
“Collecting evidence in a randomized fashion, ideally with double-blind, double-dummy, and sham-controlled design, could well be overrated. Currently, it is a declared holy grail but that may take bizarre shapes.”
There is some truth here. The SYMPLICITY HTN-3 trial, for instance, which tested renal denervation took a very bizarre shape, requiring control patients to undergo a sham procedure. Until this trial came along there was near universal belief, among interventional cardiologists especially, that renal denervation would be massively effective. The results, of course, proved otherwise, sending everyone back to the drawing board. In the end renal denervation may prove to be effective, but it was first absolutely essential to show that its initial iteration was an abject failure.
Meier goes on to rail against RCTs, though he doesn’t provide any good examples:
“Even more grotesquely misleading conclusions can emanate from poorly conceived or sloppily conducted randomized trials. Current publication criteria for randomized trials do not allow one to alter the endpoint of a study after the study has commenced. Neither do they allow one to hypothesize in the abstract or the final conclusions of the paper on how the results might have been if the design or endpoints of the study had been chosen differently.”
This is preposterous, though any defender of clinical trials must acknowledge that RCTs, like anything else, can be abused and distorted. But Meier’s points here are patently untrue. As the Oxford COMPARE group has conclusively demonstrated, endpoints are altered all the time in even the most prestigious journals. The problem, however, isn’t altering outcomes. The problem is when outcomes are altered without any notification of the change and when the original endpoints aren’t reported. Does Bernard Meier really need to be reminded of the importance of prespecified endpoints and prospective designs? And Meier must be reading different journals than the rest of us, or even the journal in which his article is published, because authors constantly speculate on how their results might have been different if the trial had been designed differently.
Stents For Everyone!
One of Meier’s main points seems to be that there was far too much resistance for far too long of drug-eluting stents. Essentially he argues that all patients (though he never defines who that group might be) should receive a DES. Back when there were concerns about stent thrombosis with DESs Meier argues that “all patients should have received them.” But of course it’s impossible to know ahead of time how a controversy like that will be resolved, and in addition to the necessity of generating evidence there is the general principle of not doing any harm. And at no point in his discussion of DES, or indeed any other technology, does Meier consider the role of cost. It’s also worth point out that there are plenty of knowledgeable experts— including even a few interventional cardiologists— who believe that DESs are overused.
Meier’s enthusiasm for stents leads him to reject fractional flow reserve (FFR) to identify lesions. ”In the realm of easy to implant and utterly safe DESs, fixing the lesion is faster and about as safe as assessing it with the FFR,” he writes. Meier is not even particularly bothered if a stent is implanted in a lesion that is “not yet haemodynamically significant,” since “it most likely would evolve to that point over time.” He then goes on to take the extreme position that “PCI might be the better choice over conservative treatment, irrespective of trials stating the opposite.” This is a breathtaking rejection of evidence. For all its problems, nearly all experts agree that the COURAGE trial exerted a powerful and much-needed brake to stent-mania. The logical conclusion of Meier’s position is universal delivery of full metal jackets. This would indeed be a utopia— for interventional cardiologists.
PFO Closure For Everyone!
Meier ends with a passionate defense of PFO closure, the benefits of which have been “underestimated because of imperfect designs of the respective randomized trials.” This is because “the assessment of the results was conducted before the point in time when it can be assumed that the majority of potentially preventable events will have happened.” According to Meier, “waiting for the results of randomized trials may preclude patients from an apparently good thing while the trials are ongoing. In particular, preventive procedures such as the closure of the PFO need studies with >10 years of follow-up to be able to prove the expected benefit. Should all patients outside such a trial be deprived of a potentially life-saving simple intervention?”
This is an extremely dangerous proposition. Meier doesn’t even mention the alternative possibility that the absence of evidence could lead to full adoption for decades of a harmful device.
It is also untrue that PFO closure can not be studied in RCTs. The PFO closure studies were plagued by slow randomization. The problem was that interventional cardiologists refused to randomize their patients because they had already adopted Meier’s mindset that the procedure was beneficial. Potential trial subjects were instead treated with off-label use of devices, to the financial benefit of manufacturers and implanters alike. As an editorialist wrote when the CLOSURE 1 trial was finally published: “During the 9 years it took for the results of this trial to be reported, approximately 80,000 patients have had a patent foramen ovale closed with the use of a device at an average cost of $10,000 per procedure. Even if only half these patients were treated by this method for the purpose of preventing stroke, it would suggest that during that period of time $400 million was spent on a procedure that had no apparent benefit, to say nothing of the potential clinical risks involved.
Interventional cardiologists who argue that devices should be used without supporting RCTs because the trials are impossible to perform remind me of the defendant accused of murdering his parents who asked the judge for mercy because he was an unfortunate orphan.
Interventional cardiologists think they know what’s good for us. They need to make sure they can back up their views with the tools of science.