Only a Minority of Stable PCI Patients Meet ISCHEMIA Criteria, Study Contends

Less than one-third of patients with stable ischemic heart disease who underwent PCI in the United States between October 2017 and June 2019 met criteria for inclusion in the ISCHEMIA trial, contend the authors of a new analysis. Moreover, the proportion of patients treated who met those rigorous criteria varied widely across the more than 1,600 hospitals included in the National Cardiovascular Data Registry (NCDR) CathPCI Registry.

As reported by TCTMD, ISCHEMIA was released at the American Heart Association (AHA) meeting in 2019 after years of anticipation, and then was published in full in early 2020. Results showed that an initial invasive strategy, as compared with an initial conservative strategy, did not reduce rates of major cardiovascular events in patients with stable coronary disease, although relief of angina was greater with the invasive approach.

“I think most folks would say that ISCHEMIA is the most-impactful study since COURAGE,” senior author Jay Giri, MD (Hospital of the University of Pennsylvania, Philadelphia), told TCTMD. “But after the initial presentation, which got a lot of press at AHA 2019, we were struck by the fact that we didn't have a clear sense of whether it really applied to the kind of patients that we felt like we were seeing, as interventional cardiologists, in everyday practice.”

Cardiology care in general, and interventional cardiology in particular, have gotten increasingly more complicated, he noted: patients are older and sicker and tend to present later in their disease course.

“Bear in mind, ISCHEMIA was designed in a way to extend the findings of COURAGE and make it more broadly applicable to a wider range of stable coronary disease patients,” Giri said. “That was the whole point of doing it.”

Lead author Saurav Chatterjee, MD (North Shore-Long Island Jewish Medical Centers, Hempstead, NY), pointed out that the NCDR analysis was conducted while ISCHEMIA was still ongoing. “We thought, once the ISCHEMIA trial was done or at least was planned, that it was going to answer all of our questions about ischemic heart disease, especially the broad spectrum of patients who are being considered for an invasive strategy,” he said. The expectation was that the majority of patients who “fit the bill that ISCHEMIA represented” could be expected to have similar kinds of outcomes as the randomized trial.

In the end, he said, “I think it is quite an interesting, and I would say a novel finding, that only 32% of the patients really matched up to what was essentially a very well thought out, very well-published, very well-publicized, and very thoughtfully executed trial,” Chatterjee said.

Chatterjee and colleagues report their results online today ahead of the November 8, 2021, issue of JACC: Cardiovascular Interventions.

Fraction of Fractions

Between October 2017 and June 2019, more than 388,000 patients underwent PCI for stable ischemic heart disease in the United States, representing nearly 42% of all patients who received PCI during this window. Of this stable subset, just 32.28% of those who received PCI would have met criteria for ISCHEMIA, representing just 13.52% of patients undergoing PCI in the United States.

Having clinical or anatomical high-risk features was the chief reason for their (theoretical) exclusion, including left main disease in 35%, LV systolic dysfunction in 44%, and end-stage renal disease in 17%.

In all, 17% of stable patients (7% of all US PCIs) had either negative or low-risk findings on functional tests, while nearly 32% either had no stress test or had no ischemic burden reported.

It remains to be seen what kind of impact ISCHEMIA will have on PCI numbers in the United States going forward.

“The ISCHEMIA trial is unquestionably one of the most-important studies in interventional cardiology over the last decade,” Giri reiterated. “But while the findings are very valuable in informing our decision-making in many patients, the population enrolled does not appear representative of the broad range of patients that modern US coronary interventionalists encounter today,” he said.

Still, Giri made it clear that he believes the ISCHEMIA findings are important: he uses them for decision-making and in discussions with patients, particularly to make the point that an intervention may make them feel better, but it will not reduce their risk of future myocardial infarction or improve their survival. Moreover, it is not without risk, he noted—particularly for the growing population of patients with complex, high-risk lesions and chronic total occlusions. While it’s a landmark trial that he draws on daily, “I have to be smart enough to know it doesn't apply to everybody,” said Giri.

Chatterjee, too, cautioned that it would be especially important not to apply ISCHEMIA’s messages to other groups of patients. “There is a definite concern for undertreatment of ‘higher-risk’ ischemic patients, who could have revascularization deferred if the results of the ISCHEMIA trial were to be applied universally.”

ISCHEMIA Investigators Call for Recount

In an accompanying editorial, leaders of the ISCHEMIA and ISCHEMIA-CKD trials offer some “context” to the NCDR numbers “with the hope that this helps readers clearly see the relevance of the trial to patients under their care,” write David J. Maron, MD (Stanford University School of Medicine, CA), Sripal Bangalore, MD, MHA, and Judith S. Hochman, MD (both NYU Langone Medical Center, New York, NY).

First off, patients that Chatterjee and colleagues deemed “excluded” from ISCHEMIA on the basis of end-stage renal disease can easily be moved between columns because these patients were included in the companion ISCHEMIA-CKD trial, which ultimately delivered a similar primary endpoint as the main ISCHEMIA trial. Patients with the other major categories of “high-risk features”—left main disease and heart failure—were appropriately excluded, they note, because “there is little controversy regarding the benefit of revascularization in patients with stable ischemic heart disease with left main coronary artery disease or left ventricular ejection fraction < 35%.”

On the other end of the spectrum, they note, lower-risk patients with negative or low-risk features on functional tests are a group in whom practice standards are already clear. “Low-risk patients were excluded from ISCHEMIA because they were shown in COURAGE and BARI 2D to not benefit from revascularization and they do not meet guideline recommendations for elective PCI in the absence of symptoms,” they write. Thus including them in the current analysis “further inflates” the group of patients “not eligible” for ISCHEMIA by including a subset for whom current practice is not in question.

That leaves the patients for whom ischemic burden on functional tests was either not reported or never known. Using a conservative estimate—that 50% of patients with missing data might have been eligible for ISCHEMIA—and rolling that into a series of other assumptions for the higher- and lower-risk groups, the editorialists conclude that the ISCHEMIA results are likely applicable to anywhere from 62.1% to 68.6% of patients in this NCDR snapshot.

“At best, the trial results apply to a far higher proportion, excluding only those at high risk (18.5%) or with unacceptable symptoms despite maximal medical therapy (percentage unknown), for whom PCI is clearly indicated,” the editorialists conclude.

Consistent Message?

Chatterjee et al’s NCDR analysis is not the first to try to apply the ISCHEMIA results to a real-life cohort. Leonardo De Luca, MD, PhD (UPMC Salvator Mundi International Hospital, Rome, Italy), and colleagues, writing last year in EuroIntervention, report that nearly 85% of stable patients who received PCI in the Italian START registry would not have met criteria for ISCHEMIA and just 4% would have definitively met the bar for inclusion. In the START analysis, however, the vast majority would have been ruled out on the grounds that they did not have evidence of myocardial ischemia.

Speaking with TCTMD, De Luca noted that this discrepancy speaks in part to the much greater use of stress testing in the US versus Europe, but emphasized that the “overall message is consistent” between the START analysis and this one.

“The ISCHEMIA-like population is a minority of our PCI population worldwide,” he said. But for a substantial proportion of other patients, he noted, best practice is already established.

So does that mean that ISCHEMIA does not provide answers for the vast majority of stable patients being considered for PCI or is it that studies consistently show that many patients undergoing PCI should not be doing so?

“I think that both conclusions are true,” said De Luca. “The main message that we can draw from the ISCHEMIA trial is that we should treat our patients on an individualized approach,” he said. “We should care for our patients, especially those at low or intermediate risk with optimal medical treatment: that is a fundamental symptomatic and prognostic treatment goal for the majority of our patients. But we should offer PCI to those patients in whom angina persists, that are not a minority of these patients. In the end, the ISCHEMIA trial confirmed that contemporary optimal medical therapy is a wonderful option for many patients with stable angina [and it] confirms the importance of revascularization when [optimal medical therapy] fails to limit the symptoms of angina in our patients.”

To TCTMD, Giri pointed again to the 17% of patients who got PCI in their stable cohort (making up 7% of all PCIs), who underwent procedures despite having a negative or low-risk stress test. This group of “inappropriate” patients has long been the target of studies and guidelines—not to mention controversy.

“If you looked at that number 15 years ago, what would it have been? It would have been much, much higher than that,” said Giri. “It just shows how much our field has put an emphasis on toning down, and it's probably not right for that number to go to zero, right? Because there probably are some hypothetical circumstances, and I can think of several, where you need to get a patient on the table, you need to act. And when you check these boxes, it makes it look like that patient doesn't fit the criteria for ‘appropriate PCI.’”

There are likely still some “bad actors” out there, said Giri, but these days they are likely the vast minority.

Indeed, a separate aim of the NCDR analysis, said Chatterjee, was to look at whether patients being taken for PCI differed across hospitals in the United States, and how closely different hospitals matched up with the ISCHEMIA population. Here, investigators found “significant variation” regarding the proportion of stable patients undergoing PCI who met ISCHEMIA criteria, with less than 1% of hospitals treating at least two-thirds of their patients on the basis of criteria that matched up with ISCHEMIA.

“For something like the treatment of stable ischemic heart disease, where you are essentially adhering to a specific protocol, we really did not imagine we would see so much variability,” Chatterjee said.