Influence of polyvascular disease on clinical outcome in patients undergoing transcatheter aortic valve implantation via transfemoral access

Background The influence of polyvascular disease (PVD) on the short- and long-term clinical outcomes of patients undergoing transcatheter aortic valve implantation via trans-femoral access (TF-TAVI) has not been fully elucidated. Methods A total of 2167 patients from the Optimized CathEter vAlvular iNtervention-TAVI (OCEAN-TAVI) registry who underwent TF-TAVI was studied. PVD was defined as the presence of at least two of the following vascular bed (VB) diseases: concomitant coronary artery disease (CAD), cerebrovascular disease (CVD), and peripheral artery disease (PAD). Results Patients with PVD (288 patients, 13.3%) had a higher incidence of in-hospital complications, such as AKI (16.3% vs. 7.0%, p<0.01) and disabling stroke (3.5% vs. 1.2%, p<0.01) than patients without PVD. These complications caused higher rates of procedural mortality (4.5% vs. 2.0%, p<0.01). PVD increased the risk of the 2-year rate of cardiovascular death (adjusted hazard ratio [HR], 1.61; 95% confidence interval [CI], 1.04–2.50; p<0.05); however, non-cardiovascular death, myocardial infarction, or ischemic stroke was not associated with PVD. Worsening heart failure (4.6% vs. 1.1%, p<0.01) was the main cause of cardiovascular death among patients with PVD. In a sub-analysis, compared with patients with AS alone, those with 2 VB diseases (CAD+PAD; adjusted HR, 1.93; 95% CI, 1.06–3.53; p<0.05) and 3 VB diseases (CAD+CVD+PAD; adjusted HR, 2.61; 95% CI, 1.21–5.62; p<0.05) had a higher risk of 2-year cardiovascular death. Conclusions The increased prevalence of concomitant atherosclerotic VB diseases before TF-TAVI may increase the rates of in-hospital complications and 2-year cardiovascular death. Given the higher rate of mortality in patients with PVD undergoing TF-TAVI, future studies focusing on medical therapy are needed to reduce long-term cardiovascular events in this high-risk subset.


Introduction
Transcatheter aortic valve implantation (TAVI) is the treatment of choice for patients with symptomatic severe aortic valve stenosis (AS) who are deemed inoperative or at a high surgical risk [1]. Comorbid conditions and frailty in routine TAVI clinical practice can affect shortand long-term outcomes. After the recent results from the PARTNER-3 and EVOLUT lowrisk trials [2,3], the indication for TAVI is expected to expand to low-risk patients. Furthermore, in light of the evidence on the long-term durability of TAVI and the corresponding improved survival, it is important to study long-term cardiac events and strategies for optimal management [4].
Risk factors for AS are similar to those for atherosclerosis. Consequently, concomitant vascular bed (VB) diseases, such as coronary artery disease (CAD), peripheral arterial disease (PAD), and cerebrovascular disease (CVD), are often identified before the procedure.
The Reduction of Atherothrombosis for Continued Health (REACH) registry showed that patients with prior ischemic events have a higher rate of subsequent ischemic event [5]. In the literature, patients with polyvascular disease (PVD) have the highest risk of long-term cardiovascular events, and this has been described in many other cardiovascular studies and in the European Society of Cardiology (ESC) guidelines [5][6][7][8][9].
The TAVI population includes patients at high surgical risk, with extremely old, frail, and inoperative patients being the most common. Furthermore, although the influence of CAD, PAD, or CVD on patients undergoing TAVI has been studied [10][11][12], the effect of PVD and combination of these VBs disease (i.e. CAD+PAD, CAD+CVD, PAD+CAD, CAD+PAD +CVD) on clinical outcome are scarcely investigated. Herein, we elucidated the effect of PVD on the short-and long-term clinical outcomes of patients with severe AS undergoing transfemoral TAVI.

Study subjects
Data were collected from the Optimized CathEter vAlvular iNtervention-TAVI registry (OCEAN-TAVI; UMIN000020423, https://ocean-shd.com). This is an ongoing prospective multi-institutional registry study conducted at 14 institutions in Japan. Of the 2588 all-comer consecutive patients with severe AS undergoing TAVI between October 2013 and May 2017, 421 patients treated via an alternative access approach (transapical, 357 patients; direct aortic, 15 patients; trans-subclavian, 19 patients; and trans-iliac, 30 patients) were excluded. We finally enrolled 2167 patients undergoing TAVI via a transfemoral approach. This study adheres to the principles of the Declaration of Helsinki, and the study protocol was approved by the medical ethics committee of all institutions. Informed consent was obtained from all patients.

Definition
PVD and the general diagnostic criteria for each disease territory were defined according to a previous review of PVD and the REACH registry [5,13]. Patients were diagnosed with CAD if one of the following criteria was present: (i) prior coronary revascularization, (ii) stable angina, and (iii) prior myocardial infarction (MI) or unstable angina [5,13]. In this study, scheduled percutaneous coronary intervention (PCI) before and during TAVI was included in prior coronary revascularization.
Patients were diagnosed with peripheral artery disease (PAD) if one of the following criteria was present: (i) intermittent claudication with documented lower artery stenosis/occlusion, (ii) an ankle brachial pressure index (ABI) < 0.9 in either leg, and (iii) previous interventions for angioplasty, stenting, atherectomy, or peripheral artery bypass grafting, or other vascular interventions, including amputation [5,13]. ABI was measured by the oscillometric method in 1908 patients (88%) before TAVI using the VaSera series (Fukuda Denshi, Tokyo, Japan) or Form series (Omron Healthcare, Kyoto, Japan).
Patients were diagnosed with CVD if one of the following criteria were present: (i) history of ischemic stroke, transient ischemic attack, carotid endarterectomy, or carotid artery stenting and (ii) carotid artery stenosis (>50%) detected with ultrasound using the method reported in the North American Symptomatic Carotid Endarterectomy Trial [5,13]. Bilateral carotid artery ultrasonography was performed in 1829 patients (84%).
PVD was defined in the 2017 ESC guidelines as clinically significant atherosclerosis in at least two major artery territories [9]. Most studies on PVD included CAD, CVD, and PAD as at-risk VBs, with aortic and arterial visceral disease significantly less consistently included and/or reported [13]. Therefore, according to a previous review [13], in this study, CAD, CVD, and PAD were considered a component of PVD, and aortic and visceral disease were not included.
Patients were categorized according to the number of atherosclerotic VBs before TAVI as follows: AS alone (AS without known CAD, PAD, or CVD); 1 VB disease: AS+CVD, AS +CAD, and AS+PAD; 2 VB diseases: AS+CVD+PAD, AS+CAD+CVD, and AS+CAD+PAD; and 3 VB diseases: AS+CAD+CVD+PAD. PVD was defined as the presence of 2 or 3 VB diseases: AS+CVD+PAD, AS+CAD+CVD, or AS+CAD+PAD, and AS+CAD+CVD+PAD, respectively.
The clinical frailty scale was measured by a trained medical professional according to the Canadian Study of Health and Aging grading criteria [14]. Before TAVI, anemia was defined as a hemoglobin level of <11 g/dL, and thrombocytopenia was defined as a platelet count <100×10 9 /L [15]. Chronic kidney disease (CKD) stage 4 or 5 was defined as an estimated glomerular filtration rate of <29 mL/min/1.73 m 2 [15].
Single antiplatelet therapy included aspirin, P2Y12 antagonist, or cilostazol, and dual antiplatelet therapy included two of the following drugs: aspirin, P2Y12 antagonist, and cilostazol. Oral anticoagulation (OAC) was defined as warfarin or direct OAC intake.

Endpoint
The rate of in-hospital complications was determined. The primary endpoint in this study was cardiovascular death within the first 2 years postoperatively, and secondary endpoints were non-cardiovascular death within the aforementioned 2 years, all-cause death, ischemic stroke, myocardial infarction (MI), major or life-threatening bleeding, and hospital readmission after TAVI.
According to on-site reports, the cause of death was determined and classified into any of the following modified Definition of Valve Academic Research Consortium-2 criteria [16]: (1) procedural mortality; (2) MI; (3) worsening heart failure (HF); (4) non-coronary vascular death, i.e., pulmonary embolism, ruptured aortic aneurysm, dissecting aneurysm, gastrointestinal ischemia, lower extremity ischemia, or other vascular diseases; (5) stroke; (6) valve-related deaths, such as structured or non-structured valve dysfunction, i.e., infected endocarditis; (7) sudden or unwitnessed death; (8) unknown or other causes of cardiovascular death; and (9) non-cardiovascular death. Procedural mortality comprised all-cause deaths within 30 days or during index procedure hospitalization if the postoperative length of stay was longer than 30 days [16], valve-related deaths included structural or non-structural valve dysfunctions or other valve-related adverse events, strokes included ischemic or hemorrhagic strokes, and non-coronary vascular deaths did not include strokes in this study.

Statistical analysis
Continuous variables are presented as mean ± standard deviation (SD), and dichotomous variables are presented as numbers and percentages. We performed the Pearson bivariate analysis, chi-square test, and Fisher's exact test for categorical analysis between groups. An unpaired ttest or Mann-Whitney U test was used for continuous variables. The Kaplan-Meier survival method was used to compare endpoints between groups by the log-rank test. Univariate and multivariate analyses were performed using a Cox proportional-hazards model to determine the adjusted risk for clinical outcomes except for MI because of the low endpoint count of MI (2 patients (0.09%)). Accordingly, the 47 clinically relevant predictors displayed in Table 1, comprising baseline clinical backgrounds, echocardiography data, medication, and procedure, were entered in the univariate analysis. All factors consistent with previous reports as predictors for endpoints [12,15,[17][18][19][20][21][22][23][24][25] having P-values of <0.01 were entered to the multivariate analysis. P-values of <0.05 were considered statistically significant. SPSS version 21.0 (Chicago, IL, USA) was used for statistical analysis.
Compared with patients without PVD, those with PVD were predominantly male and frail and had a higher body mass index and more comorbidities, including a higher society of thoracic surgeons predicted risk of mortality (STS), higher frequency of concomitant CAG, concomitant CVD, concomitant PAD, a New York Heart Association functional classification 3 Table 1. Patient background, medical therapy at discharge, baseline transthoracic echocardiography findings, and procedural characteristics. or 4, current smoking, dyslipidemia (DL), diabetes mellitus (DM), hypertension (HT), and CKD stage 4 or 5 (Table 1). For medical therapy, beta-blocker, diuretics, and statin were more likely to be administered to patients with PVD than to those without. On transthoracic echocardiography, the left ventricular ejection fraction was lower (56.8% vs. 60.0%, p<0.01), and mitral regurgitation rate was higher (15.3% vs. 10.7%, p<0.05) in patients with PVD than those without. As for procedural findings, rate of surgical cut-down did not differ among groups. The amount of contrast media was lower in patients with PVD than in those without. Table 2 shows the data of in-hospital complications. The incidence of acute kidney injury (AKI: 16.3% vs. 7.0%, p<0.001) and disabling stroke (3.5% vs. 1.2%, p<0.01) was significantly higher in patients with PVD than in those without.

Two-year clinical outcomes
The Kaplan-Meier curve of the 2-year clinical outcomes after TAVI is shown in Figs 1-4.
Compared with patients without PVD, those with PVD had a significantly higher incidence of cardiovascular death (p<0.01, Fig 1A), yet there were no differences in non-cardiovascular death occurrence between patients with PVD and those without (p = 0.42, Fig 1B). Accordingly, the 2-year rate of all-cause death was significantly higher in patients with PVD than those without (p <0.01, Fig 2). In contrast, the incidence of ischemic stroke (p = 0.73, Fig 3A) and acute MI (p = 0.60, Fig 3B) did not significantly differ between patients with PVD and   Table 3).
The findings for sub-analysis of adjusted risk for 2-year clinical outcome in AS patients with VB disease compared with AS alone are shown in Table 4.  the significant adjusted hazard ratio was not shown for AS patients with individual VB disease compared with those with AS alone. Table 5 displays the comparison of causes of death between patients with PVD and those without.
The sub-analysis of the causes of death of patients with individual VB disease is shown in Table 6.

Main findings
This study investigated the association between PVD and clinical outcomes in patients with transfemoral AS. We have found the following points: (a) patients with PVD had a higher incidence of in-hospital complications, such as AKI and disabling stroke, which resulted in higher rates of procedural mortality, than patients without PVD; (b) furthermore, compared with patients without PVD, those with PVD also had a higher 2-year rate of cardiovascular death, yet the 2-year incidence of non-cardiac death, MI, or ischemic stroke was not associated with PVD; (c) the main cause of 2-year cardiovascular deaths was worsening HF, which resulted in a higher cardiovascular death risk among patients with 3 VB (CAD+CVD+PAD) and 2 VB

In-hospital complications and procedural mortality
In our study, in-hospital complication such as stroke and AKI was significantly higher in patients with PVD than those without PVD. Vlastra reported that CVD, CAD, and CKD were independent predictors of stroke on the day of TAVI and after 30 days [12]. Moreover, a previous meta-analysis reported that HT, DM, PAD, low EF, and CKD were predictors of AKI after TAVI. HT and DM were also reported to impair renal autoregulation [26]. Patients with CKD are vulnerable to nephrotoxic drugs, renal hypotension, particularly under low EF settings, and inflammatory mediators during TAVI [26]. In our study, these risk factors were frequently observed in patients with PVD, and might be associated with the occurrence of stroke and AKI in hospital after TAVI. Furthermore, our results may account for the higher burden of atherosclerotic plaque and calcification in patients with PVD, especially in the aorta and valve, compared with that of patients without PVD, which may be disrupted by TAVI. TAVI requires the use of large-sized delivery catheters, balloon valvuloplasty, new valve positioning and implantation, and postdilation with calcified and plaque-rich aortic wall and valve manipulation. Consequently, the dislodgement and embolization of the crushed calcified native valve and aortic debris occur in the brain and kidney.
Our study has shown that patients with PVD had a higher procedural mortality rate (4.5%) than those without PVD. Previous studies have reported the association between in-hospital complications, such as stroke and AKI, and short-term mortality after TAVI [12]. AS = Aortic valve stenosis, CAD = coronary artery disease, CI = confidence interval, CVD = cerebral vascular disease, HR = hazard ratio, PAD = peripheral artery disease, TAVI = transcatheter aortic valve implantation.

Outcomes within two years after TAVI
Patients with PVD had a significantly higher 2-year cardiovascular death and all-cause death rates than those without, which is consistent with the results of previous reports on stable outpatients with acute coronary syndrome (ACS) [6]. However, in our study, there were no differences in the 2-year MI and stroke rates between patients with and without PVD, which is different from the results of previous studies on other atherosclerotic scenarios, wherein patients with PVD had more ischemic events than those without [5,6]. There may be differences between the high-risk TAVI population and other atherothrombotic middle-aged patients, in whom plaque rupture and thrombotic events are the main causes of cardiovascular events [27]. Accordingly, to the best of our knowledge, this is the first report to demonstrate that PVD has effects on clinical outcomes in a high-risk TAVI cohort. In a sub-study, patients with AS and 3 VB (CAD+PAD+CVD) and 2 VB (CAD+PAD) diseases had a higher risk of 2-year cardiovascular death than those with AS alone. A recent metaanalysis has demonstrated that CAD did not affect 1-year prognosis after TAVI [10]. Ben-Shoshan has reported that, unlike surgical bypass or aortic valve replacement, extracranial carotid artery stenosis was not associated with worse outcome following TAVI [28]. However, a previous meta-analysis has revealed that the pre-existence of PAD was related to an increased risk of all-cause death in patients undergoing TAVI [11]. In our study, patients with AS with 1 isolated VB disease (CAD, PAD, or CVD) did not have a statistically significant higher risk of cardiovascular death than patients with AS and no VB diseases. A combination of CAD and PAD in AS patients, and even more accumulated atherosclerotic complications, such as 3 VB diseases, was related to a higher risk of cardiovascular death after TAVI. Accordingly, pre-procedural carotid ultrasound screening and ABI for patients undergoing TAVI is meaningful to detect PVD. We previously reported ABI as a prognostic surrogate marker after TAVI [27]. Future studies including a larger number of PVD subsets are required to determine which VB disease primarily provides more risk with AS alone.  In our study, patients with PVD were predominantly male, which had been reported to represent an adverse factor after TAVI in previous studies [29,30]. Conrotto F reported that factors associated with atherosclerotic disease (i.e. insulin dependent DM, previous stroke, and reduced ejection fraction (EF)) were predictors of midterm mortality in men after TAVI [31]. In our TAVI registry, after adjusting confounders including male, PVD was an independent predictor of cardiovascular mortality for 2 years, suggesting that development of atherosclerotic burden in patients with PVD was related to cardiovascular death irrespective of gender difference.
In this study, the main cause of cardiovascular death was worsening HF. CAD and PAD have been known to be a greater prognostic factor of HF with preserved EF (HFpEF) than CVD [32]. The systemic atherosclerosis mechanism underlying the worsening of HF included the progression of systemic arterial stiffness, vascular endothelial dysfunction, and chronic inflammation [32]. Moreover, LV unloading after TAVI confirmed HF symptom alleviation and LV hypertrophy regression, but this process does not follow a fixed schema, and various factors may affect heart function [33]. In this study, high rates of HT, DM, CKD, and anemia, as well as frailty, in patients with PVD provoked HF. Higher frailty levels are indicative of more frequent rehospitalization in patients with HF and were reported to be mortality predictors following TAVI [14]. In our study, PVD was associated with rehospitalization after TAVI, and in previous studies, the major cause of cardiac rehospitalization was worsening chronic HF (32.6%-46.1%) [24,33]. Hospitalization itself worsened the performance of activities of daily living, resulting in a reduction in muscle strength, the development of sarcopenia, and HF progression [34].
In contrast, non-cardiovascular death was not different between patients with and without PVD. Patients with AS and only 1 VB disease, i.e.., AS+CVD or AS+PAD, had a significantly higher risk non-cardiac death than those with AS only. Atherosclerosis is considered a chronic inflammatory disease related to an abnormal stimulus of the endothelium that is mostly induced by conditions such as DL, HT, DM, and obesity [35]. These conditions have also been considered as predispositions to cancer [35]. Therefore, the elevated levels of pro-inflammatory systemic molecules, such as interleukins, C-reactive proteins, and tumor necrotic factors, have been associated with not only atherosclerosis, but also non-cardiovascular diseases, including several malignancies and infectious diseases [35]. Even patients with 1 VB disease may have a higher risk of non-cardiac death than patients with no VB disease.
The 2-year incidence of major or life-threatening bleeding tended to be higher in patients with PVD than in those without and independently the highest in patients with 3 VB diseases. The rate of CVD and CKD was also higher in patients with PVD than in patients without. These factors, as reported in a PCI population, resulted in a higher bleeding risk [15].

Management and treatment
The vascular screening and precise collection of the history of patients are important to diagnose PVD. This is not only for selecting the appropriate treatment approach, but also for performing risk stratification, even in patients undergoing TF-TAVI. The use of cerebral protection devices or switching to alternative approaches may reduce the incidence of TAVIrelated stroke among patients with PVD in case that high calcified and plaque burden were found in aortic route [12]. Minimum contrast or the use of iso-osmolor contrast media for patients with PVD in addition to CKD may reduce AKI [36,37]. Peri-operative blood transfusion has been reported to be an AKI predictor [36]; accordingly, avoiding bleeding complications is crucial in patients with PVD. Duplex echo-guided femoral puncture and the relative selection of sheaths according to the diameter of the femoral artery using computed tomography is useful to avoid bleeding in patients with advanced atherosclerotic PVD [38].
As for medical therapy, different approaches are required for patients with PVD undergoing TAVI from another atherosclerotic cohort such as CAD, PAD, and CVD without AS. Previous studies reported that ACS after TAVI was rare (<5%), and stroke occurring after 30 days was a natural event in elderly TAVI population [4,12]. PVD was reported to be associated with worsening HF in patients with HFpEF [32]. Thus, for PVD patients who underwent TAVI, according to the results of this study, medical therapy targeting HF may be reasonable. Previous registry data demonstrated that a renin angiotensin blockade, beta-blockers, and statin therapy were associated with favorable clinical outcomes after TAVI [19,39,40]. These drugs possess anti-atherosclerotic and HF-treatable effects. Future randomized study is warranted to elucidate these efficacies for TAVI patients.
In our study, patients with PVD, especially those with 3 VB diseases, had a significantly higher 2-year bleeding risk than patients without PVD. The rates of several bleeding risk factors, including HT, CKD, PAD, frailty, and anemia, were higher among patients with PVD than those without [15]. According to these baseline background data, intensive antithrombotic therapy may be avoided for patients with PVD undergoing TAVI. However, a recent metaanalysis of randomized trials of antithrombotic therapy for secondary prevention of cardiovascular events in the cohort of CAD and/or PAD has demonstrated that intensive antithrombotic therapy benefits patients with PVD [13]. Future studies on better antithrombotic strategy for patients with PVD undergoing TAVI may be required to reduce long-term cardiovascular events.

Limitations
This study has several limitations. First, baseline characteristics were markedly different between both groups, and few patients were in PVD group. Nevertheless, our registry included all-comer patients and monitored the current use of TAVI in real-world settings. This study was a post-hoc investigation. Our findings are hypothesis generating to be further explored in a future randomized trials /prospective registry investigating medical therapy after TAVI with a larger sample size. It can also showcase more specifically which combination of VB disease provides more risk and which optimal medical therapy is effective after TAVI.
Second, recent randomized trials have indicated favorable outcomes of TAVI among lowrisk patients whose comorbidities were different from those in our study and early registries [2]. Future studies regarding the clinical impact of PVD on long-term outcomes in these lowintermediate-risk patients using low-profile new generation TAVI devices should be conducted.

Conclusion
PVD, which is an increase in the number of concomitant atherosclerotic VB diseases, before transfemoral TAVI may increase the incidence of in-hospital complications and the 2-year cardiovascular mortality rate. Given the higher rate of mortality in patients with PVD undergoing TF-TAVI, future studies focusing on medical therapy are needed to reduce long-term cardiovascular events in this high-risk subset.