Respected investigators from Imperial College London have shaken the core of cardiology. The stakes could not be bigger. Millions of people have received stents for stable coronary artery disease (CAD) at a cost of billions of dollars.
The ORBITA[1] trialists, who presented their findings at TCT 2017, tested the last holdout for PCI in stable patients—relief of angina. Never before has PCI been subjected to the best test for any medical procedure—a sham control.[2,3] Angina is subjective, and subjective symptoms are susceptible to placebo and nocebo effects.[3,4]
It's important to set out that this discussion deals with the use of PCI in patients with stable disease—it does not address acute ischemic conditions where PCI delivers dramatic benefits.
ORBITA Background
In 2007, findings from the randomized controlled COURAGE[5] trial also shocked the cardiology community when it showed that PCI in addition to optimal medical therapy (OMT) did not reduce the rate of MI or death compared with OMT alone in patients with stable CAD. These findings have persisted through 15 years of follow-up.[6]
Via email, the primary investigator of COURAGE, Dr William Boden (Boston University, MA), highlighted that investigators found no subset of patients that did better with PCI vs OMT. Not those with multivessel disease and EF<50%, not those with LAD disease, and not those with nuclear studies showing moderate-severe ischemia.[7–9]
Interventional cardiologists hold that PCI delivered angina relief, but that case rests on shaky evidence. In the unblinded COURAGE trial, the small benefit in angina in the PCI arm waned at 3 years.[10] Some experts argue this was due to crossover PCI in the OMT group.
Boden believes the crossover issue is overstated, since a large percentage of crossover occurred in the first year of COURAGE. A substudy of COURAGE[11] looked specifically at outcomes of patients in the OMT group who crossed over to PCI and found no difference in end points when compared with those in the PCI group. The authors concluded that "because early crossover was not associated with an increase in irreversible ischemic events or impaired 12-month health status, these findings support an initial trial of OMT in stable ischemic heart disease with close follow-up of the most symptomatic patients."
ORBITA Findings
Patrice Wendling has an excellent news summary of the ORBITA trial. Compared with the sham-controlled group,
PCI did not significantly improve exercise time. The numerical incremental increase in average exercise time was 16 seconds (P=0.20).
PCI did not significantly improve measures on well-validated patient-centered angina questionnaires.
PCI did not significantly improve the Duke treadmill score or peak oxygen uptake.
PCI did significantly improve the dobutamine stress echo wall-motion index, indicating that stenting reduced ischemic burden.
Comments
First, the caveats. ORBITA enrolled patients with single-vessel disease. Its results do not apply to patients with multivessel disease or left main disease. Also, the findings do not suggest that patients with symptoms refractory to medical therapy should be denied PCI.
That said, it would be a huge mistake to downplay this landmark trial. The investigators bravely tested one of the core beliefs in cardiology—that "fixing" blockages benefits patients with stable disease. They did it with well-conceived methodology and meticulous efforts to maintain blinding. In doing so, not only have they called into question the benefit of an accepted procedure, they have further advanced our understanding of CAD.
Now let's look at some of the specific criticisms of ORBITA.
ORBITA Is Surely Clinically Relevant
Two major critiques of ORBITA are that patients were low risk, and thus the results don't generalize to most patients with stable disease. This is unfair. Here are two reasons:
First, the term "low-risk" as a qualifier is debatable and fluid. For instance, 98% of ORBITA patients had Canadian Cardiovascular Society Class 2 or 3 angina; the average duration of angina was 9 months; the mean area stenosis was 84% by quantitative coronary angiography; 69% had significant LAD disease, and the mean fractional flow reserve (FFR) was 0.69 (≤0.8 is considered significant). All the angiograms are published in an appendix to the Lancet paper. Take a look—would you consider them low-risk if they were your lesions?
Second, and most important, is that ORBITA enrolled patients who routinely get PCI in the real world. Rightly or wrongly, scant few operators leave these sorts of lesions untreated.
ORBITA Is Not Underpowered
Many have complained that 100 patients per arm is too small. But ORBITA senior investigator Darrel Francis disagrees.
He explained the reasoning. In prior studies, a single drug (ranolazine) improved exercise time by 46 seconds over placebo,[12] and balloon angioplasty without stenting improved exercise time by 96 seconds.[13]
Investigators used these findings to decide what margin of exercise improvement over sham would be significant. They believed that an increase in more than 90 seconds would be reasonable, but since drug-eluting stents are better than plain angioplasty, a 60-plus-second incremental increase in exercise time would be a more conservative threshold. If they were even more conservative and powered their trial for a 30-second incremental change, then surely no one could complain that the bar for PCI was set too high.
With that in mind, they settled on 30 seconds as the threshold. Plugging this value into a standard power calculation led to a sample size of 200 patients. Because the sham arm potentially exposed people to risk, it was not ethical to enroll more patients than necessary based on the power calculation.
Power calculations are complicated, but one useful thing to note is that testing for larger effect sizes means enrolling fewer patients. Francis explained that if you attempted to detect a 90-second difference in exercise time, you would only need 11 patients per arm. Obviously, that study would have a hard time getting published.
P values and null hypothesis testing often confuse me. What does not confuse me is what Dr Frank Harrell (Vanderbilt University, Nashville, TN), a respected statistician, wrote in his blog about the 28-second improvement in exercise time in the PCI arm of ORBITA. Harrell focused on the 95% upper confidence interval, which was 42 seconds. He argued that since most cardiologists do not consider a 42-second improvement in exercise time clinically significant, the fact that ORBITA shows (with 95% confidence) that PCI does not deliver this modest increase is clinically relevant.
Low Levels of Angina Hard to Improve Upon
Another criticism is that the patients enrolled had low levels of angina. It is true that the baseline burden of angina after the medical optimization phase may have influenced the outcome measure. In ORBITA, the baseline exercise tolerance of patients was greater than 8 minutes. This is good exercise tolerance for older patients, and it's hard to improve upon. Also, many patients' angina improved during the medical optimization part of ORBITA; good control of symptoms is hard to make better.
I would argue that the symptom relief seen in the medical optimization phase makes the case for better noninterventional management of angina.
The Sham Makes a Huge Difference
The dobutamine stress echo wall-motion scores confirm that PCI did its job in improving blood flow to a substantial portion of the heart. Yet ORBITA had no signal that subjective symptoms improved.
To me, this is easily explained by the placebo/nocebo effect.
Consider this typical scenario: A patient is told he should have cath because of angina or a positive stress test. He has come to understand from friends and the internet that if a blockage is found in his arteries, the doctor will "fix it." Then he has the cath, receives a stent, and the doctor says. "Sir, you had a blockage and I fixed it." After the PCI, the doctor often shows the patient pictures of the pre- and postangiogram. Everyone is happy. Doctors, nurses, patient, and family all feel relieved. That scenario induces a massive placebo effect.
The opposite effect—nocebo—is also relevant. If a patient and her clinicians knows she has a blockage that was not "fixed," then the likelihood of more angina, more repeat revascularizations are likely.
This nocebo scenario, argues three editorialists,[3] played a role in the positive results of another unblinded trial comparing FFR-guided PCI with OMT where the composite end point favoring PCI was driven by repeat urgent revascularization in the OMT group.[14]
In ORBITA, the blinding and sham procedure prevented the patient from having either a placebo or nocebo effect. This isolates the effect of improving blood flow on symptoms and exercise tolerance. No other study has done that.
Shared Decision-making Offers a False Choice
Another argument I've heard to discredit or downplay ORBITA's findings is the idea that if given a choice, patients may choose to have increased blood flow to the heart or prefer not to take medications, which could have side effects.
This thinking belies the evidence. Neither COURAGE nor ORBITA tested medical therapy against PCI. They tested whether or not PCI adds benefit on top of medical therapy.
In the debate surrounding ORBITA, I have not heard any mention of the downsides of PCI. Yes, PCI is low risk, but complications can occur. More important, in my mind, is that PCI trades one problem (flow-limiting stenosis) for another problem—having a metal cage exposed to blood platelets. Dual antiplatelet therapy (DAPT) also confers risk (and pill burden) going forward.
It's hardly controversial to call the ideal length of DAPT uncertain. Drug-eluting stents have reduced the dreaded complication of late-stent thrombosis, but it still occurs, as does neoatherosclerosis at the stent site.[15]
Conclusion
To discard the findings from ORBITA indicates hubris—a doctor's greatest enemy.
It's also wrong to overextend these findings to unstable patients or to focus only on overuse of PCI. Overuse is a problem, but I see two broader messages from this study.
One is to respect the placebo and nocebo effect. Our ability to help or harm people with our words and actions gets too little attention. We think our procedures help people, and they clearly do, but how often is it simply our caring and effort that delivers most of the benefit?
The other, and perhaps most lasting message from ORBITA, is that it could force a rethinking of the stenosis-centric frame for treating CAD.
For decades, cardiologists feared focal lesions because they cause ischemia. PCI relieves ischemia. So PCI is good. Then COURAGE came along. We understood it, we accepted it, but we didn't really believe it. And we didn't have to, we had an out: a stent may not reduce the rate of MI or death, but it relieved human suffering.
Now that ORBITA has cast great doubt on this last promise of PCI in stable CAD, maybe we can—finally—start to see atherosclerosis in a new way.
And if that happens, if we get out of the frame of mind that progress means building a better way to fix blockages, who knows what new treatment for heart disease could be discovered?
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Coronary Stents Humbled Yet Again in Stable CAD - Medscape - Nov 07, 2017.
Comments