WASHINGTON – Results from two large, multicenter comparisons of coronary stenting and coronary bypass surgery for treating patients with unprotected left main coronary disease may have superficially shown sharp differences, but the bottom line will likely be greater empowerment of percutaneous coronary intervention as an option for selected patients with less complex coronary disease.

Prior to the results from the EXCEL and NOBLE trials, reported at the Transcatheter Cardiovascular Therapeutics annual meeting, “guidelines put PCI [percutaneous coronary intervention] into a class 1, 2 or 3 status for treating left main coronary disease depending on disease complexity, but in the United States, PCI for patients eligible for CABG [coronary artery bypass grafting] has not been frequently done. I think these results, in a very circumscribed subset of patients and using a state-of-the-art stent, will affect the guidelines,” predicted Gregg W. Stone, MD, lead investigator for EXCEL and professor of medicine at Columbia University in New York.

“What the guidelines have not addressed are the patients with low- or intermediate-complexity disease who have an acceptable risk for undergoing either PCI or CABG. I think the trial results answer this question,” said David Kandzari, MD, director of interventional cardiology and chief scientific officer at the Piedmont Heart Institute in Atlanta and an EXCEL investigator.

While the EXCEL and NOBLE results don’t provide a simple answer on the relative merits of PCI and CABG, many of their outcome differences seem explicable, several experts said at the meeting.

The Nordic-Baltic-British Left Main Revascularisation (NOBLE) trial randomized 1,201 patients who had unprotected left main coronary disease and were judged by a heart team to be reasonable candidates for both PCI or CABG at 36 centers in nine European countries during 2008-2015. The primary endpoint was death from any cause, nonprocedural MIs, stroke, or repeat revascularization.

The researchers followed patients for a median of just over 3 years, but they calculated the primary endpoint based on a Kaplan-Meier estimate for 5-year outcomes, which showed the primary endpoint in 29% of the PCI patients and in 19% of the CABG patients, a statistically significant benefit in favor of CABG, Evald H. Christiansen, MD, reported at the meeting, which was sponsored by the Cardiovascular Research Foundation. Concurrently with his report, the results were published online (Lancet. 2016 Oct. 31. doi: 10.1016/S0140-6736[16]32052-9).

This difference between PCI and CABG was largely driven by an excess of postprocedural MIs and repeat revascularizations among the PCI patients, said Dr. Christiansen, an interventional cardiologist at Aarhus University Hospital in Denmark. Another notable finding was that the superior outcomes with CABG primarily occurred among patients with a SYNTAX score – a measure of coronary disease complexity – of 22 or less, which identifies patients with low complexity disease. The outcomes of patients with SYNTAX scores of 23-32, which identifies intermediate complexity disease, and of patients with scores of 33 or higher, with very complex disease, were similar in the PCI and CABG arms, he reported. This finding was “very surprising,” Dr. Christiansen said, because it reversed the finding originally made in the SYNTAX trial that PCI performed best compared with CABG in patients with the lowest scores and least-complex coronary disease.

The superiority of CABG over PCI seen in the NOBLE results, especially in patients with lower SYNTAX scores, seemed at odds with the EXCEL results, reported at the meeting by Dr. Stone and simultaneously online (N Engl J Med. 2016 Oct. 31. doi: 10.1056/NEJMoa1610227). In EXCEL, which enrolled only patients with a SYNTAX score of 32 or less (low- or intermediate-complexity coronary disease), patients had a 3-year incidence of death, stroke or MI of 15% in both the PCI and CABG arms.

But the EXCEL and NOBLE trials had several important differences, and it seemed like cumulatively these differences account for their differing results.

“One of the biggest differences” was the exclusion of procedural MIs in the NOBLE tally of adverse events, noted Dr. Stone. These were diagnosed in EXCEL using the MI definition published in 2013 by a panel of the Society for Cardiovascular Angiography and Interventions (SCAI). NOBLE disregarded procedural MIs because many of its participating hospitals did not have the laboratory resources to make these diagnoses and because the trial’s design predated the SCAI definition by several years, Dr. Christiansen explained.

Dr. Gregg W. Stone (left) and Dr. Evald H. Christiansen

Other important differences included the shorter follow-up in EXCEL, the inclusion of revascularization as an endpoint component in NOBLE but not in EXCEL, and differences in the stents used. In EXCEL, all patients undergoing PCI received Xience everolimus-eluting stents. In NOBLE, the first 11% of the enrolled patients received first-generation, sirolimus-eluting Cypher stents; the next 89% of enrollees received a biolimus-eluting Biomatrix Flex stent. Dr. Christiansen acknowledged the confounding caused by having many patients in the NOBLE PCI arm who received outmoded Cypher stents, especially because their relatively longer follow-up made them overrepresented in the primary outcome results. Plus, the Biomatrix Flex stent was disparaged by Martin B. Leon, MD, an EXCEL investigator and professor of medicine at Columbia University, who called the device “not currently widely used for PCI and more of historic interest.”

Dr. Leon added that the EXCEL and NOBLE patients also had substantially different prevalence rates of diabetes and acute coronary syndrome.

“The huge difference [between EXCEL and NOBLE] is the endpoint,” declared Marc Ruel, MD, another EXCEL investigator and head of cardiac surgery at the Ottawa Heart Institute. “The EXCEL endpoint was driven by the high rate of periprocedural MIs in the CABG arm. Once you get past 30 days, the noninferiority is not met by PCI.”

Another big endpoint difference was leaving revascularizations out of the EXCEL composite. “Once you take revascularization out of the primary endpoint, the outcome [of EXCEL] was more or less preordained,” noted Craig R. Smith, MD, chairman of surgery at Columbia University and an EXCEL investigator. “It’s the slope of events [in the PCI arm] after 3 months that’s the story. I think the CABG and PCI slopes in EXCEL will continue to diverge with time” beyond the current 3-year follow-up, Dr. Smith said.

“I agree that after 30 days surgery was the more durable procedure,” said Dr. Stone. “There is a big upfront hit for patients to take with surgery compared with PCI. If patients get through that, then they have a more durable procedure. That’s the trade-off.”

Dr. Stone hinted that future reports of EXCEL data will highlight other hits that patients must endure upfront if they choose CABG over PCI. “The early difference was quite profound not only for the primary endpoint but also for renal failure, infections, arrhythmias, and blood transfusions,” he said. Choosing between PCI and CABG for patients with left main disease and a lower SYNTAX score “is a decision that should be made by the heart team and patients. Some patients will prefer surgery, and some will prefer PCI.”

The NOBLE trial received partial funding from Biosensors, the company that markets the Biomatrix Flex stent used in the trial. The EXCEL trial was sponsored by Abbott Vascular, the company that markets the Xience stent used in the trial. Dr. Stone, Dr. Kandzari, Dr. Christiansen, Dr. Ruel, and Dr. Smith had no disclosures. Dr. Leon has been a consultant to and received research support from Abbott Vascular and Boston Scientific and has also received research support from Edwards, Medtronic and St. Jude.

mzoler@frontlinemedcom.com

On Twitter @mitchelzoler

VIDEO: Two PCI vs. CABG trials produce conflicting results