Revascularization After PCI or CABG Linked to Higher Mortality: EXCEL

The need for a repeat procedure following revascularization of the left main coronary artery was more common among patients initially treated with PCI versus CABG surgery in the EXCEL trial, but repeat revascularization—regardless of the initial treatment strategy—was associated with a significantly increased risk of death, according to a new analysis.

At 3 years, the need for any repeat revascularization following PCI or CABG surgery for left main coronary artery disease was associated with an increased risk of all-cause mortality (HR 2.05; 95% CI 1.13-3.70) and a more than fourfold increased risk of cardiovascular mortality (HR 4.22; 95% CI 2.10-8.48). The magnitude of the mortality risk associated with revascularization was smaller those associated with MI or stroke during follow-up.

“It doesn’t matter if you originally had a PCI or CABG, if you require a repeat revascularization overall, that’s associated with an increased risk of mortality,” lead investigator Gennaro Giustino, MD (Icahn School of Medicine at Mount Sinai, New York, NY), told TCTMD.

Published January 15, 2020, in the Journal of the American College of Cardiology, the analysis was meant to evaluate the incidence, timing, and prognostic impact of revascularization with PCI or CABG for left main coronary artery disease, Giustino said. Of note, the analysis involves 3-year follow-up and not the 5-year results that have been making headlines for the past month.

The aim here was to zero in on repeat procedures, Giustino stressed. “Repeat revascularization as an endpoint is often criticized or considered less important than death, MI, or stroke,” he said. “At the same time, it’s also recognized that it’s a biased endpoint to include as part of the primary endpoint of a trial comparing PCI versus CABG surgery. . . . That being said, it’s still a key secondary endpoint to analyze in a coronary revascularization trial.”

David Cohen, MD (University of Missouri-Kansas City), who wasn’t involved in the EXCEL trial, said repeat revascularization has historically been considered a relatively benign or nuisance event, one mainly related to the patient’s quality of life. Whether this was the case in complex, high-risk anatomy, such as those with left main coronary artery disease, was uncertain. 

Overall, Cohen said this analysis reinforces the idea that repeat revascularization is “part of the story” supporting the favorable long-term prognosis after CABG surgery in this and other patient populations. For that reason, it gives interventional cardiologists a target for improving clinical practice, he asserted. “If we can improve the stents, improve their geometry, improve the technique, and lower the need for repeat revascularization after PCI further, the logical conclusion is that we could narrow any survival difference between PCI and CABG.”

Led by Gregg Stone, MD (Icahn School of Medicine at Mount Sinai), and A. Pieter Kappetein, MD, PhD (now Chief Medical Officer and Vice President at Medtronic), EXCEL was a 1,905-patient study comparing PCI with a metallic everolimus-eluting stent (Xience, Abbott Vascular) versus CABG surgery in patients left main coronary artery disease. Three-year and 5-year results showed no statistically significant difference in the composite rate of death, stroke, or MI between PCI- and CABG-treated patients. The 5-year results, however, showed a survival advantage with CABG that prompted a withering attack on the EXCEL investigators, leading in turn to considerable debate between surgeons and interventionalists.

EXCEL to 3 Years

In EXCEL, the overall incidence of repeat revascularization was 9.7%, or 346 repeat revascularization procedures in 185 patients: 12.9% of patients initially treated with PCI required a repeat procedure compared with 7.6% of patients treated with CABG surgery (HR 1.73; 95% CI 1.28-2.33). The median time to repeat revascularization was 347 days after PCI and 257 days after CABG surgery. Higher body mass index, insulin-treated diabetes, and the need for hemodynamic support during the initial procedure each were associated with an increased risk of repeat revascularization after PCI, while younger age, female sex, and peripheral vascular disease were independent predictors of a repeat procedure in those initially treated with surgery.

In their multivariate-adjusted model, the researchers attempted to tease out the independent effects of not only repeat revascularization, but also the prognostic impact of MI, stroke, age, diabetes, and anemia. In these analyses, the magnitude of the association between repeat revascularization and all-cause and cardiovascular death was smaller than that of MI or stroke. For example, individuals who had an MI after coronary revascularization for left main disease had a risk of all-cause mortality more than four times greater than those who didn’t have an MI (HR 4.03; 95% CI 2.43-6.67), while the risk of death was considerably larger in patients who had a stroke during follow-up.

The reason for the smaller prognostic impact is that “any” revascularization lumps all repeat procedures together, but their analysis showed that not all repeat revascularizations are equal, said Giustino. Repeat revascularization performed for recurring disease at a lesion or artery previously treated was associated with an increased risk of all-cause and cardiovascular mortality. In contrast, neither non-target-lesion nor non-target-vessel revascularizations were linked to a higher risk of death.

“It makes sense, especially if you’re talking about a patient with left main disease because what you really care about is not having to re-intervene after PCI on the left main artery,” said Giustino. The prognostic impact of repeat revascularization, as well as its subtypes, was evident in patients initially treated with PCI or CABG, with no evidence of interaction between treatments.

To TCTMD, Cohen said the message from these observations is that recurring left main coronary artery disease, or ostial disease of the LAD or circumflex arteries, “carries a fair amount of prognostic weight whereas other lesions in more distal coronary territories are less important.” Like Giustino, Cohen said the finding is plausible given that the left main artery supplies a large percentage of the left ventricular myocardium. “If you have problems at any point in the left main, it can be fatal,” he said.

Although most repeat procedures were performed with PCI in both the PCI and CABG groups, the researchers showed that repeat revascularization with CABG surgery was associated with a significantly increased risk of all-cause death (HR 5.61; 95% CI 2.45-12.83), whereas reinterventions with PCI were not (HR 1.60; 95% CI 0.84-3.04).

Threshold for CABG Patients Is Higher

Giustino said that at the time they conducted the analysis and submitted the paper, the much-debated 5-year results weren’t available, but he doesn’t suspect their data would change substantially with longer follow-up. The increased risk of death following repeat revascularization peaked at 30 days after the repeat procedure and then declined, note the investigators.   

In an editorial, David Williams, MD, and Pinak Shah, MD (both from Brigham and Women’s Hospital, Boston, MA), write that nearly all of the deaths in the subgroup requiring repeat revascularization were attributable to cardiovascular causes and most of these occurred in the first month, which means “it is likely they were related to the repeat revascularization procedure.” Because the mortality risk was only observed in patients requiring target-lesion and target-vessel revascularization, “this finding suggests that it is not the PCI procedure itself that adds risk but rather aspects of the target lesion or vessel.”   

The EXCEL researchers state that durability of the index procedure is an optimal goal, but as Giustino pointed out the need for repeat revascularization represents a heterogenous condition. It could be a stent-related problem, which might be offset with optimal implantation techniques during the index procedure and the use of intravascular imaging, or it may involve the progression of lesions elsewhere in the coronary anatomy. For the latter, guideline-directed secondary-prevention therapy, such as the use of high-intensity statins, should be prescribed to slow the progression of atherosclerosis.

To TCTMD, Cohen said the rate of repeat revascularization with PCI in EXCEL, although higher than with CABG surgery, is comparable to other clinical trials looking at revascularization in left main disease. “It should actually be a little bit lower in EXCEL, because they excluded the patients with the highest SYNTAX scores by design,” he said. “Obviously, they’re using a stent that’s current generation and as good as any stent that we have for it.”

He also noted that physicians tend to have a much higher threshold for sending heart surgery patients for a repeat procedure. Using data from the SYNTAX trial of PCI versus CABG, Cohen, along with Suzanne Arnold, MD (Saint Luke’s Mid America Heart Institute, Kansas City), published data showing that surgical patients experienced a much greater drop-off in quality of life before they were referred for a repeat procedure.

“The threshold for people reporting their symptoms is higher after CABG and the threshold for doing something about it is undoubtedly higher after CABG,” said Cohen. “The reasons for it are speculative, but it’s likely a combination of factors. Some degree of chest pain after CABG may be attributed to the incision or other things, so people might ignore it. Secondarily, we also know that very commonly after bypass surgery the anatomy is distorted and less amenable to PCI.” Given that, as well as the complexity of disease progression in these patients, doctors may be more likely to manage the patient medically, he said.