Impact on guidelines
None of this was known at the time the European cardiology societies convened a committee to write their new guidelines on myocardial revascularization. The writing panel disagreed about whether PCI and CABG were equivalent for patients with left main coronary artery disease (CAD).
Besides EXCEL, another study, the NOBLE trial, compared PCI and CABG in left main CAD and came to opposite conclusions – conclusions that matched the leaked data. In that trial, European investigators chose a slightly different primary endpoint: a composite of death, MI, stroke, and the need for a repeat procedure. They used the universal definition of MI exclusively, and notably, they omitted procedural MI from their clinical event count. The results, published at the same time as the EXCEL 3-year findings, suggested that CABG was better.
Given the discrepant findings of two large trials, the guideline committee considered all of the available data comparing the two methods of revascularization for left main CAD. But even then, things weren’t clear-cut. One draft meta-analysis, supported by the National Institute for Health Research, suggested that results were worse for first- and second-generation drug-eluting coronary stents – including those used in EXCEL – compared with surgery.
Another meta-analysis, later published in The Lancet, drew a different conclusion and found that PCI was just as good as surgery. The main author, Stuart Head, a cardiothoracic surgeon on the ESC/EACTS guideline committee, was a research fellow with EXCEL investigator Dr. Kappetein at Erasmus. EXCEL investigators Dr. Stone, Dr. Kappetein, and Dr. Serruys were coauthors of the Lancet meta-analysis.
There was heated discussion about the committee’s draft recommendations, which gave both CABG and PCI a Class IA recommendation in patients with left main CAD and low anatomical complexity. In October 2017, the ESC commissioned an anonymous external reviewer to weigh in. James Brophy, MD, PhD, a cardiologist and professor of medicine and epidemiology at McGill University, Montreal, confirmed that he was the reviewer after he published an updated version in June 2020.
Looking at all of the data available at the time comparing the procedures for left main CAD, Dr. Brophy’s analysis suggested a 73% chance that the excess in death, stroke, or MI represents at least two excess events per 100 patients treated with PCI rather than CABG.
Dr. Brophy thought that most patients would find these differences clinically meaningful and advised against giving both procedures the same class of recommendation. He was also concerned that many readers will skip to the summary recommendation table without reading the entire guideline document.
“I feel this is misleading in its present form,” he wrote in 2017.
Despite Dr. Brophy’s review, the guideline committee stuck with its original recommendations. The final 2018 ESC/EACTS Guidelines on myocardial revascularization gave equal weight to both CABG and PCI in patients with left main CAD and low anatomical complexity. In contrast, US guidelines do not put PCI and CABG on the same footing for any group of patients with left main CAD.
The lead author of the ESC/EACTS guidelines section on left main disease, and around a third of those on the writing task force, all declared financial payments from stent manufacturers either to themselves or their institutions. The EXCEL principal investigator, Dr. Kappetein, was secretary general of EACTS and oversaw the guidelines process for the surgical organization. He left to work for Medtronic midway through the process and was later joined there by his former research fellow, Stuart Head.
Dr. Brophy said in an interview that given the final guideline recommendations, he assumed that the committee had other reviews and went with the majority opinion.
But not everyone involved in the guidelines saw Dr. Brophy’s review. Nick Freemantle, a statistical reviewer appointed by EACTS, expected to see it but didn’t. This omission calls into question the neutrality of the whole process, in his view.
Mr. Freemantle believes that the deck was stacked so that he only saw the pieces of evidence that supported the conclusions that were already decided and that he was not shown “the bits that don’t fit that neatly.”
“And without that narrative, it all feels a bit grubby, to be honest,” he said.
Professor Barbara Casadei, ESC president, disputed this, saying that the guidelines were approved by all surgical members, including the EACTS council.
Missing from Dr. Brophy’s review were the later data from EXCEL. As he had told the DSMB in 2017, Stone presented the 4-year data from EXCEL at the TCT conference in September 2018. At this point, the analysis showed that 10.3% of people had died after PCI and 7.4% after CABG.
But this presentation was not given much prominence at the conference, which Dr. Stone organized, and occurred during a didactic session in a small room rather than on one of the main stages where the 3-year data from EXCEL were announced with much fanfare. The presentation also took place 3 weeks after the European guidelines were published.