No Survival Benefit With Revascularization in Stable CAD, Meta-analysis Confirms

There is no survival advantage with invasive coronary revascularization over medical therapy in patients with stable ischemic heart disease, but revascularization does reduce the incidence of unstable angina and provides greater symptom relief, according to the results of a new study-level meta-analysis.

Although there was no significant reduction in the risk of MI overall, coronary revascularization did appear to lower the risk of spontaneous MI, albeit at the higher cost of more procedural infarctions.

“The results are very similar with ISCHEMIA,” lead investigator Sripal Bangalore, MD (New York University School of Medicine, New York, NY), told TCTMD. “I would say ISCHEMIA included a higher-risk group of patients with moderate-to-severe ischemia. Even with that, the consistent message is that there is no survival benefit [with revascularization]. If the patient is symptomatic, there will likely be better and longer lasting angina relief, and potentially you can reduce the risk of spontaneous MIs but there is an up-front risk of procedural events.”

In the National Heart, Lung, and Blood Institute (NHLBI)-sponsored ISCHEMIA trial, which was presented in 2019 and published in April, there was no significant difference in the risk of MI during the 4 years of follow-up. There was a trend toward more MIs, predominantly procedural, in the invasive-strategy arm during the first 6 months of the study, but as the trial progressed the event curves crossed such that MIs, in this case largely spontaneous infarctions, were more common in the conservative treatment arm.

The new meta-analysis, presented during a late-breaking clinical trials session at the PCR e-Course 2020 and published simultaneously in Circulation, included 14 randomized, controlled trials, among them ISCHEMIA and ISCHEMIA-CKD. Other trials included COURAGE, FAME 2, and BARI 2D, as well as several older studies, such as DEFER and MASS. The meta-analysis, led by Bangalore, also included ISCHEMIA investigators David Maron, MD (Stanford University, CA), Gregg Stone, MD (Icahn School of Medicine at Mount Sinai, New York, NY), and Judith Hochman, MD (New York University School of Medicine), as coauthors.

William Boden, MD (Boston University School of Medicine/VA New England Healthcare System, MA), who led the COURAGE trial, wasn’t very impressed by the new data, particularly since the researchers pooled results from small trials and older studies from the 1990s and 2000s right up to the era of COURAGE and BARI 2D. ACME, for example, was published in 1992 and included 212 patients, while MASS was published in 1995 and included 214 patients. Many of these trials were published before optimal medical therapy (OMT) was established: before 2000, OMT consisted only of aspirin, nitrates, and beta-blockers, while coronary revascularization involved balloon angioplasty.

“To comingle data from almost 30 years in various trials, and to lump them with the current era of studies, which include COURAGE, BARI 2D, FAME 2, and ISCHEMIA—those are the four big trials in the post-2000 arena comparing apples to apples—is a bit of contaminated meta-analysis,” Boden told TCTMD. The new study, instead of helping, he argued, does little to inform “more enlightened decision-making in the current era,” he said.

Nearly 15,000 Patients

Overall, there were 14,877 patients included in the meta-analysis and patients were followed for a weighted mean of 4.5 years. Most of the studies included patients with preserved left systolic function and mildly symptomatic patients (typically CCS class I/II), and excluded those with left main CAD. For those that underwent revascularization, PCI was the first procedure in more than 70% of cases, while CABG surgery was the first option in 16.2% of patients (in the revascularization arm, any revascularization was performed in 87.5% of patients). In eight studies, stents were used in at least 50% of patients undergoing PCI while DES were used primarily in FAME 2, ISCHEMIA, and ISCHEMIA-CKD. During the 4.5 years of follow-up, 31.9% of patients in the medical therapy arm underwent revascularization.

With mortality, routine revascularization was not associated with a reduction in risk. The results were consistent when trials were broken down by era, including only analyzing studies that looked at the use of stents.

“It’s always a top question—is there a survival advantage with a therapy?” said Bangalore. “We also did a trial sequential analysis, which is to say, if you draw stopping boundaries like we do in clinical trials, where does it put us with these 15,000 patients? Interestingly, we showed the cumulative evidence as of now crosses the futility boundary. We have a great confidence to say that with these data it rules out a 10% difference in survival between revascularization and medical therapy.”

To comingle data from almost 30 years in various trials, and to lump them with the current era of studies . . . is a bit of contaminated meta-analysis. William Boden

Additionally, there was no reduction in risk of MI with revascularization across the 14 studies. There was a borderline significant reduction in the risk of MI with revascularization when investigators analyzed data from the contemporary stent era trials (RR 0.89; 95% CI 0.80-0.998). As noted, revascularization was associated with a reduction in the risk of nonprocedural MI (RR 0.76; 95% CI 0.67-0.85), but an increase in procedural events (RR 2.48; 95% CI 1.86-3.31) compared with medical therapy. There was also a reduction in the risk of unstable angina (RR 0.64; 95% CI 0.45-0.92), a benefit more pronounced in the contemporary trials, and an increased likelihood of freedom from angina (RR 1.10; 95% 1.05-1.15) with revascularization.  

The clinical significance of these early procedural MIs is a matter of debate, said Bangalore.

“I’ve always taken the opposite stance that they might not be as prognostically important, but it depends on the definition,” he said. One drawback of looking at these data is the variability in the definition of procedural MI used across the many trials, and while some procedural MIs might not be clinically relevant, large procedural events, such as MIs resulting from side branch occlusions, would be critical. On the whole, however, “it does appear as if there is a consistent signal that the nonprocedural MIs are more prognostically important,” said Bangalore.

Davide Capodanno, MD, PhD (University of Catania, Italy), said the new data will help physicians communicate to patients with stable ischemic heart disease the merits and drawbacks of revascularization and medical therapy. “We have seen, beyond any doubt, that there is no difference in mortality,” said Capodanno, the discussant following the presentation. “It’s a little bit more mixed, more nuanced, the story of myocardial infarction.”

What’s needed, said Capodanno, is a better understanding of the prognostic implications of the spontaneous and procedural MIs. “Of course, we know that spontaneous myocardial infarction is bad, but I’m not so sure about periprocedural MIs. Is this as important as spontaneous myocardial infarction? Obviously, this is a problem and we need to do better to make our procedures safer.”

With respect to freedom from angina, Capodanno said the meta-analysis shows benefit, a finding consistent with ISCHEMIA and other trials. Speaking elsewhere at the PCR e-Course in a state-of-the-art lecture on the merits of revascularization and medical therapy in chronic coronary syndromes, Capodanno said that while the ORBITA trial showed there is a placebo effect with PCI, that placebo effect is likely attenuated given the extended follow-up. Like Boden, he also said there are drawbacks to the meta-analysis, noting that it doesn’t include patient-level data and there is a lot of heterogeneity across the studies.

Longer Follow-up For ISCHEMIA

What’s important, Boden stressed to TCTMD, is that the meta-analysis showed there was no reduction in hard ischemic events with invasive revascularization, which is what they observed in COURAGE and what was seen in ISCHEMIA. “It’s only when you start dissecting the MIs into procedural and nonprocedural MI that you start to see these differences,” said Boden. “And we all recognize that the endpoint of procedural MI in ISCHEMIA was very much influenced by the definition that was used.”

Referring to the data as published in Circulation, Boden noted that there is no mention of the lack of an effect of revascularization on the overall MI endpoint in the abstract’s conclusions. “The reader is not left with the impression that MI is neutral, but rather that revascularization reduces the risk of spontaneous MI, which, I think, is a bit misleading given the vagaries of the MI definition,” he said. 

Finally, Boden stressed revascularization includes CABG surgery, and that some of the benefit of revascularization may have been driven by favorable outcomes in the surgical arm. In BARI 2D, for example, the combined endpoint of death, MI, or stroke was significantly reduced with CABG surgery versus medical therapy—driven by a reduction in MI—but there was no benefit for PCI. In FREEDOM, a study that included patients with multivessel disease and diabetes, CABG significantly reduced the risk of MI compared with PCI.

“That’s plausible biologically because you’re bypassing a large segment of both obstructed and nonobstructed coronary lesions,” said Boden. “I can understand why CABG would reduce spontaneous MI, but with PCI you’re stenting a stenotic segment and it does nothing to protect the rest of the vessel.”

To TCTMD, Bangalore said they are hopefully planning to follow patients in ISCHEMIA for an additional 5 years to evaluate any late benefit with revascularization. If invasive treatment reduces the risk of nonprocedural MI, you would expect to see a later benefit on mortality. “Potentially, we hope to answer it in the ISCHEMIA trial—we have a grant pending to follow these patients longer term,” he said. “We want to see if there is a late separation of the mortality curves. Will there be a late separation of the mortality curves as we get more and more benefit from the reduction in spontaneous MI?”

Boden, for his part, doesn’t think longer follow-up will show a survival advantage with the invasive approach. In COURAGE, even after 15 years, there was no mortality difference between the two treatment strategies, he said.    

Note: Stone is a faculty member of the Cardiovascular Research Foundation, the publisher of TCTMD.