Surgery Outdoes Percutaneous Treatment for Patients With CAD, Aortic Stenosis

LOS ANGELES, CA—Among patients needing both aortic valve replacement and coronary revascularization, surgery is associated with lower morbidity and mortality compared with percutaneous methods, according to Medicare data from 2018 to 2022.

J. Hunter Mehaffey, MD (West Virginia University, Morgantown), who presented the results at the 2025 Society of Thoracic Surgeons (STS) Annual Meeting, told TCTMD that CABG remains "a very important therapy for coronary disease,” and that it should be considered when there is equipoise between transcatheter and surgical aortic valve replacement.

These new observational data contradict the results from the prematurely stopped TCW randomized trial, where major adverse events with SAVR and CABG were significantly higher than with TAVI and PCI. Some, however, criticized TCW for being too small and including life-threatening or disabling bleeding in its primary endpoint, which may have biased the study toward favoring percutaneous treatment.

Though the new Medicare analysis does show, as expected, higher perioperative risks of acute kidney injury (AKI) and bleeding with surgery compared with transcatheter treatment, it highlights a “real difference” in procedural mortality as well as advantages in terms of need for pacemaker and vascular complications, Mehaffey continued.

“It’s very important that we not focus just on the intervention, but also how that patient is doing 5 years from now,” he said. “If we can have fewer heart attacks, fewer episodes of heart failure readmission, fewer strokes, and superior risk-adjusted survival, . . . that’s an important consideration. These data need to be part of that conversation.”

Study co-author Vinay Badhwar, MD (West Virginia University), told TCTMD that the mortality rate observed with surgery in TCW—10% versus 0% with TAVI/PCI—was “fairly excessive and not necessarily representative of broadscale, real-world clinical practice.” Surgeons want safe operations, done less invasively when possible, “but we also want the long-term survival of the patient,” he added. In coming from a reliable, longitudinal registry, the Medicare data provide a strong accounting of current practice, Badhwar said.

Nearly 38,000 Patients

For the study, Mehaffey and colleagues included 37,822 Medicare beneficiaries who underwent PCI/TAVI (n = 17,413) or CABG/SAVR (n = 20,409) between 2018 and 2022.

Most of the percutaneously treated patients (83.3%) received elective PCI 3 months before TAVI, with 13.4% undergoing nonemergent PCI during the index TAVI admission and only 3.3% of patients undergoing PCI within 3 months following TAVI. Single versus multivessel revascularization was balanced in the CABG/SAVR cohort (53.8% vs 46.2%), but the PCI/TAVI group skewed heavily toward single-vessel revascularization (90.4% vs 9.6%).

In risk-adjusted analyses, the index admission for patients treated with PCI/TAVI compared with CABG/SAVR was linked to less major bleeding (OR 0.72; P < 0.0001), AKI (OR 0.25; P < 0.0001), and in-hospital mortality (OR 0.43; P < 0.0001), but there was a greater risk of need for new pacemaker (OR 1.59; P < 0.0001) and surgical repair of the femoral artery (OR 7.1; P < 0.0001). There was no difference seen for stroke (OR 1.00; P = 0.975).

A longitudinal analysis showed higher risks of readmission for stroke (HR 1.10; P = 0.024) with PCI/TAVI compared with CABG/SAVR, as well as increases in MI (HR 1.68; P < 0.0001), all-cause mortality (HR 1.09; P < 0.0001), and the composite outcome of stroke, MI, valve reintervention, or death (HR 1.26; P < 0.0001).

In light of the TCW trial and these new data, Mehaffey touched on the need for randomized, controlled trials in patients with coronary disease and aortic stenosis.

“We're all very excited about the highest-quality randomized evidence,” but future trials in this space should be guided by certain considerations, Mehaffey said. “It really has to have a pragmatic design. And I think if we get too much into the weeds of defining a SYNTAX score and defining this and that, one, is the trial going to be enrollable? And number two, is it going to be generalizable to the patient that shows up?”

Badhwar said they are starting to implement a validated metric of frailty, “which is something that has been missing from a lot of registry and database studies that can help provide insight and account for some of those differences in the selection bias that is inherent to these types of studies.”