Severe aortic stenosis (AS) represents a prevalent valvular heart disease in the elderly, with incidence increasing significantly with age (1). Studies have reported that 15% to 81% of individuals with AS have concomitant coronary artery disease (CAD), with a higher prevalence observed in older populations (2–4). These patients experience increased cardiovascular risk due to impaired myocardial perfusion and left ventricular dysfunction (5, 6). Determining the optimal coronary revascularization approach in individuals undergoing aortic valve replacement remains a complex and unresolved clinical issue, particularly in those with advanced age and extensive coronary disease.

Traditionally, surgical aortic valve replacement (SAVR) with coronary artery bypass grafting (CABG) has been considered the standard approach for eligible individuals, as it provides complete revascularization and durable valve function. Accordingly, SAVR combined with CABG has historically been favored in younger patients and those with acceptable surgical risk profiles (5, 6). However, SAVR + CABG is associated with higher surgical trauma, prolonged recovery, and increased perioperative complications, particularly in intermediate- and high-risk individuals. Recent advances in transcatheter aortic valve replacement (TAVR) and percutaneous coronary intervention (PCI) have facilitated the widespread adoption of TAVR + PCI as a minimally invasive alternative, especially in elderly and high-risk populations (3, 7). According to the 2025 ESC/EACTS Guidelines for the Management of Valvular Heart Disease, TAVR is recommended for anatomically suitable patients aged ≥70 years with tricuspid severe AS, while the choice between TAVR and SAVR in other candidates should be guided by a multidisciplinary Heart Team integrating clinical risk, anatomical feasibility, imaging findings, and patient preference (6). These evolving treatment strategies underscore the need for comparative evidence on the clinical outcomes of TAVR + PCI vs. SAVR + CABG in individuals with AS and concomitant CAD (8, 9).

Several studies have evaluated TAVR + PCI, with some comparing it to SAVR + CABG (10–13). However, most evidence has focused on short-term outcomes (e.g., mortality, stroke, acute kidney injury), while evidence beyond the early post-procedural period remains limited and heterogeneous (14–17). Although some studies have reported outcomes beyond two years, their findings are highly heterogeneous. From a clinical perspective, outcomes beyond two years are more appropriately considered mid-term rather than definitive long-term endpoints, particularly in the context of valve durability and coronary disease progression (17). Additionally, substantial heterogeneity exists across studies regarding patient risk stratification, revascularization strategies, and study designs, contributing to variability in reported outcomes (14). In particular, for patients with complex CAD (e.g., high SYNTAX scores), PCI may result in incomplete revascularization, potentially compromising mid- to long-term prognosis, whereas CABG may provide more sustained hemodynamic benefits (18). Against this background of heterogeneous and incomplete primary evidence, a rigorous synthesis of available data remains essential to clarify the relative benefits and limitations of contemporary interventional and surgical strategies.

Several meta-analyses have previously synthesized evidence comparing TAVR + PCI with SAVR + CABG in patients with AS and concomitant CAD, including those by Sakurai et al. (14), Baudo et al. (19), and Emara et al. (20). While these studies provided important insights into perioperative outcomes and overall survival, they differed substantially in inclusion periods, surgical risk profiles, follow-up definitions, and analytical approaches. In particular, most prior syntheses pooled heterogeneous patient populations across a broad risk spectrum or emphasized aggregate outcomes without predefined temporal stratification, thereby limiting their applicability to contemporary, guideline-based decision-making in intermediate- and high-risk patients. In contrast, the present study focuses specifically on intermediate- and high-risk individuals, as operationally defined by contemporary surgical risk models and clinical characteristics, and applies a structured comparison across predefined short-term and ≥2-year follow-up horizons. By incorporating evidence exclusively from randomized controlled trials and high-quality cohort studies, this meta-analysis aims to reduce clinical heterogeneity, enhance temporal interpretability of outcomes, and align evidence synthesis with current clinical practice patterns and evolving guideline recommendations.

This meta-analysis aims to systematically compare short- and mid- to long-term outcomes of TAVR + PCI vs. SAVR + CABG in intermediate- and high-risk patients with severe AS and concomitant CAD. By incorporating evidence from randomized controlled trials (RCTs) and cohort studies, we seek to address existing heterogeneity and generate clinically relevant, methodologically rigorous conclusions to inform treatment selection, guide personalized decision-making, and support future updates to practice guidelines.

This meta-analysis included 13 studies with a total of 53,869 intermediate- and high-risk individuals with severe AS and concomitant CAD. The analysis systematically compared short- and mid- to long-term outcomes between TAVR + PCI and SAVR + CABG. In the short term, TAVR + PCI was associated with lower risks of stroke, myocardial infarction, and acute kidney injury, with no significant difference in all-cause mortality. However, the incidence of permanent pacemaker implantation was notably higher. Over mid- to long-term follow-up (≥2 years), TAVR + PCI was associated with increased risks of all-cause mortality, myocardial infarction, and repeat revascularization, whereas stroke risk remained comparable. Subgroup analyses confirmed the short-term advantages of TAVR + PCI across study designs and sample size stratifications. However, in studies with extended follow-up durations (≥5 years), TAVR + PCI showed a persistently higher risk of all-cause mortality and myocardial infarction during follow-up, raising concerns regarding mid- to long-term durability. The GRADE assessment indicated that evidence quality ranged from very low to moderate.

These short-term findings are consistent with the PARTNER 2 (38) and SURTAVI (2) trials, which demonstrated comparable short-term mortality between TAVR and SAVR, with reduced stroke incidence favoring TAVR. However, these trials did not specifically address concomitant CAD requiring revascularization. By incorporating evidence from RCTs and high-quality cohort studies, the present analysis further supports the short-term safety benefits of TAVR + PCI. Similar results were reported by Sakurai et al. (14), who identified no significant difference in 30-day mortality but lower acute kidney injury and higher pacemaker implantation rates with TAVR + PCI. Notably, that analysis lacked stratification by surgical risk, potentially contributing to heterogeneity. By focusing specifically on intermediate- and high-risk individuals, this meta-analysis additionally identified reduced risks of stroke and myocardial infarction with TAVR + PCI, findings not clearly emphasized in previous meta-analyses. Among the included studies, the TCW trial by Kedhi et al. (26) provided RCT-level evidence supporting favorable short-term outcomes with TAVR + PCI in elderly patients with anatomically suitable CAD, though limited by short follow-up and exclusion of patients with complex coronary anatomy. The short-term benefits of TAVR + PCI may be attributed to its minimally invasive nature, avoiding cardiopulmonary bypass and reducing procedural trauma, thereby lowering the risks of stroke, hypoperfusion-related injury, and acute kidney injury. PCI also facilitates timely revascularization, potentially reducing perioperative myocardial infarction. The increased rate of permanent pacemaker implantation following TAVR is likely related to atrioventricular node injury, particularly with self-expanding valves, as documented in prior studies.

Regarding mid- to long-term outcomes, the present findings align with the EXCEL (39) and SYNTAX (40) trials, which demonstrated superior mid- to long-term outcomes with CABG compared to PCI in patients with complex CAD. Although these trials did not include patients undergoing aortic valve intervention, the current analysis extends their conclusions to a high-risk AS population. The increased rates of myocardial infarction and repeat revascularization observed with TAVR + PCI may reflect incomplete revascularization or disease progression in untreated segments. In contrast, CABG provides durable distal perfusion via bypass grafts, which may contribute to mid- to long-term myocardial protection. Valve durability may also influence mid- to long-term outcomes. While early-generation TAVR valves showed structural deterioration within 5–8 years, surgical bioprosthetic valves typically last 10–15 years (41). However, more recent studies (e.g., PARTNER 2/3 and FRANCE-TAVI studies) (38, 42, 43) suggest that newer transcatheter valves exhibit improved durability, though further validation through mid- to long-term follow-up is necessary. Paravalvular leak following TAVR may lead to chronic volume overload, potentially affecting ventricular remodeling and mid- to long-term survival. Despite the worse mid- to long-term outcomes, stroke incidence during follow-up did not differ significantly between groups, possibly reflecting differences in antithrombotic strategies, baseline atrial fibrillation burden, and sample size.

Importantly, interpretation of these findings should be contextualized within the framework of the 2025 ESC/EACTS Guidelines (6) for the management of valvular heart disease, which emphasize individualized, Heart Team–based decision-making that integrates clinical risk, anatomical complexity, imaging findings, and patient preference, rather than relying solely on chronological age or surgical risk scores. In this context, the present results underscore that while TAVR + PCI may offer short-term procedural advantages, particularly in frail or high-risk patients, it may be associated with less favorable mid- to long-term outcomes in individuals with complex coronary anatomy or a longer life expectancy. Accordingly, patients with multivessel or anatomically complex CAD, in whom complete revascularization is more reliably achieved with CABG, may derive greater mid- to long-term benefit from SAVR + CABG, provided surgical risk is acceptable. Conversely, it is equally important to recognize clinical scenarios in which TAVR + PCI may not be recommended or technically feasible, including unfavorable annular or aortic root anatomy, inadequate vascular access, the need for concomitant surgical procedures (e.g., ascending aortic repair or other valve interventions), or specific valve morphologies such as certain bicuspid aortic valve phenotypes. These considerations reinforce that TAVR + PCI should not be viewed as a universal substitute for surgery but rather as one component of a tailored treatment strategy. At the same time, surgical treatment may be infeasible in selected patient subsets—such as those with porcelain aorta, prohibitive redo sternotomy risk, or advanced frailty—in whom TAVR + PCI may represent a feasible, and in some cases the only practical, option for combined valve intervention and coronary revascularization. The present findings therefore highlight the necessity of balancing early procedural safety against longer-term risks when selecting revascularization and valve strategies in this complex population.

This meta-analysis provides an updated and comprehensive synthesis comparing TAVR + PCI and SAVR + CABG in this intermediate- and high-risk patients with severe AS and concomitant CAD. Methodological rigor—including adherence to PRISMA guidelines, GRADE assessment, and pre-specified subgroup and sensitivity analyses—strengthens the robustness of the findings. Nonetheless, several limitations warrant discussion. First, substantial heterogeneity was observed across included studies in terms of patient characteristics, procedural details (e.g., valve type, access route), and operator experience, which may have influenced effect estimates. Second, PCI timing relative to TAVR varied widely, ranging from staged to simultaneous procedures, and was inconsistently reported, limiting the feasibility of stratified analyses by timing. Third, detailed coronary anatomical data (e.g., SYNTAX score, left main disease) were inconsistently reported, constraining lesion-complexity–based comparisons. Fourth, although HRs were prespecified as the preferred effect measure for mid- to long-term outcomes, most included studies only reported crude event counts and did not provide HRs or extractable Kaplan–Meier curves. Consequently, ORs derived from event numbers were used as an alternative, which may not fully capture the temporal dimension of outcomes and could have influenced the precision of mid- to long-term risk estimates. Fifth, outcome definitions were not fully standardized across studies (e.g., some applied VARC-2 or VARC-3 criteria, while others did not specify criteria). Although we systematically summarized all reported definitions in Supplementary Tables S3 and S4, such inconsistency may have contributed to residual heterogeneity, particularly for revascularization, where most studies did not distinguish target-lesion vs. any-vessel events, planned vs. unplanned staged procedures, or graft-failure–related reinterventions. Sixth, while procedural and anatomical effect modifiers (e.g., access route, valve type, PCI timing, completeness of revascularization, and SYNTAX strata) were considered a priori, their implementation was limited by non-uniform reporting and lack of harmonized definitions across studies; as a result, only consistently available subgroup variables—follow-up duration (<5 vs. ≥5 years) and study size (<300 vs. ≥300)—could be analyzed uniformly, with study design explored where feasible. Seventh, although two RCTs were included, the evidence base was predominantly derived from observational studies and therefore remains vulnerable to selection bias and residual confounding, even when adjustment techniques—such as propensity score matching or multivariable regression—were applied at the individual study level in some cohorts. Lastly, the overall certainty of evidence ranged from low to very low, primarily due to inconsistency and imprecision, underscoring the need for additional high-quality data. In the context of the 2025 ESC/EACTS Guidelines, these limitations highlight the need for cautious interpretation when extrapolating comparative effectiveness to individual clinical decision-making. Future research should therefore prioritize well-powered randomized trials with extended follow-up, standardized time-to-event outcome reporting, and stratification by coronary anatomy and clinical risk. Equally important, harmonized reporting of procedural and anatomical variables—including access route, valve type, PCI timing with explicit staged vs. concomitant definitions, completeness of revascularization, and quantitative SYNTAX categories—as well as consistent adoption of standardized outcome definitions (e.g., VARC-3), will be essential to enhance comparability across studies. Meta-analyses incorporating patient-level data and guideline-aligned outcome frameworks may further improve evidence quality and clinical applicability.

TAVR + PCI is associated with lower early periprocedural complications but may carry higher mid- to long-term risks compared with SAVR + CABG, although the certainty of evidence is limited by observational data and heterogeneity. Treatment decisions should be individualized, considering CAD complexity, revascularization completeness, valve type, and procedural timing. These findings underscore the importance of balancing early procedural safety against potential mid- to long-term risks when selecting treatment strategies in patients with severe AS and concomitant CAD. Further randomized trials with standardized outcome definitions and extended follow-up are warranted.