Abstract
Background
The optimal revascularisation strategy for patients undergoing percutaneous coronary intervention (PCI) after transcatheter aortic valve implantation (TAVI) remains undefined. In particular, the prognostic impact of complete revascularisation (CR) versus incomplete revascularisation (ICR) in this setting is unknown.
Aims
To evaluate the long-term clinical outcomes of CR versus ICR in patients undergoing PCI after TAVI.
Methods
We analysed 447 patients from the multicentre, international Revascularization After Transcatheter Aortic Valve Implantation (REVIVAL) registry who underwent PCI after TAVI between 2008 and 2023. Patients were classified as CR or ICR according to the presence of residual significant stenosis following PCI. The primary endpoint was the 4-year incidence of major adverse cardiovascular events (MACE), a composite of cardiovascular death, myocardial infarction or stroke. The Kaplan-Meier method was used to estimate cumulative event rates, and a weighted Cox regression model employing an entropy balance approach was used to adjust for possible confounders.
Results
ICR was achieved in 132 patients (29.5%) and CR in 315 patients (70.5%). The mean follow-up after PCI was 1065±904 days (1213±938 in the ICR group; 1002±883 in the CR group). The crude 4-year incidence of MACE was 35.0% with ICR and 34.7% with CR (HR 0.93, 95% CI 0.62 to 1.39, p=0.710). Adjusted analysis confirmed no significant difference (32.5% vs 34.1%; HR 1.01, 95% CI 0.62 to 1.64, p=0.971).
Conclusion
In this registry, CR after TAVI was common but not associated with improved outcomes compared with ICR. These results support a selective rather than systematic pursuit of CR in elderly, high-risk post-TAVI patients.
Keywords: Transcatheter Aortic Valve Replacement, Percutaneous Coronary Intervention, CORONARY ARTERY DISEASE, Aortic Valve Stenosis, Cardiac Catheterization
WHAT IS ALREADY KNOWN ON THIS TOPIC
Patients who undergo percutaneous coronary intervention (PCI) after previous transcatheter aortic valve implantation (TAVI) are at an increased risk for adverse cardiovascular events.
Complete revascularisation is generally the preferred strategy in patients undergoing PCI for CAD, but the optimal strategy in TAVI patients is unclear.
WHAT THIS STUDY ADDS
Complete revascularisation was commonly achieved in patients who previously underwent TAVI. However, the achievement of complete revascularisation was not associated with different clinical outcomes compared with incomplete revascularisation.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The results of this study underline the need for an individualised approach to coronary revascularisation in TAVI patients, and further research is warranted to inform clinicians on the indications for different revascularisation strategies.
Introduction
Coronary artery disease (CAD) and aortic stenosis (AS) share common pathogenic mechanisms and frequently coexist,1 with CAD present in up to 75% of patients with AS and associated with worse clinical outcomes after aortic valve interventions.2 3
Transcatheter aortic valve implantation (TAVI) has become the standard of care for symptomatic severe AS across a broad spectrum of surgical risk.4–7 As TAVI is increasingly performed in younger patients with longer life expectancy, the incidence of unplanned percutaneous coronary intervention (PCI) after TAVI is rising, underscoring the importance of defining optimal revascularisation strategies in this expanding population.8
In the general CAD population, complete revascularisation (CR) is often associated with improved outcomes compared with incomplete revascularisation (ICR).9 In patients undergoing TAVI, most available evidence concerns revascularisation before the TAVI procedure. The NOTION-3 (Routine Revascularization With Percutaneous Coronary Intervention in Patients With Coronary Artery Disease Undergoing Transcatheter Aortic Valve Implantation - the Nordic Aortic Valve Intervention-3) trial demonstrated improved outcomes in patients randomised to PCI versus no PCI, with CR achieved in 89% of treated patients.10 In contrast, the randomised, non-inferiority ACTIVATION (Percutaneous Coronary Intervention Prior to Transcatheter Aortic Valve Implantation) trial showed that rates of death and rehospitalisation at 1 year were similar between PCI and no PCI prior to TAVI.11 Also, observational studies have produced conflicting evidence.12–14
As a result, no consensus has been reached on the optimal revascularisation strategy in patients undergoing TAVI, including timing and completeness of revascularisation.3 In particular, the post-TAVI setting remains unexplored, and it is unknown whether achieving complete revascularisation when performing PCI after TAVI would offer a clinical benefit. This is a clinically relevant gap, as achieving CR in this setting may be hindered by procedural challenges, including impaired coronary access due to the transcatheter valve frame, complex CAD anatomy and significant comorbidities.9
This study addresses this unmet need by evaluating the prognostic impact of complete versus incomplete revascularisation in patients with previous TAVI undergoing PCI.
Methods
Study population
This study uses data collected in the Revascularization After Transcatheter Aortic Valve Implantation (REVIVAL; NCT03283501) registry, a multicentre, international, observational registry using pooled individual data of consecutive patients who underwent PCI after TAVI in 21 centres across Europe from 2008 to 2023 (online supplemental table 1). Completeness of revascularisation was defined anatomically on the final angiogram as the absence of any residual stenosis judged significant by the operator (generally ≥70% in coronary vessels or segments ≥2.25 mm in diameter). ICR was defined by the persistence of at least one such lesion. Functional (physiology-guided) lesion assessment was not systematically performed, and revascularisation decisions were left to operator judgement without predefined criteria for deferring treatment of individual lesions.
The study was conducted according to the ethical principles of the Declaration of Helsinki, and each centre’s institutional review board approved the corresponding study protocols. Informed written consent was obtained from all participants before enrolment in the study.
Standard techniques were used for TAVI and PCI in all centres, and the choice of the device was left to the operator’s judgement.
All participant data, including baseline characteristics, data from the TAVI and PCI procedures, antithrombotic drug therapy and clinical outcomes, were collected by each centre in a prespecified data extraction sheet. The data were then anonymised and merged into a general core dataset. Patients were followed up using inpatient or outpatient visits or telephone interviews.
Endpoints
The primary endpoint of this study was the occurrence of a major adverse cardiovascular event (MACE), which was defined as the composite of cardiovascular (CV) death, myocardial infarction (MI) or stroke at 4-year follow-up. Secondary outcome parameters included the following: the individual components of the composite primary outcome, all-cause death, target vessel MI, periprocedural MI, target vessel revascularisation (TVR), target lesion revascularisation, target vessel failure (defined as a composite of TVR, MI and CV death), definite or probable stent thrombosis, bleeding (defined as Bleeding Academic Research Consortium type 2–5) and CV hospitalisation. All endpoints were assessed after a follow-up period of 4 years. Endpoint definitions can be found in online supplemental table 2.
Statistical analysis
Continuous variables are presented as mean ± SD or as median with IQR, whereas categorical variables are reported as absolute counts and corresponding percentages. Covariate balance between patients who underwent PCI with CR and those with ICR was evaluated using the standardised mean difference (SMD), with values below 0.1 considered indicative of good balance.
We applied entropy balancing to control for confounding and achieve covariate balance between the two groups at the time of PCI. This procedure reweights observations so that the covariate distributions in the two groups become as similar as possible. Weights were then truncated at the first and 99th percentiles to improve estimate stability and limit the influence of extreme weights.15 In our analysis, this trimming procedure did not introduce bias, as SMDs after truncation confirmed that covariates remained well balanced across groups.
We computed a weighted model adjusting for potential confounders, defined as pre-exposure variables associated with both the exposure and the outcome but not lying on the causal pathway (online supplemental figure 1). Online supplemental table 3 shows a complete list of all covariates included in the adjusted model. Online supplemental figure 2 provides the pre-weighting and post-weighting balance diagnostics for the model.
To account for potential correlations introduced by the weighting process, we used a robust variance estimator to estimate 95% CIs.
Cox regression was then performed to estimate the effect of different revascularisation strategies on clinical outcomes, and the Kaplan-Meier method was used for the estimation of the crude and weighted cumulative incidences. These incidences were compared between the two groups by using the log-rank test for the time to first event.
We performed several sensitivity analyses to assess the statistical robustness of our findings. To test for the potential overestimation of risks by the Kaplan-Meier method, we applied an adjusted Fine-Gray hazard model as an alternative to the primary Cox model to estimate the subdistribution HRs while accounting for competing risks. The competing risks were defined as follows: non-cardiovascular death was treated as the competing risk for the composite endpoint of MACE and the individual endpoint of cardiovascular death, and all-cause death was treated as the competing risk for all other non-fatal endpoints. This way, we accounted for the nature of each of the different endpoints.
Because more than half of the cohort was enrolled at a single centre, we conducted a dedicated sensitivity analysis to assess the robustness of our findings to between-centre heterogeneity (online supplemental table 4). To reduce the statistical instability associated with very small clusters, we excluded centres contributing fewer than 10 patients. In the resulting restricted cohort, we re-estimated the weighted Cox regression using the same covariates as in the adjusted model, while additionally including country as a balancing variable in the entropy weighting procedure. Country, rather than individual centre, was used because most centres contributed few patients and balancing on centre-level indicators would have produced unstable weights; country captures the main expected sources of between-centre heterogeneity (healthcare system, referral patterns and local practice). Additionally, subgroup analyses with formal tests for interaction were conducted for the primary endpoint on the basis of clinical presentation (acute vs chronic coronary syndrome), sex, median age and timing of PCI after TAVI (early vs late groups using a 3-month cut-off). Lastly, to address the difference in follow-up duration between groups, we performed a Kaplan-Meier analysis of the various outcomes with follow-up truncated at 2 years.
The p values <0.05 were considered statistically significant. All analyses were performed using R, V.4.3.1 (R Foundation for Statistical Computing, Vienna, Austria).
Results
Baseline clinical and procedural characteristics
Out of 464 patients included in the REVIVAL registry, 447 patients were included in the analysis, and 17 were excluded due to missing data. ICR was achieved in 132 patients (29.5%) and CR in 315 patients (70.5%). The mean time from TAVI to PCI was 347±489 days in the ICR group and 386±568 days in the CR group.
Tables 1 and 2 present the baseline demographic characteristics as well as clinical and procedural characteristics in the ICR and CR cohorts before and after weighting for clinical confounders at the time of the PCI.
Table 1. Baseline clinical characteristics of the population with no adjustment and after adjustment.
| Unadjusted | Adjusted | |||||
|---|---|---|---|---|---|---|
| ICR | CR | SMD | ICR | CR | SMD | |
| Number of patients | 132 | 315 | 117.4 | 266.0 | ||
| Clinical characteristics | ||||||
| Age, mean (SD) | 81.4 (6.2) | 80.8 (6.2) | 0.101 | 81.0 (6.0) | 80.9 (6.1) | 0.013 |
| Females, n (%) | 42 (31.8) | 123 (39.0) | 0.152 | 44.5 (37.9) | 101.8 (38.3) | 0.008 |
| Logistic Euroscore, mean (SD) | 15.5 (12.9) | 13.2 (11.3) | 0.187 | 14.2 (11.5) | 13.4 (11.5) | 0.068 |
| Euroscore 2, mean (SD) | 7.0 (7.5) | 5.0 (5.6) | 0.299 | 5.9 (5.7) | 5.3 (6.3) | 0.107 |
| Body mass index, mean (SD) | 26.9 (4.3) | 27.3 (5.1) | 0.078 | 27.2 (4.6) | 27.2 (4.9) | 0.002 |
| Hypertension, n (%) | 119 (90.2) | 282 (89.5) | 0.021 | 105.6 (89.9) | 238.3 (89.6) | 0.011 |
| Diabetes, n (%) | 48 (36.4) | 91 (28.9) | 0.160 | 35.9 (30.6) | 81.0 (30.5) | 0.003 |
| Dyslipidaemia, n (%) | 99 (75.0) | 224 (71.1) | 0.088 | 86.1 (73.3) | 195.3 (73.4) | 0.002 |
| Family history of coronary artery disease, n (%) | 12 (17.6) | 19 (13.3) | 0.121 | 7.4 (12.9) | 14.5 (13.2) | 0.008 |
| Previous PCI, n (%) | 58 (43.9) | 133 (42.2) | 0.035 | 48.1 (41.0) | 110.1 (41.4) | 0.008 |
| Previous CABG, n (%) | 31 (23.5) | 32 (10.2) | 0.362 | 17.1 (14.5) | 38.1 (14.3) | 0.006 |
| Previous PCI or CABG, n (%) | 71 (53.8) | 142 (45.1) | 0.175 | 55.3 (47.1) | 122.7 (46.1) | 0.020 |
| Peripheral arterial disease, n (%) | 20 (15.3) | 62 (19.7) | 0.116 | 21.5 (18.3) | 48.5 (18.2) | 0.002 |
| Estimated glomerular filtration rate, mean (SD) | 60.6 (23.5) | 62.6 (21.1) | 0.090 | 62.6 (23.1) | 62.8 (21.3) | 0.007 |
| Left ventricular ejection fraction, mean (SD) | 54.0 (11.7) | 54.9 (12.0) | 0.083 | 55.1 (12.2) | 55.0 (11.9) | 0.004 |
| TAVI characteristics | ||||||
| Year of TAVI procedure, n (%) | 0.173 | 0.006 | ||||
| 2008–2012 | 24 (18.2) | 40 (12.7) | 14.5 (12.3) | 32.6 (12.2) | ||
| 2012–2017 | 69 (52.3) | 163 (51.9) | 62.6 (53.3) | 141.3 (53.1) | ||
| 2018–2023 | 39 (29.5) | 111 (35.4) | 40.4 (34.4) | 92.1 (34.6) | ||
| Transapical approach, n (%) | 10 (7.6) | 26 (8.3) | 0.025 | 8.6 (7.3) | 19.4 (7.3) | 0.001 |
| Valve size, mean (SD) | 26.6 (2.2) | 26.4 (2.4) | 0.087 | 26.37 (2.2) | 26.36 (2.3) | 0.004 |
| Valve-in-valve, n (%) | 2 (1.5) | 3 (1.0) | 0.051 | 1.5 (1.3) | 3.5 (1.3) | 0.001 |
| Postdilatation, n (%) | 38 (28.8) | 86 (27.3) | 0.033 | 35.9 (30.6) | 81.0 (30.5) | 0.003 |
Variables included in the adjusted model are reported in online supplemental table 3.
CABG, coronary artery bypass graft; CR, complete revascularisation; ICR, incomplete revascularisation; PCI, percutaneous coronary intervention; SMD, standardised mean difference; TAVI, transcatheter aortic valve implantation.
Table 2. PCI characteristics of the population with no adjustment and after adjustment.
| Unadjusted | Adjusted | |||||
|---|---|---|---|---|---|---|
| ICR | CR | SMD | ICR | CR | SMD | |
| Number of patients | 132 | 315 | 117.4 | 266.0 | ||
| PCI characteristics | ||||||
| Days from TAVI to PCI, mean (SD) | 346.9 (488.7) | 385.9 (568.5) | 0.074 | 357.4 (481.7) | 356.8 (536.0) | 0.001 |
| PCI planned at time of TAVI, n (%) | 44 (33.3) | 65 (20.6) | 0.289 | 32.1 (27.4) | 72.0 (27.1) | 0.006 |
| ACS as index presentation, n (%) | 48 (36.4) | 102 (32.9) | 0.171 | 38.4 (32.7) | 86.6 (32.6) | 0.004 |
| Radial access, n (%) | 50 (39.4) | 153 (50.2) | 0.218 | 45.2 (39.1) | 125.5 (48.7) | 0.194 |
| Number of diseased vessels, mean (SD) | 1.3 (0.5) | 1.2 (0.5) | 0.069 | 1.3 (0.5) | 1.3 (0.5) | 0.083 |
| ACC/AHA 2B or C, n (%) | 79 (81.4) | 233 (89.6) | 0.234 | 74.4 (84.2) | 197.7 (89.5) | 0.159 |
| Severe calcifications, n (%) | 27 (20.5) | 86 (27.3) | 0.161 | 26.0 (22.2) | 66.6 (25.0) | 0.067 |
| Chronic total occlusion, n (%) | 4 (3.0) | 7 (2.2) | 0.051 | 4.4 (3.7) | 5.5 (2.1) | 0.100 |
| Evidence of thrombus, n (%) | 8 (6.6) | 9 (3.0) | 0.167 | 7.5 (6.8) | 9.3 (3.7) | 0.142 |
| Intravascular imaging, n (%) | 9 (6.8) | 23 (7.3) | 0.019 | 7.3 (6.2) | 20.8 (7.8) | 0.064 |
| Bifurcation, n (%) | 37 (28.0) | 77 (24.4) | 0.082 | 30.5 (26.0) | 68.3 (25.7) | 0.008 |
| In-stent restenosis, n (%) | 16 (12.1) | 40 (12.7) | 0.018 | 16.4 (14.0) | 34.7 (13.0) | 0.027 |
| Thrombus aspiration, n (%) | 2 (1.6) | 0 (0.0) | 0.181 | 1.4 (1.3) | 0.0 (0.0) | 0.159 |
| Use of plaque modification device, n (%) | 4 (3.4) | 15 (5.1) | 0.087 | 5.1 (4.7) | 13.9 (5.6) | 0.040 |
| PCI on left Main (coronary artery), n (%) | 21 (16.7) | 48 (15.8) | 0.024 | 17.9 (15.7) | 42.9 (16.6) | 0.026 |
| PCI on left anterior descending (coronary artery), n (%) | 51 (40.5) | 146 (48.0) | 0.152 | 49.9 (43.8) | 124.9 (48.5) | 0.093 |
| PCI with drug-eluting stent, n (%) | 112 (90.3) | 279 (93.3) | 0.109 | 101.7 (91.4) | 238.3 (93.0) | 0.059 |
| PCI with drug-eluting balloon, n (%) | 12 (9.7) | 24 (8.0) | 0.058 | 11.8 (10.6) | 20.7 (8.1) | 0.087 |
| Number of lesions per patient, mean (SD) | 1.39 (0.8) | 1.4 (0.7) | 0.042 | 1.4 (0.8) | 1.4 (0.7) | 0.060 |
| Number of stents per lesion, mean (SD) | 1.67 (1.21) | 1.62 (1.02) | 0.043 | 1.7 (1.2) | 1.7 (1.1) | 0.061 |
| Total stent length per lesion, mean (SD) | 37.29 (28.6) | 38.1 (25.2) | 0.028 | 39.2 (28.1) | 39.0 (26.6) | 0.008 |
| Minimum stent diameter, mean (SD) | 2.96 (0.62) | 2.9 (0.6) | 0.037 | 2.95 (0.61) | 3.0 (0.6) | 0.002 |
| Total number of stents per patient, mean (SD) | 1.7 (1.3) | 1.53 (1.01) | 0.144 | 1.8 (1.2) | 1.59 (1.00) | 0.139 |
| Total stent length per patient, mean (SD) | 36.6 (29.9) | 31.1 (26.3) | 0.195 | 39.0 (30.0) | 33.1 (27.5) | 0.206 |
| Complex PCI, n (%) | 50 (40.7) | 88 (30.1) | 0.221 | 41.5 (37.4) | 79.3 (31.4) | 0.126 |
| Successful PCI, n (%) | 118 (89.4) | 314 (99.7) | 0.465 | 105.9 (90.2) | 266.0 (100.0) | 0.465 |
| Angiographic success, n (%) | 121 (100.0) | 298 (98.0) | 0.201 | 110.5 (100.0) | 255.4 (99.1) | 0.136 |
| Coronary occlusion n (%) | 0 (0.0) | 4 (1.3) | 0.160 | 0.0 (0.0) | 2.6 (1.0) | 0.141 |
| Post-PCI antiplatelet therapy | ||||||
| Oral anticoagulation, n (%) | 31 (24.0) | 78 (25.4) | 0.032 | 29.8 (25.4) | 67.2 (25.3) | 0.003 |
| Short dual antiplatelet therapy, n (%) | 30 (26.3) | 74 (25.0) | 0.030 | 27.2 (25.9) | 66.6 (26.3) | 0.010 |
Variables included in the adjusted model are reported in online supplemental table 3.
ACS was defined as unstable angina, non-ST-elevation myocardial infarction or ST-elevation myocardial infarction. Complex PCI was defined as at least one among ≥3 vessels treated, ≥3 lesions treated, or bifurcation lesions treated.
ACS, acute coronary syndrome; CR, complete revascularisation; ICR, incomplete revascularisation; PCI, percutaneous coronary intervention; SMD, standardised mean difference; TAVI, transcatheter aortic valve implantation.
Before weighting, patients who received ICR were slightly older (81.4±6.2 vs 80.8±6.2 years, SMD=0.101), more likely to be men (68.2% vs 61.0%, SMD=0.152), to have a family history of CAD (17.6% vs 13.3%, SMD=0.121) and to be diabetic (36.4% vs 28.9%, SMD=0.160). Patients with ICR furthermore had more previous revascularisations (53.8% vs 45.1%, SMD=0.175). These patients also had higher rates of procedural complexity (40.7% vs 30.1%, SMD=0.221), and PCI was less likely to be deemed successful (89.4% vs 99.7%, SMD=0.465). The two groups were similar regarding hypertension (90.2% vs 89.5%, SMD=0.021), body mass index (26.9±4.3 vs 27.3±5.1, SMD=0.078) and left ventricular ejection fraction (54.0±11.7 vs 54.9±12.0, SMD=0.083).
An acute coronary syndrome (ACS) as index presentation was more common in the ICR cohort compared with the CR cohort (36.4% vs 32.9%; SMD=0.171). When broken down by clinical subtype, unstable angina occurred in 10.6% versus 7.1%, non-ST-elevation myocardial infarction in 22.0% versus 19.7%, and ST-elevation myocardial infarction in 3.8% versus 6.1% of patients with ICR and CR, respectively.
After adjusting for several demographic, clinical and TAVI-related confounders, balance was achieved for most of the included variables (table 1, online supplemental figure 2), while some differences in procedural characteristics remained (table 2).
Outcomes before weighting
Mean follow-up after PCI was 1065±904 days overall, 1213±938 days in the ICR group and 1002±883 days in the CR group. 282 patients (63.1%) completed a 4-year follow-up.
The 4-year crude incidence of MACE was 35.0% in the ICR group and 34.7% in the CR group (HR 0.93, 95% CI 0.62 to 1.39, p=0.710; table 3 and figure 1A). Similarly, no significant differences were found in all-cause death (34.6% vs 40.9%, HR 1.16, 95% CI 0.78 to 1.71, p=0.459; table 3), CV death (21.3% in the ICR group, 25.5% in the CR group; HR 1.10, 95% CI 0.66 to 1.84, p=0.709; table 3 and figure 1B), MI (13.7% vs 11.4%, HR 0.84, 95% CI 0.43 to 1.65, p=0.609; table 3 and figure 1C), stroke (5.2% vs 8.4%, HR 1.41, 95% CI 0.51 to 3.89, p=0.504; table 3 and figure 1D) or any other of the secondary endpoints (table 3).
Table 3. 4-year clinical outcomes in the unadjusted population and after adjustment.
| Endpoint | Unadjusted | Adjusted for clinical variables | ||||||
|---|---|---|---|---|---|---|---|---|
| KM incidences at 4 years | HR (95% CI) |
P value | KM incidences at 4 years | HR (95% CI) |
P value | |||
| ICR (132 pts) |
CR (315 pts) |
ICR (117.4 pts) |
CR (266.0 pts) |
|||||
| Major adverse cardiovascular events | 35.0 | 34.7 | 0.93 (0.62 to 1.39) |
0.710 | 32.5 | 34.1 | 1.01 (0.62 to 1.64) |
0.971 |
| Death | 34.6 | 40.9 | 1.16 (0.78 to 1.71) |
0.459 | 35.4 | 40.9 | 1.18 (0.75 to 1.86) |
0.466 |
| MI | 13.7 | 11.4 | 0.84 (0.43 to 1.65) |
0.609 | 10.4 | 11.6 | 1.05 (0.42 to 2.60) |
0.919 |
| Stroke | 5.2 | 8.4 | 1.41 (0.51 to 3.89) |
0.504 | 5.8 | 7.3 | 1.20 (0.41 to 3.48) |
0.744 |
| Cardiovascular death | 21.3 | 25.5 | 1.10 (0.66 to 1.84) |
0.709 | 22.1 | 25.2 | 1.09 (0.59 to 1.99) |
0.787 |
| Stent thrombosis | 2.5 | 1.1 | 0.43 (0.09 to 2.11) |
0.295 | 3.8 | 1.2 | 0.31 (0.06 to 1.67) |
0.174 |
| Periprocedural MI | 9.6 | 8.1 | 0.53 (0.21 to 1.30) |
0.164 | 9.7 | 7.1 | 0.43 (0.15 to 1.20) |
0.105 |
| Target vessel MI | 9.6 | 5.4 | 0.55 (0.23 to 1.32) |
0.181 | 8.8 | 5.9 | 0.62 (0.20 to 1.91) |
0.408 |
| Target lesion revascularisation | 13.6 | 8.9 | 0.68 (0.34 to 1.37) |
0.286 | 15.4 | 10.3 | 0.69 (0.31 to 1.54) |
0.367 |
| Target vessel revascularisation | 21.6 | 13.1 | 0.69 (0.39 to 1.24) |
0.215 | 22.2 | 15.2 | 0.80 (0.41 to 1.55) |
0.504 |
| Target vessel failure | 34.9 | 36.4 | 1.00 (0.67 to 1.49) |
0.990 | 33.5 | 38.5 | 1.14 (0.71 to 1.84) |
0.581 |
| Bleeding | 3.2 | 5.7 | 1.26 (0.40 to 3.97) | 0.690 | 4.3 | 4.4 | 1.14 (0.44 to 2.98) |
0.790 |
Variables included in the adjusted model are reported in online supplemental table 3.
CR, complete revascularisation; ICR, incomplete revascularisation; KM, Kaplan-Meier; MI, myocardial infarction; PCI, percutaneous coronary intervention.
Figure 1. Kaplan-Meier estimates of 4-year clinical outcomes in the unadjusted population (A: major adverse cardiovascular event, B: cardiovascular death, C: myocardial infarction and D: stroke). CR, complete revascularisation; ICR, incomplete revascularisation.
Outcomes after weighting
In the adjusted model, the 4-year adjusted incidence of MACE was 32.5% in the ICR group and 34.1% in the CR group (HR 1.01, 95% CI 0.62 to 1.64, p=0.971; table 3 and figure 2A). Similarly, no statistically significant differences were found regarding all-cause death (35.4% vs 40.9%; HR 1.18, 95% CI 0.75 to 1.86, p=0.466; table 3), CV death (22.1% vs 25.2%; HR 1.09, 95% CI 0.59 to 1.99, p=0.787; table 3 and figure 2B), MI (10.4% vs 11.6%; HR 1.05, 95% CI 0.42 to 2.60, p=0.919; table 3 and figure 2C), stroke (5.8% vs 7.3%; HR 1.20, 95% CI 0.41 to 3.48, p=0.744; table 3 and figure 2D) or any other of the secondary endpoints (table 3).
Figure 2. Kaplan-Meier estimates of 4-year clinical outcomes after adjustment (A: major adverse cardiovascular event, B: cardiovascular death, C: myocardial infarction and D: stroke). CR, complete revascularisation; ICR, incomplete revascularisation.
Sensitivity analyses
The statistical robustness of our findings from the adjusted model was confirmed across multiple sensitivity analyses. In the country-level analysis, in which centres and countries with statistical instability were excluded, the adjusted HR for 4-year MACE was consistent with the primary results (online supplemental table 4). A formal competing risk analysis was performed to evaluate a possible overestimation by the Kaplan-Meier method and showed results which were consistent with the main analysis (online supplemental table 5). Furthermore, formal testing for interaction showed a consistent effect of revascularisation strategies on MACE at 4 years across all subgroups of interest (all p-interaction >0.10; online supplemental figure 3). In the sensitivity analysis truncating follow-up at 2 years, the MACE rate was 19.8% in the ICR group and 18.8% in the CR group (HR 0.87, 95% CI 0.49 to 1.56, p=0.64), with no statistically significant differences in the secondary outcomes (online supplemental figure 4).
Discussion
To our knowledge, this is the first dedicated investigation assessing the prognostic impact of complete versus incomplete revascularisation in patients undergoing PCI after TAVI. The main findings can be summarised as follows:
In this elderly, high-risk cohort, CR was achieved in most cases (70.5%) yet was not associated with differences in 4-year MACE or other secondary endpoints.
This result was consistent across crude, adjusted and sensitivity analyses.
ICR was more frequent in the setting of greater procedural complexity and lower procedural success but did not translate into worse clinical outcomes.
In the general PCI population, CR is typically preferred when technically feasible, with more robust evidence supporting its benefit in ACS than in stable CAD.9 16 17 Evidence in the TAVI setting is limited and existing studies almost exclusively address PCI performed before or during TAVI, often as part of preprocedural CAD management.11 18 The clinical context of PCI after TAVI is different, frequently involving unplanned PCI for CAD progression or ACS, and thus prior findings may not be directly applicable.
The REVIVAL registry population was characterised by advanced age (median age of 81 years) and high surgical risk by both Logistic EuroSCORE and EuroSCORE II. In younger surgical cohorts, the prognostic importance of residual CAD is often greater, likely due to longer life expectancy and fewer comorbidities. This may explain why surgical guidelines advocate concomitant coronary artery bypass grafting in patients undergoing surgical aortic valve replacement.19 20 Such recommendations cannot be directly extrapolated to elderly, frail post-TAVI patients.21 The CR rate in this study (~70%) was considerably high, probably due to the low angiographic disease burden in our cohort (mean 1.25±0.49 diseased vessels). This reflects real-world practice, in which TAVI patients with significant CAD often undergo revascularisation before or concomitantly with the valve procedure,21 as supported by the similarly high rate of prior PCI in both groups (43.9% in the ICR group and 42.2% in the CR group, SMD=0.035). This introduces a degree of selection bias, limiting the generalisability of the findings to patients presenting with complex multivessel disease after TAVI.
Interpretation of ICR depends heavily on its definition, which in this study was based on operator angiographic assessment without systematic functional evaluation. ICR can range from leaving untreated high-risk lesions due to technical or anatomical constraints to deferring intervention on minor, non-flow-limiting lesions.22 23 Previous work in TAVI populations has shown that only ‘unreasonable’ ICR, defined by leaving untreated high-risk lesions, is associated with increased mortality.12 24 25 Without functional lesion assessment, our analysis cannot distinguish between these categories, which may dilute any signal for prognostic difference. A procedural success rate of 89.4% in the ICR group further indicates that this cohort is heterogeneous, including both patients in whom revascularisation was intentionally deferred on clinical or anatomical grounds and, less frequently, patients in whom CR was not achieved because of procedural complications. Notably, part of what is anatomically classified as ICR may reflect reasonable ICR, that is, anatomical ICR but functional CR. Rather than undermining our findings, this reinforces that systematically pursuing anatomical CR may not be warranted in this population.
Non-randomised data in stable CAD have not conclusively shown that anatomical CR improves outcomes over ICR, and confounding remains a major limitation.13 Indeed, similarly to our study, ICR is often unintentional, arising from complications or technical failure, both of which could themselves worsen prognosis. Conversely, aggressive pursuit of CR may increase procedural risk without guaranteeing resolution of ischaemia, especially in the presence of small-vessel or microvascular disease.26
In our study, ICR was associated with more complex anatomy, less radial access and lower procedural success, yet these factors did not result in worse outcomes. The absence of a clear crude or adjusted advantage for CR argues against aggressively pursuing CR in this frail cohort and rather supports an individually tailored revascularisation strategy.27 28 A selective ICR strategy may allow shorter procedures, reduced contrast use, lower intraprocedural risk, and potentially shorter durations of intensified antiplatelet therapy, without compromising long-term outcomes.29 30
Limitations
Several important limitations of this study must be acknowledged. First, due to the observational design of this study, the results should be considered hypothesis-generating. Randomised trials would be necessary to account for confounders and provide solid evidence on the preferable revascularisation strategy in TAVI patients. However, conducting such trials currently poses great challenges because the incidence of PCI after TAVI is still relatively low. Additionally, patient selection for PCI may be biased by comorbidities and frailty, potentially influencing clinical outcomes within the groups. Importantly, the differentiation between anatomical and functional ICR was not possible in this study. Residual CAD was assessed angiographically, and completeness was classified according to an operator-based anatomical definition, consistent with the pivotal randomised trials on revascularisation completeness,16 31–33 while functional lesion assessment was not systematically performed. Furthermore, the 15-year study period encompasses substantial advances in PCI techniques, and CR may not have been achievable in earlier enrolments because of technical limitations that would no longer apply with contemporary techniques. Finally, the modest sample size and event rate limit the power to detect more subtle differences in endpoints.
Conclusions
In this multicentre registry of elderly, high-risk patients undergoing PCI after TAVI, complete revascularisation was common but not associated with improved long-term outcomes compared with incomplete revascularisation. These findings support a selective rather than systematic pursuit of complete revascularisation in this population, balancing procedural feasibility and risk against uncertain prognostic benefit. Larger, prospective studies, ideally incorporating functional lesion assessment, are needed to define the optimal revascularisation strategy post-TAVI.
Supplementary material
Acknowledgements
We thank Giulia Frigerio for her valuable help with the statistical analysis presented in this manuscript.
Footnotes
Funding: The publication fee for this work was covered by the Italian Ministry of Health’s ‘Ricerca Corrente’ funding to the IRCCS Humanitas Research Hospital.
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient consent for publication: Consent obtained directly from patient(s)
Ethics approval: This study involves human participants. The study was conducted according to the ethical principles of the Declaration of Helsinki, and each centre’s institutional review board approved the corresponding study protocols. Informed written consent was obtained from all participants before enrolment in the study. Participants gave informed consent to participate in the study before taking part.
Presented at: The results of this study were presented in the form of a poster at the 2026 EAPCI Congress, Munich, Germany.34
Data availability statement
Data are available upon reasonable request.
References
- 1.Abdul-Rahman T, Lizano-Jubert I, Garg N, et al. The common pathobiology between coronary artery disease and calcific aortic stenosis: Evidence and clinical implications. Prog Cardiovasc Dis. 2023;79:89–99. doi: 10.1016/j.pcad.2023.06.002. [DOI] [PubMed] [Google Scholar]
- 2.Stefanini GG, Stortecky S, Cao D, et al. Coronary artery disease severity and aortic stenosis: clinical outcomes according to SYNTAX score in patients undergoing transcatheter aortic valve implantation. Eur Heart J. 2014;35:2530–40. doi: 10.1093/eurheartj/ehu074. [DOI] [PubMed] [Google Scholar]
- 3.Tarantini G, Tang G, Nai Fovino L, et al. Management of coronary artery disease in patients undergoing transcatheter aortic valve implantation. A clinical consensus statement from the European Association of Percutaneous Cardiovascular Interventions in collaboration with the ESC Working Group on Cardiovascular Surgery. EuroIntervention. 2023;19:37–52. doi: 10.4244/EIJ-D-22-00958. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Popma JJ, Deeb GM, Yakubov SJ, et al. Transcatheter Aortic-Valve Replacement with a Self-Expanding Valve in Low-Risk Patients. N Engl J Med. 2019;380:1706–15. doi: 10.1056/NEJMoa1816885. [DOI] [PubMed] [Google Scholar]
- 5.Mack MJ, Leon MB, Thourani VH, et al. Transcatheter Aortic-Valve Replacement in Low-Risk Patients at Five Years. N Engl J Med. 2023;389:1949–60. doi: 10.1056/NEJMoa2307447. [DOI] [PubMed] [Google Scholar]
- 6.Thyregod HGH, Jørgensen TH, Ihlemann N, et al. Transcatheter or surgical aortic valve implantation: 10-year outcomes of the NOTION trial. Eur Heart J. 2024;45:1116–24. doi: 10.1093/eurheartj/ehae043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Jørgensen TH, Thyregod HGH, Savontaus M, et al. Transcatheter aortic valve implantation in low-risk tricuspid or bicuspid aortic stenosis: the NOTION-2 trial. Eur Heart J. 2024;45:3804–14. doi: 10.1093/eurheartj/ehae331. [DOI] [PubMed] [Google Scholar]
- 8.Stefanini GG, Cerrato E, Pivato CA, et al. Unplanned Percutaneous Coronary Revascularization After TAVR: A Multicenter International Registry. JACC Cardiovasc Interv. 2021;14:198–207. doi: 10.1016/j.jcin.2020.10.031. [DOI] [PubMed] [Google Scholar]
- 9.Gaba P, Gersh BJ, Ali ZA, et al. Complete versus incomplete coronary revascularization: definitions, assessment and outcomes. Nat Rev Cardiol. 2021;18:155–68. doi: 10.1038/s41569-020-00457-5. [DOI] [PubMed] [Google Scholar]
- 10.Lønborg J, Jabbari R, Sabbah M, et al. PCI in Patients Undergoing Transcatheter Aortic-Valve Implantation. N Engl J Med. 2024;391:2189–200. doi: 10.1056/NEJMoa2401513. [DOI] [PubMed] [Google Scholar]
- 11.Patterson T, Clayton T, Dodd M, et al. ACTIVATION (PercutAneous Coronary inTervention prIor to transcatheter aortic VAlve implantaTION): A Randomized Clinical Trial. JACC Cardiovasc Interv. 2021;14:1965–74. doi: 10.1016/j.jcin.2021.06.041. [DOI] [PubMed] [Google Scholar]
- 12.Witberg G, Zusman O, Codner P, et al. Impact of Coronary Artery Revascularization Completeness on Outcomes of Patients With Coronary Artery Disease Undergoing Transcatheter Aortic Valve Replacement: A Meta-Analysis of Studies Using the Residual SYNTAX Score (Synergy Between PCI With Taxus and Cardiac Surgery) Circ Cardiovasc Interv. 2018;11:e006000. doi: 10.1161/CIRCINTERVENTIONS.117.006000. [DOI] [PubMed] [Google Scholar]
- 13.Costa G, Pilgrim T, Amat Santos IJ, et al. Management of Myocardial Revascularization in Patients With Stable Coronary Artery Disease Undergoing Transcatheter Aortic Valve Implantation. Circ Cardiovasc Interv. 2022;15:e012417. doi: 10.1161/CIRCINTERVENTIONS.122.012417. [DOI] [PubMed] [Google Scholar]
- 14.Minten L, Wissels P, McCutcheon K, et al. The Effect of Coronary Lesion Complexity and Preprocedural Revascularization on 5-Year Outcomes After TAVR. JACC Cardiovasc Interv. 2022;15:1611–20. doi: 10.1016/j.jcin.2022.06.019. [DOI] [PubMed] [Google Scholar]
- 15.Cole SR, Hernan MA. Constructing Inverse Probability Weights for Marginal Structural Models. Am J Epidemiol. 2008;168:656–64. doi: 10.1093/aje/kwn164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Mehta SR, Wood DA, Storey RF, et al. Complete Revascularization with Multivessel PCI for Myocardial Infarction. N Engl J Med. 2019;381:1411–21. doi: 10.1056/NEJMoa1907775. [DOI] [PubMed] [Google Scholar]
- 17.Engstrøm T, Kelbæk H, Helqvist S, et al. Complete revascularisation versus treatment of the culprit lesion only in patients with ST-segment elevation myocardial infarction and multivessel disease (DANAMI-3—PRIMULTI): an open-label, randomised controlled trial. The Lancet. 2015;386:665–71. doi: 10.1016/S0140-6736(15)60648-1. [DOI] [PubMed] [Google Scholar]
- 18.Dhoble A, Ahmed T, Mckay RG, et al. Timing and Outcomes of PCI in Conjunction With TAVR With Balloon-Expandable Valves. JACC Cardiovasc Interv. 2025;18:244–54. doi: 10.1016/j.jcin.2024.10.055. [DOI] [PubMed] [Google Scholar]
- 19.Praz F, Borger MA, Lanz J, et al. 2025 ESC/EACTS Guidelines for the management of valvular heart disease. Eur Heart J. 2025;46:4635–736. doi: 10.1093/eurheartj/ehaf194. [DOI] [PubMed] [Google Scholar]
- 20.Neumann F-J, Sousa-Uva M, Ahlsson A, et al. 2018 ESC/EACTS Guidelines on myocardial revascularization. Eur Heart J. 2019;40:87–165. doi: 10.1093/eurheartj/ehy394. [DOI] [PubMed] [Google Scholar]
- 21.Rheude T, Costa G, Ribichini FL, et al. Comparison of different percutaneous revascularisation timing strategies in patients undergoing transcatheter aortic valve implantation. EuroIntervention. 2023;19:589–99. doi: 10.4244/EIJ-D-23-00186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Nagaraja V, Ooi S-Y, Nolan J, et al. Impact of Incomplete Percutaneous Revascularization in Patients With Multivessel Coronary Artery Disease: A Systematic Review and Meta-Analysis. J Am Heart Assoc. 2016;5:e004598. doi: 10.1161/JAHA.116.004598. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Stone GW, Ali ZA, O’Brien SM, et al. Impact of Complete Revascularization in the ISCHEMIA Trial. J Am Coll Cardiol. 2023;82:1175–88. doi: 10.1016/j.jacc.2023.06.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Witberg G, Regev E, Chen S, et al. The Prognostic Effects of Coronary Disease Severity and Completeness of Revascularization on Mortality in Patients Undergoing Transcatheter Aortic Valve Replacement. JACC Cardiovasc Interv. 2017;10:1428–35. doi: 10.1016/j.jcin.2017.04.035. [DOI] [PubMed] [Google Scholar]
- 25.López Otero D, Ávila-Carrillo A, González Ferreiro R, et al. Impact of Coronary Revascularization in Patients Who Underwent Transcatheter Aortic Valve Implantation. Am J Cardiol. 2019;123:948–55. doi: 10.1016/j.amjcard.2018.12.007. [DOI] [PubMed] [Google Scholar]
- 26.Faro DC, Laudani C, Agnello FG, et al. Complete Percutaneous Coronary Revascularization in Acute Coronary Syndromes With Multivessel Coronary Disease: A Systematic Review. JACC Cardiovasc Interv. 2023;16:2347–64. doi: 10.1016/j.jcin.2023.07.043. [DOI] [PubMed] [Google Scholar]
- 27.Thiele H, Akin I, Sandri M, et al. PCI Strategies in Patients with Acute Myocardial Infarction and Cardiogenic Shock. N Engl J Med. 2017;377:2419–32. doi: 10.1056/NEJMoa1710261. [DOI] [PubMed] [Google Scholar]
- 28.Zeitouni M, Silvain J, Guedeney P, et al. Periprocedural myocardial infarction and injury in elective coronary stenting. Eur Heart J. 2018;39:1100–9. doi: 10.1093/eurheartj/ehx799. [DOI] [PubMed] [Google Scholar]
- 29.Khan MS, Siddiqi TJ, Usman MS, et al. Meta-analysis Comparing Culprit Vessel Only Versus Multivessel Percutaneous Coronary Intervention in Patients With Acute Myocardial Infarction and Cardiogenic Shock. Am J Cardiol. 2019;123:218–26. doi: 10.1016/j.amjcard.2018.09.039. [DOI] [PubMed] [Google Scholar]
- 30.Böhm F, Mogensen B, Engstrøm T, et al. FFR-Guided Complete or Culprit-Only PCI in Patients with Myocardial Infarction. N Engl J Med. 2024;390:1481–92. doi: 10.1056/NEJMoa2314149. [DOI] [PubMed] [Google Scholar]
- 31.Wald DS, Morris JK, Wald NJ, et al. Randomized trial of preventive angioplasty in myocardial infarction. N Engl J Med. 2013;369:1115–23. doi: 10.1056/NEJMoa1305520. [DOI] [PubMed] [Google Scholar]
- 32.Gershlick AH, Khan JN, Kelly DJ, et al. Randomized trial of complete versus lesion-only revascularization in patients undergoing primary percutaneous coronary intervention for STEMI and multivessel disease: the CvLPRIT trial. J Am Coll Cardiol. 2015;65:963–72. doi: 10.1016/j.jacc.2014.12.038. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Stähli BE, Varbella F, Linke A, et al. Timing of Complete Revascularization with Multivessel PCI for Myocardial Infarction. N Engl J Med. 2023;389:1368–79. doi: 10.1056/NEJMoa2307823. [DOI] [PubMed] [Google Scholar]
- 34.Pivato CA, Gramss L, Joner M, et al. Complete versus incomplete coronary revascularization in patients undergoing PCI after TAVI: insights from the multicenter REVIVAL-PCI registry. Eur Heart J Suppl. 2026;28:039. doi: 10.1093/eurheartjsupp/suag056.039. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data are available upon reasonable request.