Abstract
Transcatheter aortic valve replacement (TAVR) is a less-invasive alternative to surgical aortic valve replacement (SAVR) for the management of symptomatic severe aortic stenosis in patients with inoperable, high, or intermediate surgical risk. Females undergoing TAVR exhibit distinct physiology and baseline characteristics from their male counterparts. Although rates of procedural complications are higher, medium-term survival appears better for females than males. These sex-based differences are discussed in depth in this review of TAVR in the female population.
Introduction
Surgical aortic valve replacement (SAVR) has been the mainstay of therapy for symptomatic severe aortic stenosis. However, transcatheter aortic valve replacement (TAVR) has emerged as a less-invasive treatment alternative, particularly for patients who are inoperable, high, or intermediate surgical risk.1–5 In contrast to most cardiovascular studies, females represent approximately half of patients undergoing TAVR. Distinctions in baseline characteristics and procedural outcomes have already emerged between male and female patients. These sex-based differences in patients receiving TAVR are discussed in this review.
Differences in Left Ventricular Geometry and Physiology
In aortic stenosis, sex-related variances are evident on a molecular and structural level in the left ventricle. Myocardial tissue biopsies from male and female patients undergoing aortic valve replacement revealed differences in collagen 1, collagen 3, and matrix metalloproteinase-2 gene expression. 6 Villari et al. also showed that endocardial fibrosis and abnormal collagen architecture were more common on biopsies from males than from females. 7 Distinct sex-related patterns in hypertrophy may be driven by differences in the upregulation of estrogen receptors in the setting of aortic stenosis and pressure overload. 8
From older hemodynamic studies to more-recent meta-analyses, females with aortic stenosis have demonstrated higher left ventricular systolic function, peak and mean gradients, and pulmonary artery pressures.9–16 Females more often displayed concentric remodeling or hypertrophy, with relatively thickened or hypertrophied ventricles, compared to their left ventricular dimensions.11–13,15,17 This type of remodeling has been associated with lower end-systolic left ventricular wall stress. In comparison, males were more likely to exhibit eccentric remodeling with increased left ventricular size and greater end-systolic wall stress. 11
In addition, concentric remodeling results in a smaller left ventricular cavity size and thus a lower stroke volume. Indeed, Kodali et al. showed that, despite higher systolic function, females have lower stroke volume than males (59 mL/beat vs. 71.7 mL/beat). 10 Even when indexing stroke volume to body size, females with preserved ejection fraction (EF) have a higher incidence of American Heart Association/American College of Cardiology (AHA/ACC) Guideline Stage D3 paradoxical low-flow (defined as a stroke volume index of <35 mL/m2), low-gradient (PLFLG) aortic stenosis. 18
While PLFLG aortic stenosis has been associated with improved 1-yr survival compared to low-flow aortic stenosis with depressed EF (22.3% vs. 32.3%), 19 mortality was worse when compared to patients with normal flow (i.e., stroke volume index >35 mL/m2). 18 Utilizing data from the Placement of Aortic Transcatheter Valves (PARTNER) I trial, Herrmann et al. showed that low flow, seen in 55% of inoperable and high-risk patients, was associated with higher 2-yr mortality (hazard ratio (HR) = 1.5; 95% confidence interval (CI): 1.25–1.89; p = 0.006) and was an independent predictor of mortality, while EF and gradient were not. 20 Currently 4 groups of patients with aortic stenosis have been described based on flow (low or normal) and gradient (low or high), and outcomes for these groups may indeed differ.21,22 How sex affects flow-gradient patterns and outcomes has yet to be studied.
Differences in Aortic Valve Pathology and Annular Dimensions
Sex-related differences are seen across a spectrum of aortic valve disease. Bicuspid aortic valve is the most common congenital heart defect, and is more often diagnosed in males. Michelena et al. demonstrated that bicuspid aortic valve-related complications such as aortic regurgitation, aortic aneurysm and infective endocarditis were more common in males across a range of patient cohorts. Among those patients with bicuspid aortic valve undergoing SAVR, the indication was more often aortic stenosis in females, versus aortic regurgitation in males. Despite these differences in comorbidities, mortality was not statistically different between females and males with bicuspid valves. However, mortality was higher for both sexes compared to the general population, and this decreased survival was more pronounced in females. Aortic regurgitation, although more common in males, was a predictor of mortality only in females with bicuspid aortic valves. 23
The pathology of trieaflet aortic stenosis also differs between the sexes. For any given severity of aortic stenosis, aortic valves in females have less calcium than males have, even after adjusting for smaller body surface area or aortic annular area in female subjects.24,25 Accordingly Clavel et al. have employed and recommended sex-specific cutoffs for aortic valve calcification, with ≥1274 arbitrary units (AU) in females and ≥2065 AU in males used to identify severe aortic stenosis when discordance arises between small aortic valve area (AVA) and low gradient. 26 In fact, severe absolute aortic valve calcification and calcium density are independent predictors of mortality after aortic stenosis diagnosis. 27
Multiple studies also indicate that females have smaller annular and left ventricular outflow tract (LVOT) dimensions.9,10,14,16,28,29 Buellesfeld et al. demonstrated on computed tomography (CT) scans that females have smaller LVOT areas (374.0 ± 94.2 mm2 vs. 478.2 ± 131 mm2; p < 0.0001) and annular areas (386.9 ± 58.5 mm2 vs. 483.1 ± 75.6 mm2; p < 0.0001), although ascending aorta dimensions were similar between the sexes. 28 Rodés-Cabau et al. showed that females represented 58.2% of the cohort with small annular dimensions (i.e., aortic annulus <18 mm), but only 18.5% with large dimensions (annulus >20 mm). 29 Similarly, Michelena et al. determined that 91% of patients with small LVOT dimensions (LVOT diameter 1.7–1.9 cm) were females. 30 To reduce the over-diagnosis of aortic stenosis in smaller hearts, the American Society of Echocardiography has recommended indexing aortic valve area to body size in the setting of a height <135 cm, body surface area <1.5 m2, or body mass index <22 kg/m2. In these circumstances an indexed AVA <0.6 cm2/m2 can be used to define severe stenosis. 31 Nonetheless a small LVOT and annulus may have a significant effect on the choice of valve size for either surgical or TAVR.
Differences in Pre-Procedural Characteristics
Females and males undergoing TAVR have distinct baseline characteristics. The European System for Cardiac Operative Risk Evaluation (EuroSCORE) risk score, which is used to predict operative mortality, is lower in females, even though female gender is a risk factor in the score.9,14,16 This likely reflects lower incidences of other patient-related factors. Indeed despite older age, females undergoing TAVR have lower rates of cardiovascular disease, including myocardial infarction, impaired left ventricular EF, percutaneous coronary intervention, coronary artery bypass surgery, peripheral vascular disease, and stroke. Females also have a lower incidence of diabetes, hypertension, hyperlipidemia, smoking, and renal or pulmonary disease.9,10,14,16
Transcatheter Aortic Valve Replacement
The TAVR procedure has been well-described in numerous guidelines and reviews.32–34 Transfemoral access is the most common implantation approach, although other methods include trans-aortic, trans-apical, trans-subclavian, trans-axillary, and even trans-caval access. Given that large-diameter catheters are used when implanting transcatheter heart valves (THV), the size of the femoral arteries, typically measured by CT, is the determining factor for deciding appropriate access. Smaller vessel size not only contributes to bleeding complications in females during transfemoral TAVR,10,35,36 but also results in greater use of alternative access routes, which have been associated with greater mortality. 37 In addition smaller sized THVs are typically placed in females due to their smaller annular size. In a meta-analyses by O'Connor et al., > 90% of females received 23 mm and 26 mm TAVRs; in contrast >30% of males received 29-mm and 31-mm valves. 9
Differences in Procedural Complications
Several procedural complications are more common in females undergoing TAVR, including vascular complications. The Valve Academic Research Consortium update defines vascular complications to include aortic or annular rupture, access-related injury, distal embolization, limb ischemia, and percutaneous closure device failure. 38 In the PARTNER I trial vascular complications were higher in females after TAVR, compared to females after SAVR (15% vs. 4.6%; p < 0.01). 39 Data from both the randomized control trial (RCT) and continued access registry of the PARTNER study demonstrated higher vascular complications in females, compared to males receiving TAVR (17.3% vs. 10%; p < 0.001). 10 These vascular outcomes are reflected in other studies and meta-analyses9,14,16,40 and might be due to smaller vascular diameter (compared to sheath dimension) and greater vascular calcification in female subjects.
Females also exhibit higher rates of bleeding than males after TAVR. Data from the PARTNER trial again suggested a sex-based difference, with 10.5% of female subjects suffering from bleeding events, compared to 7.7% of their male counterparts (p = 0.012). 10 This echoes similar outcomes in prior TAVR meta-analyses,9,14 as well as higher rates of bleeding seen in females after percutaneous coronary interventions. 41
Some studies suggest that females suffer from higher rates of stroke. Data from the PARTNER trial demonstrated higher stroke rates in females who underwent TAVR, compared to their counterparts receiving SAVR (6.8% vs. 0.7%; p = 0.02), and a trend toward higher stroke rates when compared to males undergoing TAVR (4.5%; p = 0.07). 39 While O'Connor et al. also demonstrated higher stroke rates in females (4.4% vs. 3.6%; p = 0.029), 9 meta-analyses by Conrotto et al. and Stangl et al. failed to demonstrate a significant sex-related difference.14,16 Growing experience and improved technology have curtailed the incidence of stroke with TAVR. Indeed when comparing the original and newest iterations of transcatheter valves in the PARTNER trials, stroke rates have decreased from 5% to <1%.1,42
Females are also at higher risk for other rare complications. Females accounted for 84% of all coronary obstruction cases collected from 81 centers worldwide from 2007 to 2013. 43 Females are also at higher risk of annular or aortic root rupture, which might be due to greater prosthesis over-sizing, balloon post-dilatation, or sub-annular calcification. 44
Despite higher rates of procedural complications, several important procedural outcomes are better in females, including aortic regurgitation or paravalvular leak (PVL). According to Kodali et al., mortality risk is proportional to severity of regurgitation, with even mild aortic insufficiency associated with increased mortality, 45 although a more-recent study using a more granular grading scheme suggests at least moderate PVL is required to affect survival. 46 Lower rates of moderate and PVL were seen in female subjects in the PARTNER trial (6.0% vs. 14.3%; p < 0.001) 10 as well as across multiple meta-analyses.9,14 Because females tend to have smaller annular sizes, females are more likely to experience over-sizing of their THV, thus decreasing the risk of PVL. Less aortic valve calcification, a known predictor of PVL, 47 may also contribute to lower PVL in females.
When the effective orifice area (EOA) of a prosthetic valve is too small for a given patient's body surface area, prosthesis–patient mismatch (PPM) occurs and has been associated with worse outcomes after SAVR. 48 Pibarot et al. demonstrated that in patients with annular diameters <20 mm, severe PPM (defined as an indexed EOA <0.65 cm2/m2) occurred in 33.7% patients undergoing SAVR versus 19% undergoing TAVR (p = 0.002). 49 Lower rates of PPM may be explained by the presence of sewing rings in SAVRs but not in TAVRs. The surgical prosthesis sewing ring is fixed in size, which may result in a smaller final annular diameter than the native annulus; in comparison the transcatheter heart valve can expand to fill the native annulus. Importantly, rates of PPM were the same for females and males, thus highlighting the importance of indexing valve areas to body size. Furthermore while severe PPM was an independent predictor of mortality in SAVR, it was not in the randomized control or continued access registry cohorts of the PARTNER trial (except in registry patients without aortic insufficiency).
Females also demonstrate greater reverse remodeling of their myocardial hypertrophy after TAVR. Stangl et al. showed that, although left ventricular mass index decreased after TAVR in both sexes, only females had a significantly improved left ventricular systolic function and reduced relative wall thickness. 17 Prevalence of left ventricular hypertrophy decreased more in females (from 86% to 45%) than in males (from 56% to 36%). 6
Differences in Survival
Survival rates for females have been better with TAVR than with SAVR. With SAVR, 30-d and 1-yr mortality rates were higher in females, 50 with one contemporary registry implicating female sex as an independent predictor of 30-d mortality. 51 In the PARTNER trial mortality rates were lower for females who underwent TAVR, compared to SAVR, both at 6 mo (12.2% vs. 25.8%; p < 0.01) and 2 yr (28.2% vs. 38.2%; p = 0.049). 39 This improvement was largely driven by the cohort who had undergone transfemoral TAVR. In contrast males did not demonstrate a survival advantage with TAVR, compared to SAVR. Similarly, 1-yr survival for females who received self-expanding TAVRs was better than for females undergoing SAVR (HR: 0.56; 95% CI: 0.33–0.95).4,52
Although a limited number of studies suggested failed to show a survival advantage53–55 most studies demonstrated improved survival after TAVR for female subjects, compared to their male counterparts. In both the RCT and continued access registries of the PARTNER trial, Kodali et al. showed in post-hoc analyses that mortality was lower in females than in males at 1 yr (19% vs. 25.9%; HR: 0.72; p < 0.001). 10
This survival advantage was confirmed across multiple meta-analyses. From twelve observational studies, Stangl et al. demonstrated survival benefit for females over males both at 30 d (odds ratio [OR]: 0.78; 95% CI: 0.64–0.96) and at follow-up of >3 mo (OR: 0.7; 95% CI: 0.59–0.82); of note this meta-analysis included two of the studies demonstrating decreased or no benefit of TAVR in females. 16 O'Connor et al. showed a survival advantage for females at a median of 387 d (HR: 0.79; p < 0.001) regardless of valve type or route of access. 9 At a similar median follow-up of 365 d, mortality was 24% in females versus 34% in males in the meta-analysis by Conrotto et al. 14
Discussion
There are significant sex-related differences in response to aortic stenosis. Females might have more favorable cellular and hemodynamic physiology that contributes to greater reverse remodeling after TAVR. Females demonstrate lower myocardial expression of collagen and matrix metalloproteinase but show higher expression of estrogen receptors—both of which likely contribute to less myocardial fibrosis.6–8 Females have more concentric hypertrophy, compared to males,13,56 and end-systolic wall stress might be lower in females, particularly in the setting of high relative wall thickness.11,12 Finally females have less calcification of the aortic valve for any given severity of aortic stenosis25–27 and severe aortic valve calcium independently predicts excess mortality. 27
Despite higher rates of procedural complications, females exhibit improved survival, compared to males, after TAVR but not after SAVR. Improved survival in females relative to males in TAVR is likely due to multiple factors, including fewer comorbidities and longer life expectancy. Indeed, males undergoing TAVR had significantly higher baseline rates of myocardial infarction, coronary revascularization, peripheral vascular disease, stroke, diabetes, and renal and pulmonary disorders.9,10,14,16 Recent studies from randomized TAVR trials, however, show that unadjusted and adjusted outcomes favor females who have better 1-yr survival, compared to their male counterparts. 10 Females have smaller body size and annuli that might contribute to the survival benefit with less PPM in TAVR, compared to SAVR,29,51 and less PVL.9,14 All these differences likely contribute to improved survival.
The survival disadvantage in SAVR is less well-understood. The surgical literature is mixed; some studies suggest worse outcomes in females57,58 but not after adjusting for confounders, whereas other studies suggest females have better long-term survival, compared to males.59,60 Recent studies comparing TAVR and SAVR have shed more light on sex-related outcomes with SAVR. Large amounts of registry data show lower baseline rates of cardiovascular disease, non-cardiac disorders, frailty, and Log EuroSCOREs in females who underwent SAVR, relative to females who underwent TAVR. Yet female sex remains an independent predictor in the SAVR population for risk-adjusted 30-d mortality (OR: 2.34; p = 0.043). 51 Post-hoc analyses of randomized trials confirm the sex-related differences in survival in the SAVR patients.10,37
Conclusions
These findings support the continued use of TAVR in females, while highlighting the need to refine transcatheter devices to reduce sex-related procedural complications. Finally, sex-based differences in outcomes for lower risk patient cohorts will require further study.
Footnotes
Author Disclosure Statement
Dr. Jennifer Chen does not have any consultancies, honoraria, or other vested interested to disclose. Dr. Rebecca T. Hahn has Core Lab contracts with Edwards Lifesciences, for which she receives no direct compensation, and she is a speaker for Philips Healthcare, St. Jude Medical, and Boston Scientific.
