Highlights
-
•
Stroke remains uncommon following TAVI, but its incidence is not decreasing over time.
-
•
Previous radiotherapy is a novel predictor of post-TAVI stroke.
-
•
Self-expanding valves predicted stroke, an association that requires further evaluation.
-
•
Stroke is associated with higher rates of in-hospital death and nonhome discharge.
-
•
At 12 months, stroke was a powerful predictor of all-cause mortality.
Stroke is a serious complication following transcatheter aortic valve implantation, associated with increased disability and mortality. Despite this, there is a lack of patient-level data examining stroke in contemporary, lower-risk transcatheter aortic valve implantation (TAVI) populations and limited literature exploring the temporal evolution in stroke incidence among real-world cohorts. We aimed to assess the incidence, predictors, and outcomes of stroke in a modern group of TAVI patients across the clinical risk spectrum. We performed a retrospective analysis of prospectively collected data from 2014 to 2025, included in an Australian multicenter registry. Stroke and other outcome measures were defined in accordance with the Valve Academic Research Consortium-3 guidelines. Of 3,823 patients, median age 82 (interquartile range [IQR]: 77, 86), 60% male, and median STS score 3.3 (2.0, 5.3), 54 (1.4%) patients experienced a stroke before hospital discharge. Stroke incidence did not reduce across the study period (2014 to 2017: 0.8% vs 2018 to 2021: 1.4% vs 2022 to 2025: 1.5%, p trend = 0.651). Independent predictors of stroke were previous radiotherapy (adjusted odds ratio [aOR]: 2.8, 95% CI 1.1 to 7.3, p = 0.033), use of a self-expanding valve (aOR: 2.9, 95% CI 1.4 to 6.1, p = 0.004), and nonfemoral access (aOR 3.6, 95% CI 1.1 to 12.2, p = 0.040). Post-TAVI stroke was a strong predictor of 12-month all-cause mortality in Cox proportional hazards regression modelling (adjusted hazard ratio: 7.8, 95% CI 3.9 to 15.7, p < 0.001). In conclusion, stroke remains an uncommon but serious TAVI complication in contemporary practice. Despite demographic and procedural evolution, there was no reduction in stroke incidence across the study period. Further research investigating novel strategies for stroke prevention is required.
Graphical Abstract
Transcatheter aortic valve implantation (TAVI) has transformed the management of severe, symptomatic aortic stenosis (AS), now considered a safe procedure across the clinical risk spectrum of patients. Stroke is a serious complication following TAVI, associated with increased mortality, disability, and healthcare expenditure. , The majority of post-TAVI strokes are embolic, commonly occurring early postprocedure. The etiology is multifactorial and patient-specific, with causes including embolization of cholesterol debris, thrombus formation within guidewires and catheters, air embolism, and endothelial dysfunction secondary to turbulent blood flow. While significant efforts have gone into developing novel therapies for stroke prevention, strategies including cerebral embolic protection (CEP) devices have not resulted in meaningful clinical benefits. With the lack of effective preventative therapies, further research to identify novel stroke risk factors is imperative. Despite its significant impact on patients, there is currently limited patient-level data available on post-TAVI stroke, with most research drawn from meta-analyses and systemic reviews. Importantly, a large proportion of these systematic reviews are not updated, typically including intermediate to high-risk patients. With the rapid advancements in patient eligibility and valve platforms, and the introduction of novel procedural techniques, including device recapture, renewed analysis on the risk factors and outcomes following stroke is required. What also remains largely unknown is how the incidence of stroke has changed over time. While a recent analysis by van Nieuwkerk et al. provides valuable insight, there remains scope for further research in this area using real-world, lower-risk populations. Therefore, this study aimed to provide a perspective on the temporal evolution of stroke incidence and examine predictors and outcomes of post-TAVI stroke using a large, multicenter Australian population of contemporary practice patients.
Methods
The Alfred-Cabrini-Epworth (ACE) TAVI Registry included consecutive patients undergoing TAVI from 2014 to 2025 for severe, symptomatic AS across three high-volume Australian tertiary centers. The registry prospectively collected demographic, procedural, and outcomes data using REDCap (Research Electronic Data Capture) electronic data capture tool. This is a secure web-based software platform hosted at the Alfred Hospital. ,, ACE Registry data are submitted to the National Australasian Cardiac Outcomes Registry TAVI database. This database is governed by the Cardiac Society of Australia and New Zealand and undergoes regular audit. AS severity was defined as an aortic valve area <1 cm 2 or aortic valve mean pressure gradient (MPG) >40 mm Hg on a transthoracic echocardiogram. When the MPG was <40 mm Hg, AS severity was confirmed through additional testing such as dobutamine stress echocardiography or aortic valve calcium scoring. Symptoms were graded using the New York Heart Association (NYHA) classification, and all patients were reviewed by a multidisciplinary heart team before procedure. All participants were provided a written opt-out consent for inclusion into the ACE registry before procedure. The study complied with the Declaration of Helsinki and was approved by the Alfred Hospital ethics committee (Project 146/25).
Stroke was defined as an acute onset of focal neurological signs or symptoms conforming to a focal or multifocal vascular territory, confirmed by pathology or neuroimaging evidence of central nervous system infarction. This definition is in accordance with the Valve Academic Research Consortium-3 (VARC-3) criteria. Other outcome measures were also defined using the VARC-3 criteria.
Categorical variables were compared using the Pearson chi-square test or Fisher’s exact test when expected cell counts were <5, with results expressed as number (%). Continuous variables were analyzed using the Mann–Whitney U test and are expressed as median (interquartile range). A multivariable logistic regression model was developed to identify predictors of in-hospital stroke. Model calibration was assessed using the Hosmer and Lemeshow goodness-of-fit test. Unadjusted mortality rates were compared using Kaplan–Meier curves and the log-rank test. Cox proportional hazards regression modelling was used to estimate hazard ratios with 95% confidence intervals for all-cause mortality at 12 months. Both models included clinically relevant variables that returned p < 0.1 in univariable analysis. The threshold for statistical significance was a two-tailed p value <0.05. IBM SPSS Statistics for macOS, version 30 (IBM Corp., Armonk, NY, USA) was used for statistical analysis.
Results
A total of 3,823 patients undergoing TAVI from 2014 to 2025 were analyzed, with a median age of 82 (interquartile range [IQR]: 77, 86), 60% male, and median STS score 3.3 (2.0, 5.3). A total of 54 (1.4%) patients experienced a stroke before hospital discharge, with the majority (88.9%) ischaemic in nature. The median interval between TAVI and stroke was 1 day (IQR 0, 2), with more than two-thirds (68.5%) occurring within 24 hours of the procedure. Stroke incidence did not reduce across the study period (2014 to 2017: 0.8% vs 2018 to 2021: 1.4% vs 2022 to 2025: 1.5%, p trend = 0.651).
Baseline and procedural characteristics are compared in Table 1 . Preprocedural STS risk-score (stroke: 3.8 vs no-stroke: 3.3, p = 0.504) and the prevalence of common cardiovascular comorbidities including coronary artery disease (74.1% vs 72.9%, p = 0.848), diabetes (27.8% vs 28.3%, p = 0.929), chronic kidney disease (30.8% vs 40.8%, p = 0.142), and atrial fibrillation (27.8% vs 31.6%, p = 0.545) was not different between groups. A previous history of aortic valve endocarditis was present in two patients in the no-stroke group and in no patients who experienced stroke. On transthoracic echocardiography, stroke patients had a smaller aortic valve area (0.7 cm 2 vs 0.8 cm 2, p = 0.043), but there was no difference in other markers of AS severity or frequency of concomitant moderate-severe aortic or mitral regurgitation. Use of nonfemoral access (7.4% vs 1.6%, p = 0.011) was more prevalent in the stroke cohort. Among nonfemoral cases, transaxillary (54%), transcarotid (20.6%), and transaortic (19%) approaches were most used, while transapical (4.8%) and transcaval (1.6%) access was infrequent. Three-quarters of the strokes in the nonfemoral cohort occurred in patients undergoing a transaxillary approach. While use of self-expanding (SE) valves (79.6% vs 61%, p = 0.005) was also more common in the stroke group, other procedural characteristics including procedural time (63 minutes vs 62 minutes, p = 0.698), use of predilation (63% vs 59.1%, p = 0.566), postdilation (17.6% vs 19.8%, p = 0.707), and device recapture of SE valves (23.8% vs 27.6%, p = 0.589) were not different.
Table 1
Baseline demographic, echocardiographic, and procedural characteristics stratified according to stroke status
| Stroke n = 54 | No Stroke n = 3,769 | p Value | |
|---|---|---|---|
| Male | 27 (50%) | 2,251 (59.7%) | 0.148 |
| Age | 83 (78.8, 88) | 82 (77, 86) | 0.237 |
| STS Score | 3.8 (2.3, 5.2) | 3.3 (2, 5.3) | 0.504 |
| Body Mass Index | 26 (22.6, 30.5) | 27.4 (24.3, 31.2) | 0.051 |
| Coronary artery disease | 40 (74.1%) | 2,742 (72.9%) | 0.848 |
| Acute myocardial infarction | 3 (5.7%) | 436 (12.1%) | 0.152 |
| Coronary artery bypass graft | 8 (14.8%) | 501 (13.3%) | 0.745 |
| Surgical aortic valve replacement | 3 (5.6%) | 206 (5.5%) | 1.000 |
| Atrial fibrillation | 15 (27.8%) | 1,190 (31.6%) | 0.545 |
| Heart failure | 25 (50%) | 1,350 (38.4%) | 0.094 |
| Stroke | 9 (16.7%) | 326 (8.7%) | 0.050 |
| Peripheral vascular disease | 7 (13%) | 511 (13.7%) | 0.872 |
| Hypertension | 45 (83.3%) | 2,902 (77.2%) | 0.283 |
| Diabetes | 15 (27.8%) | 1,066 (28.3%) | 0.929 |
| Chronic kidney disease | 16 (30.8%) | 1,499 (40.8%) | 0.142 |
| Radiotherapy | 5 (9.4%) | 148 (4.1%) | 0.068 |
| NYHA III–IV | 22 (43.1%) | 1,517 (41%) | 0.759 |
| Bicuspid Valve | 2 (3.8%) | 45 (1.3%) | 0.144 |
| Mean pressure gradient (mm Hg) | 42 (37.5, 53.5) | 42 (36, 50) | 0.539 |
| Aortic valve area (cm 2) | 0.7 (0.6, 0.9) | 0.8 (0.7, 0.9) | 0.043 |
| Aortic valve peak velocity (m/s) | 4.1 (3.9, 4.6) | 4.2 (3.8, 4.5) | 0.646 |
| Left ventricular ejection fraction (%) | 60 (55, 65) | 60 (55, 65) | 0.300 |
| General anaesthesia | 8 (14.8%) | 363 (10.2%) | 0.265 |
| Nonfemoral access | 4 (7.4%) | 59 (1.6%) | 0.011 |
| Pre dilation | 34 (63%) | 2,184 (59.1%) | 0.566 |
| Self-expanding valve | 43 (79.6%) | 2,297 (61%) | 0.005 |
| Postdeployment dilation | 9 (17.6%) | 726 (19.8%) | 0.707 |
| Contrast volume (mL) | 77.5 (56.3, 100) | 70 (50, 100) | 0.142 |
| Procedure time (minutes) | 63 (50, 79) | 62 (51, 78) | 0.698 |
| Valve-in-valve | 3 (5.7%) | 227 (6.1%) | 1.000 |
NYHA = New York Heart Association.
The multivariable model predicting in-hospital stroke had excellent calibration (Hosmer and Lemeshow goodness-of-fit test chi-square = 5.994, p = 0.648). Previous radiotherapy (adjusted odds ratio [aOR]: 2.8, 95% CI 1.1 to 7.3, p = 0.033), use of a SE valve (aOR: 2.9, 95% CI 1.4 to 6.1, p = 0.004), and nonfemoral access (aOR 3.6, 95% CI 1.1 to 12.2, p = 0.040) were identified as independent predictors of stroke ( Figure 1 ).
Predictors of stroke.
In-hospital outcomes are compared in Table 2 . Stroke patients had a prolonged hospital length of stay (8 days vs 4 days, p < 0.001) and were more likely to experience nonhome discharge (62.5% vs 9.2%, p < 0.001) and in-hospital mortality (9.3% vs 0.2%, p < 0.001). Incidence of other complications, including acute kidney injury and major vascular injury, and predischarge echocardiographic parameters were comparable between groups. At 1-month post-TAVI, residential location was available for 25 of the 30 stroke patients with initial nonhome discharge. Of this group, 20 (80%) had successfully returned home, with the remaining 5 patients still residing in a supportive care facility.
Table 2
In-hospital outcomes stratified by stroke status
| Stroke | No Stroke | p Value | |
|---|---|---|---|
| Hospital length of stay (Days) | 8 (5, 12.3) | 4 (3, 5) | <0.001 |
| Acute kidney injury | 1 (1.9%) | 175 (4.7%) | 0.517 |
| New-onset atrial fibrillation | 3 (6.1%) | 78 (2.2%) | 0.098 |
| Major vascular complication | 0 (0%) | 41 (1.1%) | 1.000 |
| Prosthetic valve endocarditis | 0 (0%) | 1 (0.03%) | 1.000 |
| Intensive care unit admission | 3 (5.6%) | 118 (3.1%) | 0.245 |
| Mean pressure gradient (mm Hg) | 9 (6, 11) | 9 (6, 12) | 0.466 |
| Aortic valve area (cm 2) | 2 (1.6, 2.3) | 1.8 (1.5, 2.2) | 0.248 |
| Nonhome discharge | 30 (62.5%) | 332 (9.2%) | <0.001 |
| Death | 5 (9.3%) | 6 (0.2%) | <0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
