Coronary artery disease (CAD) is commonly found in patients with severe aortic stenosis (AS) and combined surgical aortic valve replacement (SAVR) and coronary artery bypass grafting (CABG) is currently recommended as the preferred treatment in this setting. Transcatheter aortic valve replacement (TAVR) and percutaneous coronary intervention (PCI) represent a valid alternative. This study sought to investigate clinical outcomes after TAVR+PCI versus SAVR+CABG in patients with severe AS and CAD. A systematic review was conducted from inception until December 2024 for randomized controlled and propensity score-matched studies comparing TAVR+PCI and SAVR+CABG for patients with severe AS and CAD. Kaplan–Meier-derived individual patient data was retrieved when available. Primary endpoint was all-cause mortality. The study was registered with PROSPERO (CRD42025642206). Six studies met our inclusion criteria with a total of 1,998 patients: 1,007 in the TAVR+PCI group and 991 in the SAVR+CABG group. The hazard ratio of all-cause mortality varied over time between groups: TAVR+PCI showed a lower incidence of all-cause mortality in the first 19 days, with a reversal at 73 days favoring SAVR+CABG. Patients undergoing TAVR+PCI group experienced lower rates of stroke-free survival (p = 0.039), postoperative reintervention (p = 0.020), atrial fibrillation (p <0.001), and acute kidney injury (p = 0.001) rates, while they were at higher risk of postoperative moderate/severe aortic regurgitation (p <0.001), permanent pacemaker implantation (p = 0.005) and major vascular complication (p <0.001). Major bleeding didn’t differ (p = 0.358). A percutaneous treatment approach offered an early survival benefit over a surgical approach in patients with severe AS and CAD but was associated with worse survival at mid-term follow-up.
Coronary artery disease (CAD) is frequently present in patients suffering from severe aortic stenosis (AS). The preferred treatment for these patients is currently concomitant surgical aortic valve replacement (SAVR) and coronary artery bypass grafting (CABG). , However, concomitant SAVR and CABG has been associated with increased risks of complications and mortality compared to undergoing either procedure individually.
In the last decade, transcatheter aortic valve replacement (TAVR) has become an increasingly common alternative to SAVR for patients suffering from AS. This is particularly true considering that TAVR is increasingly being performed in younger and low-risk patients. However, there is limited evidence supporting percutaneous coronary intervention (PCI) alongside TAVR, resulting in inconsistent clinical practice across centers. Given the risk of complications associated with SAVR+CABG, a fully percutaneous strategy for managing both AS and CAD represents an attractive alternative. Although PCI may be considered in patients undergoing TAVR who have greater than 70% stenosis in proximal coronary segments, the comparative effectiveness of TAVR+PCI and SAVR+CABG remains uncertain, and whether 1 strategy should be preferred over the other in patient’s candidate to both is still unknown. The existing body of evidence remains inconclusive, with mixed results reported to date, and no clear consensus exists on the superior strategy for treating patients with AS and concurrent CAD.
We conducted a large-scale meta-analysis of randomized controlled trials (RCT) and propensity score matched (PSM) studies comparing clinical outcomes after TAVR+PCI versus SAVR+CABG in patients with severe AS and CAD to determine whether a percutaneous approach has a differential impact on all-cause mortality as compared to surgery at the longest follow-up.
Methods
This systematic review was registered with PROSPERO (ID: CRD42025642206). As the analysis was based solely on data from previously published research and did not involve direct patient interaction or identifiable personal data, institutional review board approval and informed consent were not necessary. Data compiled or analyzed throughout the study can be made available by the corresponding author, pending reasonable request.
Search strategy and eligibility criteria
A comprehensive literature search was carried out in alignment with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) and PRIMA of individual patient data (PRISMA-IPD) guidelines. , The PRISMA flowchart detailing the study selection process can be found in Figure 1 . PubMed, ScienceDirect, DOAJ, SciELO, and Cochrane Library databases were queried from inception through December 2024 to identify comparative studies examining TAVR+PCI versus SAVR+CABG. The review was guided by the PICOS framework: (1) Population: Adult patients diagnosed with AS and CAD requiring procedural treatment; (2) Intervention: fully transcatheter aortic valve replacement and coronary revascularization, that is, TAVR+PCI; (3) Comparator: fully surgical aortic valve replacement and coronary revascularization, that is, SAVR+CABG; (4) Outcomes: Comparisons of clinical outcomes between the 2 approaches; (5) Study Design: Randomized controlled trials, or propensity score matched studies only.
PRISMA flowchart of the included studies. A PRISMA flow diagram offers a graphical representation of the sequential steps involved in conducting a systematic review or meta-analysis. It outlines the stages of identification, screening, eligibility assessment, and inclusion of studies, aiding both researchers and readers in comprehending the selection process for the review, along with the rationale behind excluding specific studies.
Bibliographies of included articles and relevant reviews were also manually screened (reverse snowballing) to identify additional eligible studies. The detailed search strategy is included in Supplementary Table 1 . Two reviewers (M.B., P.P.L.) independently screened all titles, abstracts, and full texts. Any disagreements regarding study eligibility were resolved by discussion with senior reviewers (A.L., B.R.) until consensus was reached.
We included only English-language studies. Studies were excluded if they were non propensity score matched observational studies, abstracts, editorials, conference proceedings, commentaries, or review articles. When more than 1 publication originated from the same research group or institution, the timeframe of data collection was reviewed to identify overlapping populations. In such cases, the study with the largest cohort was selected.
Study quality was appraised using the ROBINS-I tool for nonrandomized studies and the RoB 2 tool for randomized trials. To evaluate the strength of evidence, the GRADE (Grading of Recommendations Assessment, Development and Evaluation) framework was applied.
Data collection and quality assessment
Data were extracted using Microsoft Excel (Office 365; Microsoft Corp., Redmond, WA). Categorical outcomes were recorded as absolute counts and percentages, while continuous variables were summarized as presented (mean or median, and standard deviation, range, or interquartile range). Approximation of the means from sample sizes, medians and other statistics followed the method described by Luo et al. Key study characteristics including institution, study design, study duration, and sample size were documented. Notably, heterogeneity existed in reported variables across studies, requiring inferential steps to standardize denominators and extract patient-level data.
Individual patient data (IPD) were reconstructed from published Kaplan–Meier survival plots using the methodology by Liu et al. This involved digitizing the curves with WebPlotDigitizer ( https://apps.automeris.io/wpd/ ) to obtain time-to-event coordinates for each treatment group. These digitized plots were then cross-checked with the originals for consistency. Differences were evaluated and reduced by applying root mean square error (RMSE ≤0.05), mean absolute error (MAE ≤0.02), and maximum absolute error (MaxAE ≤0.05) as performance benchmarks. These statistical measures demonstrated the high accuracy of the reconstructed data. The extracted time-event data were further processed using the R package IPDfromKM, incorporating numbers at risk and total participants to reconstruct IPD across studies. The compiled data formed the basis of the final analytic dataset.
Statistical analysis
Continuous baseline patient characteristics reported in varying formats were standardized to mean and standard deviation using the method for unknown nonnormal distributions (MLN) proposed by Cai et al. Final estimates were weighted by sample size and compared using the standardized mean difference (SMD). Categorical variables were compared using the risk difference (RD).
The primary analysis focused on overall mortality as determined by the reconstructed individual patient data. For group comparison, the risk ratio (RR) or SMD with 95% confidence interval (CI) were computed for categorical or continuous data, respectively. While for single group outcomes the pool estimated rate (PER) or pooled estimated mean (PEM) were calculated. A random-effects model using the Der Simonian–Laird method was employed to account for interstudy variability, and inverse variance weighting was applied. In studies reporting zero events, a continuity correction was used. Statistical heterogeneity was assessed using the Cochran Q statistic and I² index. Funnel plot symmetry was visually inspected to assess publication bias; Egger’s test was not conducted due to the small number of eligible studies (<10).
Survival analysis was conducted using Kaplan–Meier curves, with group comparisons made via the log-rank test. Cox proportional hazards models were used to derive the hazard ratio (HR) with 95% CI, and the proportional hazards assumption was evaluated using the Grambsch–Therneau test and Schoenfeld residual plots. When this assumption was violated, we employed flexible parametric survival models (Royston-Parmar models) using B-splines to model hazard functions without assuming proportionality. The restricted mean survival time (RMST), which reflects the expected survival duration within a fixed time horizon, was calculated, and the difference (DRMST) evaluated across study arms. The SAVR+CABG group served as the reference category.
A 2-sided p-value <0.05 was considered statistically significant. All statistical analyses were executed in R version 4.4.2 using RStudio.
Results
An outline of the systematic review process is shown in Figure 1 . The literature search identified 1,974 potentially eligible studies. After removal of duplicates, 1,911 studies were screened. Among these, 14 full text articles were assessed for eligibility. Six articles ,,,,, met our inclusion criteria with a total of 1,998 patients: 1,007 in the TAVR+PCI group and 991 in the SAVR+CABG group. All studies were published between 2018 and 2024. Details of the individual studies are shown in Table 1 . The studies included 2 RCTs, and 4 PSM studies. The critical appraisal of nonrandomized and randomized included studies is shown in Supplementary Tables 2 and 3 , respectively. The baseline patient characteristics are shown in Table 2 and show well-balanced parameters between the 2 groups with a SMD or RD <10%.
Table 1
Included studies
| Paper | Study period | Study type | Country | Patients | TAVR+PCI | SAVR+CABG |
|---|---|---|---|---|---|---|
| Alperi et al. | 2007-2019 | PSM | International | 312 | 156 | 156 |
| Amat-Santos et al. | 2018-2021 | PSM | Spain | 508 | 254 | 254 |
| Barbanti et al. | December 2010-June 2012 | PSM | Italy | 472 | 236 | 236 |
| Elderia et al. | Since 2017 | PSM | Germany | 202 | 101 | 101 |
| Kedhi et al. | May 31, 2018-June 30, 2023 | RCT | International | 172 | 91 | 81 |
| Søndergaard et al. | June 2012-June 2016 | RCT subgroup | International | 332 | 169 | 163 |
CABG = coronary artery bypass grafting; PCI = percutaneous coronary intervention; PSM = propensity-score matching; RCT = randomized controlled study; SAVR = surgical aortic valve replacement; TAVR= transcatheter aortic valve replacement.
Table 2
Baseline patient’s characteristics
| Characteristic | TAVR+PCI | SAVR+CABG | SMD/RD |
|---|---|---|---|
| Age, yrs | 78.9 ± 1.4 | 78.5 ± 1.4 | −0.0695 |
| Female sex | 48.2% (404/838) | 48.6% (403/828) | 0.0022 |
| BMI, kg/m 2 | 27.1 ± 0.9 | 27.1 ± 0.8 | −0.0015 |
| COPD | 18.3% (135/737) | 17.5% (127/724) | −0.0139 |
| Diabetes | 39.0% (327/838) | 38.9% (321/825) | −0.0007 |
| Hypertension | 85.1% (435/511) | 83.6% (427/511) | −0.0144 |
| NYHA III/IV | 52.6% (441/838) | 52.5% (432/823) | 0.0143 |
| Previous Valve | 5.9% (30/511) | 1.2% (6/511) | −0.0477 |
| History of AF | 27.9% (168/602) | 24.8% (146/589) | −0.0308 |
| PVD | 26.0% (218/838) | 22.3% (184/825) | −0.0494 |
| GFR <60 mL/min | 30.9% (138/446) | 19.2% (83/433) | −0.0826 |
| STS-PROM, % | 4.4 ± 1.1 | 4.0 ± 1.2 | −0.0530 |
| EuroSCORE II, % | 5.5 ± 2.2 | 5.4 ± 2.0 | −0.0082 |
| LVEF, % | 55.3 ± 1.9 | 54.1 ± 1.4 | 0.0918 |
| Previous MI | 30.2% (146/483) | 30.2% (142/470) | −0.0076 |
| Left main disease | 24.3% (125/515) | 28.4% (141/496) | 0.0278 |
| CAD | |||
| 1 vessel | 27.2% (113/416) | 21.7% (86/396) | −0.0495 |
| 2 vessels | 41.8% (174/416) | 39.9% (158/396) | −0.0169 |
| 3 vessels | 27.6% (115/416) | 35.6% (141/396) | 0.0671 |
| SYNTAX | 20.1 ± 8.1 | 20.8 ± 8.3 | 0.0700 |
AF = atrial fibrillation; BMI = body mass index; CABG = coronary artery bypass graft; CAD = coronary artery disease; COPD = chronic obstructive pulmonary disease; LVEF = left ventricular ejection fraction; MI = myocardial infarction; NYHA = New York Heart Association; PVD = peripheral vascular disease; RD = risk difference; SMD = standardized mean difference; STS-PROM = Society of Thoracic Surgeons predicted risk of mortality.
Meta-analysis
Intraoperative outcome including type of valve, timing between TAVR and PCI, and number of vessels and lesions treated are summarized in Table 3 , as well as the postoperative complications. TAVR+PCI were associated with lower rates of reintervention (RR: 0.29, 95%CI [0.10; 0.82], p = 0.020), postoperative atrial fibrillation (POAF, RR: 0.23, 95%CI [0.17; 0.32], p <0.001), and acute kidney injury (AKI, RR: 0.39, 95%CI [0.22; 0.69], p = 0.001) compared to SAVR+CABG. On the other hand, TAVR+PCI reported higher rates of postoperative moderate/severe aortic regurgitation (RR: 18.1, 95%CI [5.08; 64.6], p <0.001), permanent pacemaker implantation (PPI, RR: 2.92, 95%CI [1.37; 6.20], p = 0.005), and major vascular complications (RR: 6.57, 95%CI [2.95; 14.6], p <0.001) compared to SAVR+CABG. Major bleeding did not differ between groups (RR: 0.62, 95%CI [0.58; 4.50], p = 0.358). There were no other significant postoperative differences, as shown in Table 3 and Supplementary Figures 1-20 .
Table 3
Perioperative outcomes
| Outcome | Studies | Patients | Effect (95%CI), p-value | TAVR+PCI (95%CI) | SAVR+CABG (95%CI) | Heterogeneity (I 2, p-value) |
|---|---|---|---|---|---|---|
| Valve | ||||||
| BEV | 4 | 680 | – | 45.8% (25.1; 68.1) | – | 94.4%, p <0.0001 |
| SEV | 4 | 680 | – | 54.2% (31.9; 74.9) | – | 94.4%, p <0.0001 |
| Surgical bioprosthetic | 4 | 592 | – | – | 97.2% (95.4; 98.3) | 0.0%, p = 0.4520 |
| Surgical mechanic | 4 | 592 | – | – | 2.8% (1.7; 4.7) | 0.0%, p = 0.4520 |
| Mean vessels treated | 4 | 1,017 | −0.59 (−1.34; 0.16), p = 0.122 | 1.6 (1.4; 1.8) | 2.0 (1.4; 2.7) | 97.0%, p <0.0001 |
| Mean lesions treated | 3 | 416 | – | 1.6 (1.3; 1.8) | – | 91.9%, p <0.0001 |
| Complete revascularization | 3 | 796 | 0.83 (0.47; 1.41), p = 0.511 | 60.4% (39.1; 57.8) | 78.4 % (45.3; 94.1) | 96.0%, p <0.0001 |
| PCI before TAVR | 4 | 662 | – | 77.3% (34.4; 95.7) | – | 98.0%, p <0.0001 |
| PCI during TAVR | 4 | 662 | – | 18.4% (3.5; 58.2) | – | 97.8%, p <0.0001 |
| PCI after TAVR | 5 | 753 | – | 0% | – | – |
| Valve embolization | 3 | 446 | – | 2.9% (1.3; 6.1) | – | 20.9%, p <0.0001 |
| Reintervention | 3 | 878 | 0.29 (0.10; 0.82), p = 0.020 | 1.1% (0.5; 2.8) | 4.2% (1.6; 10.6) | 0.0%, p = 0.6183 |
| Pacemaker | 5 | 1,681 | 2.92 (1.37; 6.20), p = 0.005 | 13.6% (9.3; 19.4) | 4.8% (2.6; 8.7) | 73.7%, p = 0.0043 |
| Postoperative AF | 3 | 1,007 | 0.23 (0.17; 0.32), p <0.001 | 6.6% (3.1; 13.5) | 28.3% (16.0; 44.9) | 0.0%, p = 0.6370 |
| Acute kidney injury | 4 | 1,513 | 0.39 (0.22; 0.69), p = 0.001 | 8.5% (2.8; 23.1) | 22.4% (13.9; 34.0) | 75.7%, p = 0.0063 |
| Major vascular complication | 5 | 1,681 | 6.57 (2.95; 14.6), p <0.001 | 6.0% (3.4; 10.2) | 0.9% (0.4; 1.8) | 0.0%, p = 0.6058 |
| Major bleeding | 4 | 1,479 | 0.62 (0.58; 4.50), p = 0.358 | 5.6% (3.1; 9.8) | 9.1% (5.2; 15.5) | 81.6%, p = 0.0010 |
| Cerebrovascular accident | 4 | 1,209 | 0.56 (0.25; 1.23), p = 0.146 | 2.5% (1.5; 4.2) | 4.5% (3.0; 6.5) | 18.3%, p = 0.2991 |
| Myocardial infarction | 3 | 1,011 | 0.37 (0.10; 1.46), p = 0.156 | 0.6% (0.2; 2.0) | 1.6% (0.8; 3.2) | 0.0%, p = 0.7673 |
| Postop mod/sev AR | 3 | 986 | 18.1 (5.08; 64.6), p <0.001 | 9.0% (4.7; 16.3) | 0.5% (0.2; 1.8) | 0.0%, p = 0.7183 |
| 30-day mortality | 4 | 1,209 | 0.88 (0.36; 2.15), p = 0.786 | 2.1% (0.9; 4.6) | 2.6% (1.2; 5.7) | 11.1%, p = 0.3375 |
Bold font indicates statistical significance.
AF = atrial fibrillation; AR = aortic regurgitation; BEV = balloon-expandable valve; CABG = coronary artery bypass grafting; CI = confidence interval; PCI = percutaneous coronary intervention; SAVR = surgical aortic valve replacement; SEV = self-expandable valve; TAVR= transcatheter aortic valve replacement.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
