A temporal distribution of neurological events following transcatheter aortic valve replacement (TAVR) has been recognised. Acute cerebrovascular events (CVE) occurring in the first 24 hours after a TAVR procedure are usually of ischaemic origin, primarily embolic. Commonly recognised factors contributing to procedural risk include the manipulation of rigid, large-bore delivery catheters, inflation of valvuloplasty balloons and positioning of valve frames within a diseased or calcified aortic root. They lead to the dislodgment of thrombotic, atheromatous or calcific debris. As such, it has been postulated that cerebral embolic protection devices (CEPD) could reduce periprocedural strokes.
Late-onset strokes, however, are associated with patient factors, such as diabetes, peripheral vascular disease and previous stroke, among others. The role of CEPD in such high-risk populations is less certain. Short- and long-term sequelae of stroke, including worse quality of life and a higher mortality, remain a challenge to health services.1
The purpose of this review is to explore any sex differences in the pathophysiology, preventive measures and therapeutic options for stroke following TAVR.
Pathophysiology of Stroke after Transcatheter Aortic Valve Replacement
Stroke has been described in terms of severity and subsequent disability according to the National Institutes of Health Stroke Scale and the modified Ranking Scale, respectively.2 Several studies noted that the incidence peaks during the acute periprocedural phase (24 hours) after the TAVR procedure, reaching half of the total events.3 Thereafter, patients remain most vulnerable for a period up to 1 month (subacute periprocedural events) before transitioning to a late, constant, but decreased, hazard phase.4 The inclusion of women in pooled meta-analyses and randomised trials ranged from 30% to 40% and does not allow for the identification of definitive sex-related differences in outcomes at any phase.
Periprocedural Events
Acute CVEs after TAVR are mostly embolic (Figure 1 ).5 Any manipulation of rigid and large-bore delivery catheters, inflation of valvuloplasty balloons, and positioning of valves frame within a diseased aortic root and calcified aortic valve could potentially lead to the dislodgment of thrombotic, atheromatous or calcific debris. The role of postdilatation remains less definitive in increasing the risk of an embolic event. In general, pre- and postdilatation are reserved for special conditions, such as bicuspid valves, heavily calcified annuli and self-expanding platforms with a lower radial force.6–10
An embolic phenomenon has been supported by several theories, including the higher incidence of high-intensity signals in the middle cerebral artery during valve deployment, as observed by transcranial doppler, and the direct detection of fibrin/thrombotic materials in the histopathological analysis of debris retained in embolic protection devices during TAVR.11 The published trials do not allow a granular analysis of the main driver of embolic strokes in women.
Non-embolic aetiologies of ischaemic stroke include severe acute aortic regurgitation, major bleeding and prolonged rapid ventricular pacing, which may lead to haemodynamic instability and prolonged hypotension.12 A subclinical inflammatory response leading to prothrombotic activation has been described as another less well-understood factor.13 Bioprosthetic leaflets may induce platelet deposition or activation and aggregation, thereby increasing the thrombotic periprocedural risk in a fashion similar to the endothelial injury. The potential role of short-term dual antiplatelet therapy after TAVR ought to be addressed in future studies, particularly in small anatomies frequently encountered in women.14
Subacute and Late Events
Subacute and late CVEs have a multifactorial origin that may be attributed to patient-related risk factors, such as age and comorbidities (i.e. peripheral vascular disease, chronic AF, kidney disease and new-onset AF).15,16 New-onset AF occurs in approximately half of the cases within 24 hours, and in >80% of patients within 3 days after the TAVR procedure.
Strokes occurring after 90 days from TAVR are not considered procedure-related and generally entail management of atherosclerotic risk factors and comorbidities. Prior stroke, carotid disease, prior cardiac surgery or a history of peripheral arterial disease are well-documented factors that increase the risk of stroke. The OBSERVANT-II registry, an observational, prospective, multicentre cohort study including 2,753 patients, the majority of whom were women (52.3%), reported a stroke rate of 1.3% (with 37.1% of disabling stroke) at 30 days and 2.4% at 6 months after transfemoral (91.4%) TAVR with the four commonly used transcatheter heart valves (THVs) on the market (Evolut R/Pro, Sapien 3, Acurate Neo, Portico). Aortic valve predilatation, bicuspid valve, diabetes and a reduced left ventricle ejection fraction (<50%) were recognised as independent predictors of 30-day stroke.17
The role of female sex as an independent risk factor for late CVEs has been established.18–21 The lifetime risk of stroke is higher in women and they have worse stroke outcomes than men. Several unique factors, such as a history of pregnancy-related hypertensive disorders, gestational diabetes, early menopause, exogenous oestrogen and oral contraceptive pills, contribute to the higher risk.22 Women have a lower age-adjusted incidence of AF compared with men.23 However, due to the longer life expectancy for women, the lifetime risk of AF in women and the absolute number of women with AF are similar to those of men men.24 Female sex itself is considered a variable in AF risk stratification models, such as the CHA2DS2-VASc score.25 More recently, it has been suggested that female sex acts as a risk modifier rather than an independent risk factor for stroke in AF.26
Current Evidence
The incidence of cerebrovascular events following TAVR varies significantly across studies.2 Contemporary evidence showed that the rate of CVE is highly dependent on the systematic evaluations by neurologists and the use of routine imaging studies, both of which have been inconsistently applied in recent trials.3,4 The use of diffusion-weighted MRI (DW-MRI) documented new ischaemic CVEs in up to 90% of those undergoing a TAVR procedure. However, there is poor correlation between clinically apparent neurological deficits and these DW-MRI findings.5 To standardise the CVE definition and adjudicate it, the Valve Academic Research Consortium and Neurologic Academic Research Consortium recommended combining appropriate assessment of neurological symptoms with tissue-based criteria for defining stroke, identifying several subcategories according to the type of tissue damage, ascertaining the aetiology and pinpointing the time of onset.6
The SMART trial compared self-expanding valves (SEV) and balloon expandable valves in small annuli. The investigators enrolled patients with a mean age of 80 years, 87% of whom were women.27 Disabling stroke (a component of the composite endpoint of mortality, disabling stroke or rehospitalisation for heart failure up to 12 months) was low, occurring in 3.1% of SEV and 2.6% of balloon expandable valves. The trial concluded that SEV was non-inferior to balloon expandable valves for the primary endpoint. Despite the lack of data on predilatation in both groups, the more frequent need for balloon dilatation before SEV could explain the difference in disabling stroke occurrence in the two groups.27
The WIN TAVI registry identified stroke as one of the independent determinants of the primary clinical endpoint at 1 year. The rate of stroke was 2% at 1 year. Interestingly, newer-generation THVs did not demonstrate significant differences in disabling stroke between men and women.28
Role of Embolic Protection
The goal of CEPD prior to the THV deployment is to protect the cerebral circulation by capturing embolic debris released during the procedure. Different types of debris have been retrieved from the majority of patients who have undergone TAVR with a filter-based CEPD, such as acute and organising thrombus, calcium, atheroma, valve tissue, arterial wall, and foreign material. A sex-based difference in the type of debris has been reported. There is a trend towards more valve tissue being captured observed in women compared with men (63 versus 43%; p=0.0688).29,30 Female sex (p=0.0287) and diabetes (p=0.0116) were predictive of higher rates of debris in the filters.29–31
There are two categories of CEPD: devices that capture (totally or partially) debris within the aorta; or deflectors of debris from the aortic arch and its branches. They can also be classified into partial capture devices (SENTINEL), full capture devices (Emblok, Innovative Cardiovascular Solutions; Emboliner, Emboline; FLOWer, AorticLab), primarily deflective devices (with small capture capacity; TriGUARD 3, Keystone Heart/Venus Medtech; ProtEmbo, Protembis; POINT-GUARD, Transverse Medical), and deflection and capture devices (CAPTIS, Filterlex; Table 1 ).32
The Sentinel Cerebral Protection System (Claret Medical) is the most commonly used CEPD with the most extensive trial data.33 The system consists of two polyurethane filters with 140 mm diameter pores fixed in a flexible nitinol radiopaque frame, preferably introduced through a 6 Fr sheath through the right radial or brachial artery over a 0.014" guidewire, and deployed into the ostia of brachiocephalic trunk and left common carotid artery.
Studies using the Sentinel CEPD system have yielded inconsistent results.34–36 The MISTRAL-C study was the first multicentre, double-blind, randomised trial conducted to determine whether use of the Sentinel Cerebral Protection System during TAVR decreases the incidence of new brain lesions as assessed by DW-MRI and prevents neurocognitive decline.37 DW-MRI performed 1 day before and 5–7 days after TAVR demonstrated that new brain lesions were found in 78% of patients with follow-up MRI.37 Patients with the Sentinel Cerebral Protection System had numerically fewer new lesions and a smaller total lesion volume.37 The study population consisted of 48% of women with equal representation in both arms: Sentinel 47% versus non-Sentinel 49% (p=0.897). The primary limitation of the study is a small sample size and a higher-than-expected MRI dropout rate, making the study grossly underpowered for any sex-specific subgroup analysis.37
The SENTINEL study was a larger multicentre study (n=363 patients), with a 1:1:1 randomisation into a safety device arm (n=123), an imaging device arm (n=121) and an imaging control arm (n=119). Despite a numerical reduction in all-cause stroke at 30 days, statistical significance was not met (5.6% for the SENTINEL group versus 9.1% in the control group; p=0.25). However, the authors reported debris in 99% of the filters and no change in neurocognitive function. The study population was 52.1% female, with no significant differences between three subgroups: safety arm (55.3% female patients), device arm (52.1%) and control arm (48.7%).32 A separate analysis regarding sex outcomes was not performed.
An analysis by Alkhouli et al., which included a total of 10,985 patients, showed that the use of CEPD was not associated with a statistically significant reduction in post-TAVR stroke or transient ischaemic attack.38 There was a higher number of women in the no-CEPD subgroup in unmatched cohorts, yet no sex-specific analysis was conducted.38 Three large-scale, non-randomised, retrospective studies analysed the outcomes of CEPD use in patients undergoing TAVR in real-world practice and the results were conflicting.32–38
Isogai et al. evaluated the Nationwide Readmissions Database from 2018 to 2019, including 136,382 TAVR recipients (10,201 CEPD users and 126,181 CEPD non-users).35 Patients with stroke after TAVR with CEPD had significantly lower in-hospital mortality than those with stroke after TAVR without CEPD (6.3% versus 11.8%; p=0.023).35 The rate of stroke after TAVR in those who received a CEPD compared with those who did not was higher in men compared with women (56.1% versus 43.9% and 44.8% versus 55.2%, respectively; p=0.003).35
However, Butala et al.’s analysis of 123,186 patients from 599 sites did not find an association between CEPD use for TAVR and in-hospital stroke.36 There were no meaningful differences in the baseline characteristics, including sex, with 40.8% of female patients.36 Megaly et al. noted that CEPD use was independently associated with a lower risk for ischaemic stroke (p=0.032).37 Variables associated with a higher risk for ischaemic stroke after TAVR included a history of carotid artery disease, peripheral artery disease, AF or flutter, older age, bicuspid aortic valve and female sex.37
PROTECTED TAVR was a randomised, open-label, multicentre, all-comer trial powered for clinical endpoints: stroke within 72 hours after TAVR or before discharge.39 No significant effect of CEPD on the incidence of periprocedural stroke (CEPD group versus the control group 2.3% versus 2.9%; p=0.30) was reported.39 Although there was a higher percentage of women in the CEPD group than in the control group (42.0% versus 37.8%), the overall results did not support use of CEPD.40
More recently, a meta-analysis combined individual patient data from the two largest randomised studies, PROTECTED TAVR (3,000 patients) and BHF PROTECT-TAVI (7,635 patients), found no significant difference in stroke rates between the cerebral embolic protection and control arms. Secondary analyses, including per-protocol and complier average causal effect adjustments for non-adherence, confirmed these findings.41
The TriGuard™ cerebral protection device (Keystone Heart) was examined in the DEFLECT III trial, and demonstrated fewer DW-MRI-detected ischaemic brain lesions and less cognitive decline in the group with the CEPD.42 Similar findings were observed in the CLEAN-TAVI randomised trial using the Claret Montage Dual Filter System (Claret Medical).43 The meta-analysis of seven randomised controlled trials comparing the outcomes of CEPD versus no-CEPD in patients undergoing TAVR showed that the use of CEPD was not associated with a statistically significant reduction in the risk of clinical, neurological and safety outcomes.13,44 None of the randomised trials or meta-analyses were able to detect sex disparities with use of CEPD. The results of these studies have not demonstrated a significantly lower incidence of periprocedural strokes, and the minor differences in the characteristics of the retrieved debris do not translate into clinically relevant outcomes in women.
Conclusion
Despite the substantial decrease in the reported rate of complications after TAVR in recent randomised trials, sex disparities do persist in early outcomes; however, there is no impact on long-term survival. Furthermore, although female sex has been identified as an independent predictor, there is a paucity of data correlating it with preventive measures. Therefore, therapeutic and management strategies guided by sex, such as the type of THV selected and use of CEPD, cannot be incorporated into current guidelines.