Intra‐Annular Versus Supra‐Annular Self‐Expanding Valves for Valve‐in‐Valve TAVR

ABSTRACT

Background

A limitation of Valve‐in‐Valve (VIV) transcatheter aortic valve replacement (TAVR) is patient‐prosthesis mismatch (PPM), which is associated with worse quality of life and heart failure hospitalizations. As such, strategies to avoid PPM are desired. We compared the clinical and hemodynamic results of VIV TAVR with intra‐annular self‐expanding valves (IA SEV, Navitor, Abbott Vascular) versus supra‐annular self‐expanding valves (SA SEV, Evolut, Medtronic).

Aims

To evaluate the hemodynamics and clinical outcomes of self‐expanding valve platforms for VIV TAVR.

Methods

Patients were treated at two sites. The primary endpoint was the 30‐day mean transvalvular gradient. Secondary endpoints included 30‐day effective orifice area (EOA); and key clinical events including major vascular complication, coronary obstruction, 30‐day all‐cause mortality, stroke, need for reintervention, and new requirement of permanent pacemaker. A linear effects model was fit to adjust for factors related to PPM including surgical valve true inner dimension (TID), balloon post‐dilation, body surface area, and etiology of surgical valve failure.

Results

Consecutive patients who underwent VIV TAVR with IA SEV (n = 48) and SA SEV (n = 52) are reported; 42% were women, the mean age was 79.2 ± 6.7 years, and the mean STS Predicted Risk of Mortality was 6.0 ± 4.1%. The mean surgical valve TID was 21.5 ± 1.5 mm. At 30 days, there was no difference in mean transvalvular gradient in IA SEV (10.6 ± 3.6 mmHg) and SA SEV (12.3 ± 6.9, p = 0.44). EOA was slightly larger in IA SEV (1.69 ± 0.6 cm²) than in SA SEV (1.40 ± 0.5 cm², p = 0.04) in the unadjusted analysis. After adjustment, there was no significant effect of IA SEV versus SA SEV (p = 0.28), surgical valve true inner diameter (p = 0.79), or balloon post‐dilation (p = 0.37) on gradient. A subset of IA SEV patients who underwent bioprosthetic valve fracture (n = 5) experienced no annular injury, valve leaflet injury, or valve dysfunction at 30 days.

Conclusions

VIV TAVR using an IA SEV was safe in this case series with no difference in hemodynamics between IA SEV and SA SEV. These findings provide support for prospective evaluations of an IA SEV for VIV TAVR.

1. Introduction

Valve‐in‐Valve (VIV) transcatheter aortic valve replacement (TAVR) is an established alternative to repeat surgical aortic valve replacement (SAVR) in patients with failed bioprosthetic aortic valves who are at high or prohibitive risk of surgical mortality [1, 2, 3]. An important limitation of VIV TAVR is patient‐prosthesis mismatch (PPM) [2, 4]. Severe PPM after VIV TAVR is associated with worse quality of life and ongoing heart failure sequelae [5]. As such, strategies to avoid PPM and optimize hemodynamics in patients who undergo VIV TAVR are desired [6].

Outcomes of VIV TAVR with Sapien (Edwards Lifesciences, Irvine, CA), an intra‐annular (IA) balloon‐expandable transcatheter heart valve (THV), and Evolut (Medtronic Inc.), a supra‐annular (SA) self‐expanding THV, have been reported extensively [2, 7, 8, 9, 10]. Matched analyses suggest that self‐expanding TAVR with Evolut achieves larger effective orifice area (EOA) and lower gradients than VIV TAVR with balloon‐expandable valves [8, 10], a finding that reflects hemodynamics achieved in native small annuli with these valves [11, 12]. There are data suggesting that the Portico THV, a self‐expanding THV with IA leaflets, achieves similar EOA and gradients compared to Evolut, and favorable EOA and gradients compared with Sapien, in native annuli [13, 14]. However, there are limited data on VIV TAVR for Portico, and there are no data for the third‐generation Navitor valve design (Abbott Vascular) in VIV TAVR. To define the hemodynamic performance of the Navitor THV in VIV TAVR, we report a case series comparing hemodynamic and clinical outcomes between the Navitor THV and the Evolut THV.

2. Methods

All patients were treated at two high‐volume TAVR sites: the University of Pittsburgh Medical Center (Pittsburgh, PA) and the University of Virginia Health (Charlottesville, VA). Typical TAVR volume at the sites is between 250 and 500 patients per year. Patients with a history of SAVR who developed severe bioprosthetic aortic valve dysfunction were evaluated by the local heart team and deemed suitable for VIV TAVR. Only stented surgical bioprostheses were included; patients with surgical homograft or surgical aortic root replacement were excluded. A waiver of informed consent was approved by the Institutional Review Board at the University of Virginia and the University of Pittsburgh Medical Center. Baseline characteristics, procedural characteristics, echocardiogram data, and clinical outcomes were obtained from the local electronic medical record and the local Transcatheter Valve Therapeutics (TVT) database. Echocardiographic outcomes are reported according to interpretations by site cardiologists board‐certified in the interpretation of echocardiograms. Consecutive VIV TAVR cases using IA self‐expanding valves (SEV)—the Navitor THV—were collected contemporaneously with a comparator group of SA SEV—specifically the Evolut (Medtronic, Inc.) THV. Use of Navitor THV versus Evolut THV and the choice to perform cerebral embolic protection or leaflet modification were at the discretion of operators. Post‐procedural management, including the use of oral anticoagulation, was also at the discretion of treating physicians. The study period included cases from both sites from November 2022 to April 2025.

2.1. Outcomes

The primary endpoint of this study was the 30‐day mean transvalvular gradient. Secondary outcomes included 30‐day EOA, as well as key clinical events including major vascular complications, coronary obstruction, 30‐day all‐cause mortality, stroke, need for reintervention, and new requirement of a permanent pacemaker. A small group of patients underwent bioprosthetic valve fracture (BVF) during the index VIV TAVR procedure; we also report the specific procedural characteristics of this group of patients who underwent VIV TAVR with an IA SEV.

2.2. Statistics

Continuous variables are expressed as means plus or minus standard deviations, and comparisons are made with the Wilcoxon rank sum test. Categorical variables are expressed as percentages and were analyzed via Chi‐square testing or Fisher's exact test, if the expected values were less than 5 for any given variable. A linear fixed effects model was utilized to assess whether the type of valve (IA SEV vs. SA SEV) contributed to varying changes in mean gradient or EOA at 30 days while controlling for additional covariates. Covariates were included in the model based on their known clinical effect on valve hemodynamics, including surgical valve true inner diameter (TID), body surface area (BSA), the use of balloon post‐dilation after VIV TAVR, and etiology of bioprosthetic valve failure (aortic stenosis, aortic insufficiency, or mixed aortic stenosis and insufficiency). Given the presence of repeated measures, robust standard errors were calculated utilizing the CR2 method from the R‐studio package (ClubSandwich). An interaction term between the type of valve utilized and the change in the outcome variable was employed to determine if valve type affected the rate of change in the outcome variable. Collinearity of included covariates was checked via variance inflation factor with < 3 as the cutoff. Marginal means were calculated from the linear fixed effects model to determine the adjusted mean gradient and EOA at both preoperative measurement and 30 days depending on the type of valve device used. Statistical analysis was performed using R Software version 4.4.3 (Posit, Boston, MA); any p < 0.05 was considered statistically significant.

3. Results

The IA (intra‐annular self‐expanding valves [IA SEV]) group consisted of 48 patients who underwent VIV TAVR with a Navitor THV, and the SA (supra‐annular self‐expanding valves [SA SEV]) group consisted of 52 patients who underwent VIV TAVR with an Evolut THV, for an overall population of 100 patients. Baseline characteristics are shown in Table 1. The groups are well‐balanced; there are no significant differences in any major characteristics. Of the overall study population, 42% were women; the mean age was 79.2 ± 6.7 years; and the mean STS Predicted Risk of Mortality (STS PROM) for redo SAVR was 6.0 ± 4.1%. Left ventricular ejection fraction was 50.8 ± 11.6%. The mean labeled surgical valve diameter was 23.5 ± 1.7 mm corresponding to a mean TID of 21.5 ± 1.7 mm, with no significant difference between groups (p = 0.31 and p = 0.25, respectively). Surgical valve and TAVR characteristics for each case are included as Table S1. The baseline mean gradient was 36.7 ± 18.2 mmHg with a mean EOA of 1.02 ± 0.6 cm² (p = 0.64 and p = 0.69, respectively). Aortic stenosis (52.0%) was the most common etiology of bioprosthetic valve failure, followed by aortic insufficiency (29.0%) and mixed etiology (19.0%). Of the surgical valves treated with VIV TAVR, 56.3% (27/48) in the IA SEV group and 48.1% (25/52) in the SA SEV group were not fracturable (p = 0.41).

Table 1.

Baseline characteristics.

	Overall (N = 100)	Intra‐annular (N = 48)	Supra‐annular (N = 52)	p value
Age (years)	79.2 ± 6.7	78.2 ± 7.4	80.0 ± 5.9	0.29
Female (%)	42.0 (42)	39.6% (19)	44.2% (23)	0.79
Body surface area (m2)	2.0 ± 0.3	2.0 ± 0.3	2.0 ± 0.3	0.91
STS PROM (%)	6.0 ± 4.1	6.1 ± 4.2	5.8 ± 4.1	0.87
LVEF (%)	50.8 ± 11.6	49.5 ± 12.1	51.9 ± 11.0	0.26
Mean surgical valve diameter (mm)	23.5 ± 1.7	23.8 ± 1.7	23.6 ± 1.7	0.31
Mean true inner diameter (mm)	21.5 ± 1.7	21.8 ± 1.8	21.6 ± 1.7	0.25
Age of surgical valve (years)	11.6 ± 4.0	11.7 ± 4.0	11.5 ± 4.2	0.71
Surgical valve failure mode (%)
Aortic stenosis	52.0 (52)	56.3 (27)	48.1 (25)	0.54
Aortic insufficiency	29.0 (29)	27.1 (13)	30.8 (16)	0.85
Mixed etiology	19.0 (19)	16.7 (8)	21.2 (11)	0.75
Baseline hemodynamics
Mean gradient (mmHg)	36.7 ± 18.2	35.7 ± 18.2	36.9 ± 18.4	0.64
EOA (cm2)	1.02 ± 0.6	1.07 ± 0.6	1.00 ± 0.5	0.69

Abbreviations: EOA = effective orifice area, LVEF = left ventricular ejection fraction, STS PROM = Society of Thoracic Surgeons predicted risk of mortality score, THV = transcatheter heart valve.

Procedural characteristics and 30‐day unadjusted outcomes are shown in Table 2. The mean THV diameter was 25.6 ± 1.9 mm in the IA SEV group and 25.8 ± 2.1 mm in the SA SEV group (p = 0.81). There were similar rates of balloon post‐dilation (48.0% overall) in the IA SEV (47.9%) and SA SEV (48.1%) groups (p > 0.99). BVF was performed in 10.4% of fractures in IA SEV cases and 11.5% in SA SEV cases (p > 0.99). Leaflet modification, access route, and incidence of vascular access complications were all balanced (Table 2).

Table 2.

Procedural characteristics and 30‐day unadjusted outcomes.

	Overall (N = 100)	Intra‐annular (N = 48)	Supra‐annular (N = 52)	p value
Mean THV diameter (mm)	25.7 ± 2.2	25.6 ± 1.9	25.8 ± 2.1	0.81
Balloon post‐dilation (%)	48.0 (48)	47.9 (23)	48.1 (25)	> 0.99
Bioprosthetic valve fracture (%)	11.0 (11)	10.4 (5)	11.5 (6)	> 0.99
Leaflet modification	18.0 (18)	14.6 (7)	21.2 (11)	0.59
Access (%)
Femoral	93.0 (93)	87.5 (42)	98.1 (51)	0.79
Carotid	6.0 (6)	10.4 (5)	1.9 (1)	0.10
Subclavian	1.0 (1)	2.1 (1)	0.0 (0)	0.48
Access complication	1.0 (1)	2.1 (1)	0.0 (0)	0.48
30‐day outcomes
Mean gradient (mmHg)	11.5 ± 5.6	10.6 ± 3.6	12.3 ± 6.9	0.44
EOA (cm2)	1.58 ± 0.5	1.69 ± 0.6	1.40 ± 0.5	0.04
Residual mean gradient > 20 mmHg	6.0 (6)	4.2 (2)	7.7 (4)	0.15
Mortality	1.0 (1)	2.1 (1)	0.0 (0)	0.48
Stroke	0.0 (0)	0.0 (0)	0.0 (0)	> 0.99
LVEF (%)	52.7 ± 11.6	52.2 ± 11.4	53.0 ± 11.9	0.44
Paravalvular leak (%)
None	75.0 (75)	70.8 (34)	78.8 (41)	0.36
Mild	15.0 (15)	27.1 (13)	21.2 (11)	0.48
Moderate or severe	1.0 (1)	2.1 (1)	0.0 (0)	0.48

Unadjusted outcomes at 30 days are also shown in Table 2. Mean gradient at 30 days (11.5 ± 5.6 mmHg overall) was similar in IA SEV (10.6 ± 3.6 mmHg) and SA SEV (12.3 ± 6.9, p = 0.44; Figure 1). EOA was larger in the IA SEV (1.69 ± 0.6 cm²) group than in the SA SEV group (1.40 ± 0.5 cm², p = 0.04; Figure 2). Residual mean gradient > 20 mmHg was balanced: 4.2% in IA SEV versus 7.7% in SA SEV (p = 0.15). The most common category of paravalvular leak was none (70.8% IA SEV vs. 78.8% SA SEV, p = 0.36). There was one case of moderate PVL in the IA SEV group at 30 days versus none in IA SEV (p = 0.48); and there was no difference in mild PVL at 30 days (p = 0.48).

Figure 1.

Transvalvular mean gradient at baseline and 30‐days in the intra‐annular SEV and supra‐annular SEV groups. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 2.

Effective orifice area (EOA) at baseline and 30‐days in the intra‐annular SEV and supra‐annular SEV groups. [Color figure can be viewed at wileyonlinelibrary.com]

Predictors of the 30‐day mean gradient in the linear effects multivariable model are shown in Table 3. The only significant predictor of 30‐day change in mean gradient as compared with baseline was aortic insufficiency as the primary mechanism of surgical valve failure, as compared with aortic stenosis (estimate of change in mean gradient −26.6 mmHg; 95% CI −31.5 to –21.6 mmHg; p < 0.001). There was no significant effect of IA SEV versus SA SEV (p = 0.28), surgical valve TID (p = 0.79), BSA (p = 0.25), or balloon post‐dilation (p = 0.37; Table 3). Similarly, the only predictor of 30‐day EOA was aortic insufficiency (p < 0.001; Table 4), and all other variables were not significant predictors of change (Table 4).

Table 3.

Predictors of 30‐day mean gradient.

Predictor	Estimate [95% CI]	Standard error	p value
Intra‐annular SEV (vs. supra‐annular SEV)	–3.04 [–8.5 to 2.4]	2.79	0.28
Surgical valve TID (mm)	–0.17 [–1.4 to 1.1]	0.62	0.79
Body surface area (m2)	4.37 [–2.9 to –11.6]	3.72	0.25
Etiology aortic insufficiency (vs. AS)	–26.6 [–31.5 to –21.6]	2.54	< 0.001
Balloon post‐dilation	1.56 [–1.9 to 5.0]	1.75	0.37

Table 4.

Predictors of 30‐day effective orifice area.

Predictor	Estimate [95% CI]	Standard error	p value
Intra‐annular SEV (vs. supra‐annular SEV)	0.12 [–0.01 to 0.26]	0.07	0.07
Surgical valve TID (mm)	0.04 [–0.02 to –0.09]	0.03	0.18
Body surface area (m2)	0.10 [–0.15 to 0.34]	0.12	0.42
Etiology aortic insufficiency (vs. AS)	0.87 [0.66 to 1.08]	0.11	< 0.001
Balloon post‐dilation	–0.06 [–0.21 to 0.08]	0.07	0.38

Characteristics of BVF are shown in Table 5. In five cases in the IA SEV group, BVF was performed; in three cases, BVF followed VIV TAVR deployment (Table 5, Figure 3, Videos S1–S3); and in two cases, with BVF preceded VIV TAVR deployment. In all BVF cases, the post‐dilation balloon was sized 3 mm greater than the surgical valve TID (Table 5). In cases in which BVF was performed after VIV TAVR, the final mean gradient was reduced in comparison with the gradient achieved after VIV TAVR alone (Table 5). There was no incidence of leaflet injury, annular injury, or bioprosthetic valve dysfunction at 30 days in the cases in which BVF followed VIV TAVR deployment. Surgical valve fracture thresholds are reported in Table 5 (Central illustration 1).

Table 5.

Procedural characteristics of bioprosthetic valve fracture in the intra‐annular self‐expanding group.

BVF case	Surgical valve	TID (mm)	Prosthesis age (years)	Baseline MG (mmHg)	Baseline EOA (cm2)	BVF first or VIV first	MG after VIVa	BVF balloon	Fracture threshold (atm)	Final MG	Final EOA (cm2)
1	21 mm Magna	19	19	50	0.8	BVF	n/a	22 mm True	18	6	1.7
2	21 mm Mitroflow	18	13	26	0.9	BVF	n/a	22 mm True	11	9	1.9
3	25 mm Magna	23	9	40	0.6	VIV	13	26 mm True	18	9	2.2
4	23 mm Magna	21	17	35	0.7	VIV	16	24 mm True	18	8	2.1
5	21 mm Magna	19	15	52	0.7	VIV	19	22 mm True	20	4	1.2

Abbreviations: Atm = atmospheres, BVF = bioprosthetic valve fracture, EOA = effective orifice area, LVEF = left ventricular ejection fraction, MG = mean gradient, THV = transcatheter heart valve, TID = true inner diameter, VIV = valve‐in‐valve.

^a MG after VIV TAVR reflects invasive gradients measured intra‐procedural.

Figure 3.

Fluoroscopy images before (A) and immediately after (B) bioprosthetic valve fracture (BVF) to optimize hemodynamics after VIV TAVR. Note the release of the valve waist (arrow) after fracture.

Central Illustration 1.

Mean gradient at baseline and 30‐days in the intra‐annular SEV and supra‐annular SEV groups. [Color figure can be viewed at wileyonlinelibrary.com]

4. Discussion

In this study, we report several important observations. First, VIV TAVR using an IA self‐expanding prosthesis was safe and effective: procedural success was excellent and procedural complications and 30‐day clinical events were low. Second, in adjusted analyses, there was no significant difference in hemodynamic valve performance at 30 days between IA and SA self‐expanding THVs. IA SEV was not associated with a higher mean gradient or lower orifice area after adjustment for multiple factors. Finally, we report the world's first cases of BVF to facilitate VIV TAVR in Navitor THV. In all cases of BVF, there was no leaflet injury, annular injury, or early dysfunction of the new THV.

ViV TAVR has gained widespread adoption as an alternative to reoperation in patients at high risk of mortality [3]; however, there are important limitations, including a high incidence of PPM [2]. Historically, the use of SA SEV (namely, the Evolut family of THVs) has been associated with less PPM and lower gradients after VIV‐TAVR [2, 15]. The design feature most often emphasized is the “supra‐annular” leaflet design of Evolut valves, which has been widely accepted as the mechanism of larger EOA and lower gradients in VIV TAVR with Evolut as compared with Sapien [11, 16]. Although “supra‐annular” and “self‐expanding” are used as interchangeable terms—and conversely, “intra‐annular” and “balloon‐expandable” have come to refer exclusively to Sapien valves—newer generations of THV disrupt this conceptual framework. It is important that the structural heart community understands the clinical performance of newer THVs, especially in challenging clinical scenarios like VIV TAVR.

Our analysis suggests that Navitor and Evolut achieve similar EOA and hemodynamics in VIV TAVR as has been demonstrated in native small annuli [13, 14]. In other words, the “intra‐annular” construction of Navitor does not affect the hemodynamic performance, given the advantages of the valve design itself. What's more, there may be other advantages of Navitor, specifically in VIV TAVR. The IA leaflet design, with larger cells, may allow for a lower risk of coronary obstruction, ease of coronary reaccess, and favorable anatomy for a future valve‐in‐valve‐in‐valve TAVR—if that proves necessary after VIV degeneration. These are promising possibilities but will need to be evaluated with expanded clinical experience, longer‐term clinical follow‐up, and ultimately clinical trials of Navitor in VIV TAVR.

Finally, this initial experience performing BVF in Navitor THV is an important addition to the current knowledge base. BVF and BVR (bioprosthetic valve remodeling) in surgical valves which can be remodeled or fractured has been shown to improve the procedural results of VIV TAVR alone [17, 18, 19]. In cases in which BVF or BVR follows VIV TAVR, the stepwise reduction in invasively measured gradient—from just after the THV is deployed, to the final gradient measured after BVF/BVR—shows that BVF and BVR independently improve upon the hemodynamic result that can be achieved with stand‐alone VIV TAVR, since patients serve as their own controls [9, 18, 20]. The concept that Navitor THV can withstand high‐pressure balloon inflation and maintain valve function without injury to the leaflets has been previously demonstrated on bench testing in Portico valves [21]. In the present study, we reinforce this concept in the world's first reported clinical cases of BVF in Navitor THV. Surgical valve fracture thresholds were consistent with previously published reports of BVF [17, 18]. It is important to emphasize that although our experience is promising, and BVF is safe and effective in large clinical series with a low incidence of valve injury [9, 22], there remains the possibility that any THV leaflets can be injured with high‐pressure inflation, which may result in acute, severe aortic insufficiency and hemodynamic instability [23]. It is also important to caution that the long‐term impact of high‐pressure inflation on the new THV leaflets is not known. We urge operators to exercise clinical judgment when BVF and BVR are considered. Careful consideration of root anatomy, risk of coronary obstruction, LVOT dimensions, and the presence of annular and LVOT calcium is critical whenever BVF or BVR are planned.

There are several limitations to this analysis. First, these cases were performed in two centers, and generalizability is not certain. Second, larger case series and/or clinical trials are needed to further demonstrate the success of Navitor in VIV TAVR. Finally, long‐term follow‐up is not available, and further data are needed to ensure the durability of Navitor THV in cases of VIV TAVR, with or without BVF/BVR.

5. Conclusions

In summary, VIV TAVR using an IA self‐expanding prosthesis was safe and effective in a large case series. There was no difference in hemodynamics between IA and SA self‐expanding THV after multivariable adjustment. BVF in Navitor THV was safe in a small number of cases; there was no incidence of leaflet injury, annular injury, or dysfunction of Navitor THV subjected to high‐pressure balloon dilation. These data provide support for ongoing prospective evaluations of the Navitor THV for VIV TAVR.

Funding

The authors received no specific funding for this work.

Conflicts of Interest

Dr. John T. Saxon is a proctor for Edwards Lifesciences, Medtronic Inc., and Abbott Vascular, and has received consulting fees from Edwards Lifesciences, Medtronic Inc., Johnson & Johnson, and W.L. Gore. Dr. Dustin Kliner has received consulting fees from Medtronic Inc. and Abbott Vascular. Dr. Keith B. Allen reports that Edwards, Medtronic, and Abbott provide institutional research support, consulting, and proctor fees. Dr. Adnan K. Chhatriwalla reports that Abbott Vascular—consultant, speakers bureau; Boston Scientific—research grant; Edwards Lifesciences—consultant, speakers bureau; Medtronic Inc—consultant, speakers bureau. Dr. Rishi Puri reports consulting fees from Abbott Vascular, Medtronic Inc., and Anteris. The other authors declare no conflicts of interest.

Supporting information

Supporting Table S1: Surgical valve and TAVR characteristics by case.

BVF_1.

BVF_2.

BVF_3.