Author + information
- Received January 13, 2014
- Revision received March 24, 2014
- Accepted April 23, 2014
- Published online October 1, 2014.
- Didier Tchetche, MD∗∗ (, )
- Bruno Farah, MD∗,
- Leonardo Misuraca, MD∗,
- Adele Pierri, MD∗,
- Olivier Vahdat, MD∗,
- Corinne Lereun, MSE, BA†,
- Nicolas Dumonteil, MD‡,
- Thomas Modine, MD§,
- Marc Laskar, MD‖,
- Helene Eltchaninoff, MD¶,
- Dominique Himbert, MD#,
- Bernard Iung, MD, PhD#,
- Emmanuel Teiger, MD∗∗,
- Karine Chevreul, MD∗∗,
- Michel Lievre, PhD††,
- Thierry Lefevre, MD‡‡,
- Patrick Donzeau-Gouge, MD‡‡,
- Martine Gilard, MD, PhD§§ and
- Jean Fajadet, MD∗
- ∗Clinique Pasteur, Toulouse, France
- †Carrigaline Cork, Ireland
- ‡Centre Hospitalier Universitaire Rangueil, Toulouse, France
- §Centre Hospitalier Universitaire Lille, Lille, France
- ‖Centre Hospitalier Universitaire Limoges, Limoges, France
- ¶Centre Hospitalier Universitaire Rouen, Rouen, France
- #Centre Hospitalier Universitaire Bichat, Paris, France
- ∗∗Centre Hospitalier Universitaire Henri Mondor, Creteil, France
- ††University of Lyon, France
- ‡‡Institut Jacques Cartier, Massy, France
- §§Centre Hospitalier Universitaire Brest, Brest, France
- ↵∗Reprint requests and correspondence:
Dr. Didier Tchetche, Clinique Pasteur, 45 avenue de Lombez, 31076 Toulouse, Cedex 3, France.
Objectives The aim of this study was to analyze the incidence, impact, and predictors of cerebrovascular events (CVEs) in patients undergoing transcatheter aortic valve replacement (TAVR).
Background Several issues remain unresolved post-TAVR, including CVEs.
Methods The FRANCE-2 (French Aortic Nation CoreValve and Edwards-2) registry prospectively included all patients who underwent TAVR in France and Monaco from January 2010 to October 2011. A total of 3,191 patients were analyzed. Six-month follow-up data were obtained. Events were adjudicated according to Valve Academic Research Consortium (VARC)-1 definition.
Results Of the cohort, 3.98% experienced a CVE: 55% were major strokes, 14.5% minor strokes, and 30.5% transient ischemic attacks. The mean delay for CVE occurrence was 2 days (interquartile range: 0 to 7 days) with 48.5% of CVEs occurring within 2 days. There was no statistically significant difference in CVE rate with regard to the type of valve (p = 0.899) and the access route (p = 0.128). Patients with a CVE more frequently had new-onset paroxysmal atrial fibrillation (13.6% vs. 7.6%; p = 0.015). During follow-up, the unadjusted mortality rate was higher in patients with a CVE (26% vs. 16.5%; p = 0.002). By multivariate analysis, only advanced age (odds ratio: 1.05; 95% confidence interval: 1.02 to 1.08; p = 0.02) and having 2 valves implanted (odds ratio: 3.13; 95 confidence interval: 1.40 to 7.05; p = 0.006) were associated with a significant risk of CVEs.
Conclusions CVEs occur frequently after TAVR and are associated with an increased mortality rate. No difference exists in the CVE rate when exploring the type of valve or the access route. Advanced age and multiple valves implanted during the same procedure are predictors of CVE.
Transcatheter aortic valve implantation (TAVR) is a valid option for symptomatic patients with aortic stenosis and deemed inoperable or at high risk for conventional surgery. TAVR has proved to be superior to medical treatment for inoperable patients and to be noninferior to surgery in selected high-risk patients (1,2). Despite worldwide growing experience, several TAVR-related issues remain unresolved and insufficiently explored. Among these issues, cerebrovascular events (CVEs) are one of the most serious complications, potentially affecting patient survival, autonomy, and quality of life. CVEs before and after TAVR may be multifactorial and include embolic debris or thrombus, aortic dissection, hemodynamic instability, and bleeding. The randomized PARTNER (Placement of Aortic Transcatheter Valves) trial raised concern about higher stroke rates post-TAVR compared with medical therapy or surgical aortic valve replacement (1,2). In other reports, stroke rates range from 1.5% to 20% (3,4). The FRANCE-2 (French Aortic National CoreValve and Edwards-2) registry is the largest TAVR registry published thus far on consecutive patients (5). We aimed to explore the incidence, timing, impact, and predictors of CVEs in this large cohort of patients.
The FRANCE-2 registry is a prospective database that included all patients who underwent TAVR from January 2010 to October 2011 at 33 selected French centers and 1 center in Monaco. Patients were eligible for TAVR if they had a combination of symptomatic severe aortic stenosis, comorbidities contraindicating surgery or EuroSCORE (Logistic European System for Cardiac Operative Risk Evaluation) ≥20% or a Society of Thoracic Surgery score ≥10%, and a life expectancy of >1 year. The final decision for TAVR and access route was made on the basis of an on-site multidisciplinary heart team discussion. The FRANCE-2 registry was established under the French Societies of Cardiology and Thoracic and Cardiovascular Surgery. All patients provided written informed consent before the procedure, including consent for anonymous processing of their data. The registry was approved by the institutional review board of the French Ministry of Health. Valve manufacturers funded the registry but did not have any role in data collection or analysis or in the preparation of the manuscript. All patients were followed at 1 month and 6 months and then annually for 5 years.
Each heart team could choose to implant 1 of 2 commercially available transcatheter aortic valves: the balloon-expandable Edwards SAPIEN (ES) or SAPIEN XT (ESXT) valves (Edwards Lifesciences, Irvine, California) or the self-expandable Medtronic CoreValve (MCV) and MCV Accutrak (Medtronic, Minneapolis, Minnesota). ES and ESXT valves were implanted via the transfemoral, transapical, or direct aortic routes, whereas the MCV valves were implanted via the transfemoral, subclavian, or direct aortic approaches. Procedures were performed in a hybrid room (surgical operating room with integrated fluoroscopy and imaging modalities), a catheterization laboratory, or a conventional operating room according to the policy of each center. General anesthesia or conscious sedation were used according to the heart team’s preference. Patients were treated with aspirin and clopidogrel before the procedure (aspirin alone before transapical or direct aortic approaches). Unfractionated heparin was given at the start of the procedure aiming at an activated clotting time of 200 to 300 s. After TAVR, long-term aspirin therapy (75 to 300 mg) was prescribed, and clopidogrel was prescribed for 1 to 6 months according to each center’s policy. For patients taking warfarin, aspirin was added (75 to 300 mg) for 1 to 6 months according to each center’s preference.
Definitions of a CVE
Valve Academic Research Consortium (VARC)-1 definitions (6) were used. A CVE was defined as a rapid onset of a neurological deficit.
No other readily identifiable nonstroke cause had to be present at the clinical presentation.
The confirmation of the diagnosis of stroke was made by at least 1 of the following: neurological or neurosurgical specialist, neuroimaging procedure (magnetic resonance imaging, computed tomography, or cerebral angiography), or lumbar puncture (i.e., spinal fluid analysis diagnostic of intracranial hemorrhage). CVEs were divided into 3 categories:
• Transient ischemic attack (TIA): new focal neurological deficit with rapid symptom resolution, always within 24 h. Neuroimaging without tissue injury.
• Minor stroke: diagnosis as above with a modified Rankin Scale score <2 at 7 days post diagnosis or before discharge and National Institutes of Health Stroke Scale (NIHSS) score <3 above baseline at 7 days or before discharge and at 30-day assessment.
• Major stroke: modified Rankin Scale score >2 at 7 days or before discharge and NIHSS score >3 above baseline at 7 days or before discharge and at 30-day assessment.
If there was discordance between modified Rankin Scale score and NIHSS score determinations of major versus minor stroke, final adjudication was made by a qualified neurologist.
New-onset atrial fibrillation
After TAVR, patients stayed in an intensive care unit for at least 2 days and then were followed with 24-h electrocardiographic monitoring until discharge.
During the hospital monitoring, new-onset paroxysmal atrial fibrillation was defined as a new episode of atrial fibrillation lasting >30 s but <1 day. New-onset persisting atrial fibrillation was defined as a new episode of atrial fibrillation lasting >1 day.
During the outpatient clinic follow-up, new-onset persisting atrial fibrillation was defined as a new episode documented on 12-lead electrocardiography. The follow-up modalities did not allow for identification of episodes of paroxysmal atrial fibrillation.
A total of 3,191 patients of the 3,195 recorded in the FRANCE-2 registry were included in the analysis (date of TAVR was not reported for the 4 excluded patients). The study population was divided into 2 groups: patients who experienced a CVE during the follow-up period and those who did not. Patient characteristics and details of the procedure were described using summary statistics: median and interquartile range for continuous variables, as none of them were normally distributed, and frequencies and percentages for categorical variables. Differences between the 2 groups were tested using Wilcoxon rank sum test for continuous variables and the chi-square or Fisher exact test for the categorical ones. Patient characteristics and details of the procedure were explored as potential predictors of occurrence of CVEs using logistic regression. Each potential predictor was tested in a univariate model, and those that were significant (p ≤ 0.05) were combined in a multivariate model. The results of these models are reported as odds ratios. Mortality was analyzed using the Kaplan-Meier method. All analyses are performed using Stata SE version 8.2. (StataCorp., College Station, Texas).
The baseline characteristics of the study population are reported in Table 1. Patients experiencing CVEs were older (84.7 ± 6 years vs. 82.6 ± 7.2 years; p < 0.001). There were significantly more women experiencing CVEs (59.1% vs. 48.6%; p = 0.02). The patients without CVEs more frequently had chronic atrial fibrillation (7.6% vs. 0.8%; p = 0.004). There was no other difference at baseline between both groups of patients. Of note, a history of stroke, porcelain aorta, and peripheral vascular disease were, respectively, observed in 10%, 8%, and 28.5% of the population.
Patients were followed for an average of 5.2 ± 4.7 months after TAVR (median, 4.3 months), and 13% were followed for at least 12 months. Of the cohort, 127 patients had a CVE (3.98% of the population). A total of 123 patients had 1 event, and 4 patients had 2 events (3 patients with 2 major strokes and 1 patient with 2 TIAs). During follow-up, 2.2% of the patients had a major stroke, 0.6% had a minor stroke, and 1.2% had a TIA. These events occurred shortly after the procedure: 2 days (range, 0 to 7 days) (Table 2, Figure 1). Most of CVEs (48.5%) were diagnosed within 48 h after TAVR, whereas 36.9% occurred between 2 and 30 days and 14.6% after 1 month.
Transfemoral access was used in 74.6% of the population, and 66.9% of the patients received ES or ESXT valves (Table 3). There was no statistically significant difference in stroke rate with regard to the type of valve (p = 0.07), the transfemoral approach using the ES/ESXT or MCV (p = 0.968), or the procedural success rate (p = 0.259). For nontransfemoral access, there was no difference between approaches (p = 0.279). No significant difference was identified between groups of patients with and without CVEs in terms of procedural details except the need for multiple valves. In patients experiencing CVEs, 2 valves were more often used during the procedure (5.5% vs. 2.1%; p = 0.012). There were more patients requiring 2 MCVs than patients treated with 2 ES/ESXTs (3.5% vs. 1.7%; p = 0.001), without any statistical difference in terms of the CVE rate. Valve success was similar between patients with and without CVEs (95.3% vs. 97%; p = 0.259), despite a higher rate of multiple valves used in the first group. The main reason for using a second valve was valve embolization resolved by either retrieval or impaction in an extra-annular location.
Post-procedural outcome and survival
The post-procedural outcome is detailed in Table 3. The length of intensive care unit stay and hospital stay were significantly longer for patients with CVEs (4.9 ± 5.1 days vs. 3.9 ± 4.7 days, respectively; p = 0.018 and 13.2 ± 9.5 days vs. 9.9 ± 7.9 days; p < 0.001). During the in-hospital period, patients with CVEs more frequently had new-onset paroxysmal atrial fibrillation (13.6% vs. 7.6%; p = 0.015). During the whole follow-up period, new-onset persisting atrial fibrillation occurred more frequently in patients without CVEs (7.6% vs. 0.8%; p = 0.004). Of patients who had a neurological event, 26.0% died during the follow-up period compared with 16.5% of patients who did not (p = 0.005). Kaplan-Meier survival curves confirmed a significant and sustained increase in all-cause unadjusted mortality rate in patients with a CVE during follow-up (Figure 2).
Predictors of CVEs
On univariate analysis, female sex, older age, and 2 valves used in the same patient were the only predictors of CVE after TAVR (Table 4). On multivariate analysis, only advanced age and 2 valves placed were associated with a significant risk of CVE (Table 5). Pacemaker implantation was not associated with an increased risk of CVE (odds ratio: 0.15; 95% confidence interval: 0.05 to 0.47; p = 0.001).
The FRANCE 2 registry is, to our knowledge, one of the largest prospective registries exploring CVEs. These events occur frequently and have a negative impact on the patient survival after TAVR.
The stroke rate reported in the FRANCE 2 is concordant with data from the literature. Eggebrecht et al. (3) performed a meta-analysis of 53 studies including 10,037 patients and identified a stroke rate of 5.2 ± 3.4% during the first year. The main limitation of this analysis is the lack of uniformity in the definitions used for stroke. Jilaihawi et al. (7) analyzed the outcome of 5,024 patients who underwent TAVR and found a stroke rate of 2.6% at 30 days. Genereux et al. (8) performed another meta-analysis of 16 studies reporting Valve Academic Research Consortium outcomes and identified a 30-day risk of major stroke of 3.2% (95% confidence interval: 2.1% to 4.8%). In cohort A of the PARTNER 1 trial, rates of stroke and TIA were higher among patients undergoing TAVR compared with surgical aortic valve replacement at 30 days (4.6% vs. 2.4%; p = 0.12) and up to 2 years of follow-up (11.2% vs. 6.5%; p = 0.05) (9). In cohort B of the PARTNER 1 trial comparing TAVR with medical treatment in inoperable patients, the stroke/TIA rate was significantly higher in patients treated with TAVR: 6.7% vs. 1.7%, p = 0 .02 at 30 days; 11.2 vs. 5.5%, p = 0.06 at 1 year; and 13.8% vs. 5.5%, p = 0.01 at 2 years of follow-up (10).
We could not precisely determine the ischemic or hemorrhagic nature of the CVEs observed in the FRANCE 2 because our database did not contain that information. Nombela-Franco et al. (11) identified, in a collaborative retrospective registry of 1,061 patients, hemorrhagic events in only 2 patients. Nuis et al. (12) identified a 5% hemorrhagic stroke rate in a single-center series of 214 patients treated with the MCV. Therefore, the vast majority of TAVR-related CVEs seem to be ischemic. More precisely, Makkar et al. (10) analyzed the stroke rate at 2 years in cohort B of the PARTNER trial and observed that the stroke rate was higher after TAVR than after medical therapy (13.8% vs. 5.5%; p = 0.01). In the first 30 days, there were more ischemic CVEs after TAVR (6.7% vs. 1.7%; p = 0.02), and beyond 30 days, there were more hemorrhagic CVEs after TAVR (2.2% vs. 0.6%; p = 0.16).
We identified a period of stroke vulnerability, extending to the first month after TAVR with a mean delay of occurrence of 22.9 ± 59.5 days (median, 2 days). Almost half of the CVEs were acute and periprocedural events, occurring within 48 h after valve replacement. This is consistent with findings of Nombela-Franco et al. (11) and Nuis et al. (12), who identified acute stroke (within 24 h) in, respectively, 54% and 42% of their patients. These acute events can occur at various steps of the procedures, for example, crossing the aortic valve, balloon valvuloplasty, placement of the valve, deployment of the valve, and withdrawal of the delivery system. Kahlert et al. (13) performed an interesting study of serial transcranial Doppler imaging in 83 patients to identify high-intensity transient signals (HITSs) as a surrogate for microembolization. Periprocedural HITSs were detected in all patients, mainly during manipulation across the native aortic valve and implantation of stented valves. The balloon-expandable ES valve caused significantly more HITSs during positioning and the self-expandable MCV valve during deployment. Overall, there were no significant differences between transfemoral and transapical TAVR or between the MCV and ES valve. In this series, only 1 subsequent major stroke and 1 minor stroke were diagnosed.
CVEs are associated with an increased mortality. In elderly and frail patients, stroke has a clear negative impact on autonomy and quality of life. In a series of 389 patients, Stortecky et al. (14) reported that patients with CVEs had an increased risk of all-cause and cardiovascular mortality compared with patients without CVEs at 30 days of follow-up. Nombela-Franco et al. (11) confirmed the same findings at 30-day and 1-year follow-up with major stroke and its permanent functional disability being associated with the worst outcome.
On multivariate analysis, age and multiple valves used in the same procedure were predictors of CVEs. Female sex was a predictor of CVEs only on univariate analysis. A significant proportion of CVEs associated with female sex may be explained by bleeding. The PRAGMATIC (Pooled-RotterdAM-Milano-Toulouse In Collaboration) registry reported a higher stroke rate in patients experiencing bleeding after TAVR and identified female sex as a risk factor for bleeding and transfusion. Complex procedures with multiple valves may increase the risk of stroke by increasing the risk of debris embolization. The embolic nature of periprocedural stroke was elegantly demonstrated by Van Mieghem et al. (15) by using a dual filter–based embolic protection device in 40 patients undergoing TAVR. Macroscopic material liberated during the procedure was captured in the device filter baskets in 30 patients (75%), varying in size from 0.15 to 4.0 mm and consisting of fibrin or amorphous calcium and connective tissue derived most likely from either the native aortic valve leaflets or the aortic wall. Although new-onset atrial fibrillation was more frequently observed in patients with CVEs, it was not identified as a predictor of stroke in the FRANCE 2. This event was mainly captured during in-hospital monitoring, and its real frequency may be underestimated because of the difficulty in recording it during a longer follow-up with electrocardiograms obtained only at 1 month, 6 months, and then yearly. Interestingly, persisting atrial fibrillation occurred more frequently in patients without CVEs. One explanation could be a more thorough and prolonged anticoagulant treatment in those patients, most of the time in combination with aspirin alone. Conversely, it is likely that patients with new-onset atrial fibrillation did not all receive adequate anticoagulant treatment at the time of the event. Amat-Santos et al. (16) recorded new-onset atrial fibrillation during the in-hospital follow-up in 31.9% of their TAVR patients. Nombela-Franco et al. (11) identified post-dilation balloon and valve dislodgment/embolization as predictors of acute CVEs (<24 h), whereas new-onset atrial fibrillation determined a higher risk for subacute CVEs (1 to 30 days), and the number of late events (>30 days) were higher in patients with a history of chronic atrial fibrillation, cerebrovascular disease, and peripheral vascular disease. Pacemaker implantation was not associated with an increased risk of CVEs. Only 3 patients experienced the sequence of having a pacemaker and then a CVE during follow-up. Potential explanations could be a sustained sinus rhythm after pacemaker placement or a more meticulous assessment and management of rhythm disturbances (e.g., new-onset atrial fibrillation) in patients receiving a pacemaker.
Considering these risk factors for CVEs, protective maneuvers could be associated with the use of an embolic protection device, limiting manipulation across the native aortic valve, avoiding balloon valvuloplasty, maintaining effective anticoagulant treatment, or using antiarrhythmic drugs in the stroke vulnerability period post-TAVR. The anticoagulant treatment should be tailored to the bleeding risk of each patient because bleeding and transfusion have been reported to carry an intrinsic increased risk of stroke (17).
There was no central core laboratory adjudication of the events. CVEs and other complications were reported by each participating heart team.
Several predictors of neurological events could not be explored because they were not collected in the database. The type of anticoagulant treatment at the time of stroke was not consistently reported. The use of Valve Academic Research Consortium 2 definitions was not retrospectively possible because the FRANCE 2 database does not contain modified Rankin Scale scores at 90 days. The follow-up at the time of the analysis was relatively short and will be extended by 5 years to provide more robust results of the FRANCE 2 registry.
CVEs occur frequently after TAVR and mainly occur within 48 h, with a stroke vulnerability period extending to 1 month post-procedure. These events are associated with an increased unadjusted mortality rate. No difference exists in the CVE rate with regard to the type of valve or the access route. Advanced age and multiple valves used during the same procedure are predictors of CVEs.
Dr. Tchetche is a consultant for Edwards Lifesciences and Medtronic. Dr. Eltchaninoff is a consultant for Edwards Lifesciences. Dr. Iung has received speaker’s fees from Edwards Lifesciences; and is a consultant for Abbott Vascular, Boehringer Ingelheim, and Valtech. Dr. Teiger is a consultant for Edwards Lifesciences and Medtronic. Dr. Lievre is a consultant for Medtronic. Dr. Lefevre is a proctor for Edwards Lifesciences. Dr. Dumonteil has received proctoring fees from Edwards Lifesciences, Medtronic, and Boston Scientific; and is a consultant for Biotronik. Dr. Modine is a consultant for Medtronic and Boston Scientific. Dr. Himbert is a proctor and consultant for Edwards Lifesciences; and a consultant for Medtronic. Dr. Lievre has served on the Advisory Board of Novo Nordisk, Merck Serono, and Novartis; and has received an institutional grant from Medtronic. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- cerebrovascular event
- Edwards SAPIEN
- Edwards SAPIEN XT
- high-intensity transient signal
- Medtronic CoreValve
- National Institutes of Health Stroke Scale
- transcatheter aortic valve replacement
- transient ischemic attack
- Valve Academic Research Consortium
- Received January 13, 2014.
- Revision received March 24, 2014.
- Accepted April 23, 2014.
- American College of Cardiology Foundation
- Chieffo A.,
- Buchanan G.L.,
- Van Mieghem N.M.,
- et al.
- Leon M.B.,
- Piazza N.,
- Nikolsky E.,
- et al.
- Jilaihawi H.,
- Chakravarty T.,
- Weiss R.E.,
- Fontana G.P.,
- Forrester J.,
- Makkar R.R.
- Genereux P.,
- Head S.J.,
- Van Mieghem N.M.,
- et al.
- Nombela-Franco L.,
- Webb J.G.,
- de Jaegere P.P.,
- et al.
- Kahlert P.,
- Al-Rashid F.,
- Dottger P.,
- et al.
- Van Mieghem N.M.,
- Schipper M.E.,
- Ladich E.,
- et al.
- Amat-Santos I.J.,
- Rodes-Cabau J.,
- Urena M.,
- et al.
- Tchetche D.,
- Van der Boon R.M.,
- Dumonteil N.,
- et al.