In elderly population, aortic stenosis (AS) coexists with significant coronary artery disease (CAD) in up to 50% of the cases and both diseases share common pathophysiological pattern associated with ageing including oxidative stress, endothelial dysfunction, enhanced inflammation, diabetes mellitus, or chronic kidney disease [1
]. The noxious impact of CAD burden on survival in patients with AS was first suggested by surgical studies in which aortic valve replacement (AVR) combined with coronary artery bypass graft portended higher mortality risk than AVR alone in patients without CAD [5
In patients scheduled for transcatheter aortic valve replacement (TAVR) procedures, current European Society of Cardiology (ESC) guidelines for myocardial revascularization suggest that percutaneous coronary intervention (PCI) should be considered before the index procedure in case of a coronary artery diameter stenosis of >70% affecting a proximal segment [6
]. However, the evidence basis for such method remains limited and the impact of incomplete revascularization remains poorly investigated. Moreover, in this frail population characterized by a high bleeding risk, the noxious impact of subsequent antithrombotic therapies associated to PCI on bleedings events remains unexplored. Tailoring the antithrombotic therapy after TAVR is particularly challenging in this high-risk elderly population with significant overlap of both ischemic and bleeding events.
The residual SYNTAX score (rSS) is an angiographic score that assesses residual CAD burden after PCI [7
]. A recent study [8
] has established that staged PCI by achieving reasonable complete revascularization (rSS ≤ 8) improves mid-term survival and reduces the incidence of repeat PCI in patients with STEMI and multiple vessel disease. By contrast, residual CAD measured by higher rSS confers a worsened prognosis in patients undergoing PCI. In the present study, we sought to evaluate the impact of incomplete revascularization (rSS > 8) on late outcomes in TAVR patients including ischemic but also bleeding events.
2. Materials and Methods
397 patients were enrolled for TAVR with severe AS and high or intermediate surgical risk according to Logistic EuroScore at our institution (Nouvel Hôpital Civil, Université de Strasbourg, France) from November 2012 to December 2013 and then from June 2015 to June 2017. In all patients, aortic annulus diameter and area were determined using cardiac computerized tomography (CT). The aims of the study were explained to all participants and they gave their informed written consent before the procedure and agreed to anonymous processing of their data (France 2 Registry). The study was approved by the CNIL’s (Commission Nationale de l’Informatique et des Libertés) committee (ethical code number 911262). In the case of PCI, patients were pre-treated by P2Y12 inhibitors (mainly Clopidogrel), intravenous aspirin (125–250 mg), and 50–100 IU/kg of unfractionated heparin to target an ACT > 250 s. The use of GPIIbIIIa inhibitors, was left to the operators’ discretion.
Before the TAVR procedure, all patients received aspirin (75–160 mg) and Clopidogrel (loading dose 300 mg, 75 mg/day maintenance dose). The double antithrombotic therapy was ongoing after the procedure for 3 months. Only commercially available valves such as the Edwards SAPIEN XT or S3 prosthesis (Edwards Life sciences LLC, Irvine, CA, USA) and the CoreValve or Evolut-R (Medtronic CV, Irvine, CA, USA) were used as previously described. During the intervention, 100 international units/kg of unfractioned heparin were administered to achieve an activated clotting time of 250 to 350 s. At the end of the procedure, heparin was antagonized with protamine (100 UI/kg). All procedures were performed under analgesic sedation with Ultiva (remifentanil hydrochloride—0.10 to 0.15 microgram/kg/min).
2.1. Calculation of Baseline SYNTAX Score (bSS), Residual SYNTAX Score (rSS) and Syntax Revascularization Index (SRI)
The baseline SYNTAX Score (bSS) was calculated from the pre-procedural angiogram, in which each coronary lesion producing >50% diameter stenosis in vessels >1.5 mm by visual estimation was scored separately using the SS algorithm and added to obtain the overall SS. In patients with angiographic stenosis ≥70% or demonstrated residual ischemia assessed either by fractional flow reserve (FFR) or by perfusion myocardial tomography, a staged PCI was performed. The rSS was defined as the SS recalculated after staged PCI. The rSS was calculated in all patients enrolled in this study. The final post-PCI angiogram was scored to assess untreated disease after staged PCI and to calculate residual SYNTAX scores (rSS). Post-procedural angiograms were reviewed by a dedicated interventional cardiologist who was blinded to both baseline characteristics and clinical outcomes. Likewise, the Syntax revascularization index (SRI), an angiographic index tool designed to quantify the proportion of revascularized myocardium, was calculated and defined as: 100 (1—rSS/baseline SS) (%).
2.2. Collection of Data
Clinical outcomes were recorded and entered into a secure database. Follow-up information was obtained using a written questionnaire via a telephone interview with the cardiologist, referring physician or patient. In the absence of response, the patient’s electronic medical file was consulted. Endpoints were adjudicated by two physicians who were blinded to treatment allocation.
2.3. Study Endpoints
The primary endpoint was the major adverse cardiac event rate (MACE) defined as the composite of cardiovascular death, myocardial infarction (MI) (STEMI or NSTEMI or type 2 myocardial infarction), stroke, and rehospitalization for heart failure (HF). All clinical events were adjudicated by an events validation committee according to the VARC-2 criteria [9
]. ST-segment elevation myocardial infarction (STEMI) was defined as a new ST-segment elevation in two consecutive leads with increased biochemical myocardial necrosis markers and non-ST-segment elevation myocardial infarction (NSTEMI) as the occurrence of ischemic symptoms associated with ST-segment depression or T-wave abnormalities and increased biochemical myocardial necrosis markers. Post-PCI troponin (Tn) elevations were not considered indicative of recurrent myocardial infarction. Stroke was defined as a focal loss of neurologic function caused by ischemic or hemorrhagic events with residual symptoms lasting >24 h. Secondary analyses were performed for each primary endpoint component.
The secondary endpoint was the occurrence of major bleeding and staged according to the BARC (Bleeding Academic Research Consortium) classification [10
]. Major bleeding was defined as a BARC score ≥ Type 3b and minor bleeding as a BARC score < Type 3b.
2.4. Statistical Analysis
Categorical variables are all expressed as count and percentages. Continuous variables are reported as median and interquartile range (25th–75th). The normality of the distribution was assessed graphically with QQ (quantile-quantile) plots and using Shapiro–Wilk tests. Categorical variables were compared with a chi-square test or Fisher’s exact test. Continuous variables were compared using a non-parametric Mann–Whitney test. Event-free survival was calculated with the cumulative incidence function estimated using the competing risk approach of Kalbfleisch and Prentice [11
]. Cumulative incidence was compared between rSS groups using the tests proposed by Gray [12
]. Time to event was defined as the time from TAVR to the date of event, with patients censored at the end of the study and considering death as a competing risk. Multivariable survival analysis was realized using Fine and Gray’s sub-distribution hazard models. Variables with a p
-value < 0.1 in univariate analyses were included in the multivariable model. To prevent expected collinearity between several variables (for instance CTADP post-TAVR >180 s and significant PVL at 1 month (Cramer’s V coefficient ¼ 0.51)), two separate multivariable analyses were performed. Results are presented as sub-distribution hazards ratios (sHR) with their 95% confidence intervals. All tests were 2-sided. A p
value < 0.05 was considered significant. All the analyses were performed using R software version 3.6.0. R Core Team (2019). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/
The current report drawn from a cohort of 311 patients who underwent TAVR is the first study to specifically evaluate the impact of incomplete revascularization on thrombotic but also late bleeding events. The salient results of the present study are as follows: (1) incomplete revascularization (rSS > 8) was observed in a small proportion of the cohort and had no impact on overall and cardiac mortality. (2) Baseline CAD extent and incomplete revascularization were associated with increased MI rates. (3) Baseline CAD extent and incomplete revascularization were predictors of periprocedural bleedings and MLBCs regardless of anti-thrombotic treatment allocation or duration.
Altogether, our findings suggest that baseline CAD extent or incomplete revascularization, could identify a subset of patients with high bleeding diathesis after TAVR.
The first analysis of the eventual detrimental impact of CAD or non-revascularized myocardium on outcomes after TAVR showed no differences between groups on one-year all-cause mortality [13
]. Although the analysis was made on a cohort of limited size (136 patients) and in the early stage of the TAVR era (2005–2007), the low rate of ischemic complication observed following TAVR has substantiated the view that the lowering of ischemic load afforded by PCI would not be mandatory in most of the cases. Likewise, in a cohort of similar size (124 TAVR patients), Van Mieghem and coworkers have emphasized that the completeness of the revascularization by PCI (32% of the cases) did not impact ischemic outcome [14
]. More recent data by Stefanini and coworkers in a larger cohort (n
= 445) have emphasized that both baseline and rSS were associated with enhanced rates of ischemic endpoint including cardiovascular death, stroke, and myocardial infarction, mostly driven by enhanced cardiac mortality. In this study, high rSS tertile (>14) was associated with the higher rate of ischemic endpoints at one year (no CAD 12.5%, low rSS: 16.5%, high rSS: 26.3; p
= 0.043). Of interest, patients with lower rSS (0–14) had comparable outcome to that one observed with complete revascularization or no CAD suggesting that this threshold may constitute an acceptable extent of residual CAD after PCI [15
]. This paradigm of an acceptable residual extent of CAD was recently challenged by the report by Shamekhi demonstrating a stepwise increase in 3-years mortality even observed for low rSS (no CAD 25.9%, low rSS (0–3) 31.4%, high rSS (>3) 41.5; p
= 0.01) [16
]. However, no association between CAD extend and MI rates, stroke, or major vascular complications could be observed at 30 days and longer follow-up endpoints were not investigated. In the present cohort, although a significant increase of MI could be evidenced in bSS > 22 or rSS > 8 sub-groups, no impact on overall or cardiac mortality could be established. The present findings are in line with the Italian CoreValve registry reporting similar one-year MACE (16.8%, 22.7%, 18.5%; p
0.594) and mortality (15.8%, 19.3%, 17.4%) rates in patients with complete, incomplete or no revascularization [17
]. Altogether, these data suggest that the impact of initial or residual atherosclerotic coronary burden if existing appears to a very limited extent and did not impact patient survival.
In the setting of cardiac surgery, the detrimental role of concomitant CAD in patients with aortic stenosis (AS) has been described for many years. Although the combination of surgical aortic valve replacement (SAVR) and CABG increases the risk of periprocedural mortality as compared to the sole aortic valve replacement [18
], there is also compelling evidence underlining that CABG in combination with SAVR reduces the long-term risk of myocardial infarction and mortality [19
]. According to this view recent ESC guidelines recommend performing CABG with SAVR when a coronary artery lesion is ≥70% [6
To the best of our knowledge, the impact of the SYNTAX score in patients treated with SAVR remains unexplored and little described. Regardless of the patient’s risk, large studies comparing TAVR and conventional SAVR have emphasized that the rate of recurrent myocardial infarction appeared comparable [24
]. By contrast, important controversies remain concerning the extent of the bleeding risk. For instance, more bleeding was evidenced in PARTNER I in the SAVR group while SURTAVI showed no difference [24
]. However, it should be emphasized that the direct comparison of the bleeding risk between SAVR and TAVR remains difficult since high heterogeneity in the antiplatelet treatment exists (DAPT vs. SAPT). Periprocedural bleeds may account for a substantial proportion of the risk in surgical procedures, whilst part of the bleeds in the TAVR group occurred at mid- and long-term follow-up as a possible consequence of ongoing DAPT or residual primary hemostasis disorders [28
Another side effect of the liberal use of PCI in TAVR patients worth considering concerns the administration and the duration of dual antiplatelet therapy (DAPT). While DAPT is usually restricted to 3 months in TAVR patients, increased duration of DAPT up to 6 months is generally observed when PCI is performed. Safety concerns and the assessment of bleeding events are key elements when assessing any revascularization and antiplatelet strategy’s net beneficial effects. This crucial aspect was neglected in previous studies questioning the impact of revascularization strategies in TAVR. Several groups including ours have emphasized the high frequency and the noxious impact on mortality of the late bleeding after TAVR [29
]. In that circumstance, improved patient care appears to be in the potential field of bleeding prevention. In patients treated by PCI, several studies have underlined the view that the Syntax Score could not be routinely used for the assessment of bleeding risk [31
]. However, data from the large-scale ACUITY trial have established that high SS remained an independent predictor of 30-days major bleeding [33
]. Accordingly, in a post hoc
analysis of the PLATO trial, the extent of CAD was demonstrated to be an important determinant of the bleeding risk [34
]. In line with this paradigm, the present study suggests that high bSS and rSS could behave as integrate markers of associated comorbidities or global patient sickness known to interfere with the bleeding risk in TAVR setting. The first hint demonstrating the paramount role of PVL in the determination of the bleeding risk was given by Genereux and co-workers. In the PARTNER cohort, the strongest predictor of bleeding events between 30 days and one year after TAVR was PVL [35
]. Several recent studies demonstrated that HMW-multimers defects were induced by significant PVL that was associated with the increase of flow turbulences and the high shear stress forces [36
]. This acquired primary hemostasis disorder was then proposed to explain why patients presenting PVL were more prone to bleed [38
]. We have recently demonstrated that prolonged CT-ADP (>180 s), a surrogate marker of HMW-multimers Von Willebrand defect, as measured during the course of the procedure allows a very accurate identification of the presence of paravalvular leak in patients undergoing TAVR [41
] but could also identify patients at higher risk of periprocedural but also late MLBCs [29
]. In the present study, owing to the high collinearity between PVL, CT-ADP > 180s on one hand and bSS and RSS on the other hand, several models of multivariate analysis were built. In the different models, beyond the importance of PVL (or primary hemostasis disorders reflected by CT-ADP > 180s), initial and residual CAD extent could be pointed out as important determinants of the bleeding risk. The question of whether CAD burden acts primarily as a determinant of bleeding diathesis (i.e., calcifications) or represents only a marker of more intense or sustained anti-thrombotic strategies remains to be determined in dedicated studies. In our hand, we could not exclude that the lower PRI, as a marker of P2Y12 inhibition extent, value observed in patients with high bSS or rSS could have an impact on the rate of periprocedural complications. However, given the limited duration of DAPT in the present study, we do not believe that DAPT allocation could have substantially contributed to the enhanced risk of MLBCs.
This study displays several limitations. First, calculation of bSS and rSS are retrospective and assessed visually by an interventional cardiology expert without using computer software such as the quantitative coronary analysis system. This can cause a variation in the calculation of the Syntax score. Second, only a small proportion of patients with high rSS and bSS could be evidenced which may impact the interpretation of the data. Third, CT-ADP measurement was performed 24 h after TAVR only without being repeated during the follow-up. As a consequence, CT-ADP measurement could not be obtained at the time of the bleeding event. Fourth, anticoagulation treatment at the time of the bleeding event was unknown. Fifth, because of the little size of the cohort and the important proportion of the patients excluded from the analysis, multivariate analysis should be interpreted with caution, and the findings viewed as exploratory and hypothesis-generating. Regarding the number of myocardial infarction and MLBC events that occurred during the follow-up study period, we could not exclude a degree of overfitting in multivariate analyses as well as a loss of power to identify independent predictors. While these data limit to some extent the validity of our comparison, it must be emphasized that registries are mandatory for collecting real-life data on unselected patients.