Previous reports of transcatheter aortic valve implantation (TAVI) have focused on short- and mid-term outcomes ((1),(2),(3),(4),5). A few studies have reported outcomes up to 2 or 3 years ((6),(7),(8),9); however, long-term in vivo durability of transcatheter heart valves (THVs) and long-term clinical outcomes are unknown. The aim of this study was to evaluate clinical and hemodynamic outcomes 5 years after TAVI.

Between January 2005 and March 2007, 111 patients underwent TAVI for the treatment of severe symptomatic aortic stenosis at St Paul's Hospital (Vancouver, British Columbia, Canada) with the balloon-expandable Cribier-Edwards or Edwards SAPIEN valve (Edwards Lifesciences, Irvine, California). All patients provided written informed consent for the procedure. After TAVI, patients were treated with aspirin (lifelong) and clopidogrel (for 3 months). If oral anticoagulation was indicated, patients were treated with warfarin and dual antiplatelet therapy (for 3 months) if tolerated.

These initial 111 patients represent the first in-human experience with transarterial and transapical TAVI in truly inoperable patients, and many steps of the procedure were not standardized at that time. Therefore, peri-procedural mortality was relatively high. Because the focus of this paper was to evaluate 5-year outcome following TAVI, patients with unsuccessful valve implantation and patients who died within 30 days were excluded from analysis, leaving a total of 88 patients who were analyzed (Figure 10_gr1).

Clinical and echocardiographic data were prospectively entered into a dedicated database at baseline, at hospital discharge, at 30 days, and annually. Transthoracic echocardiography was performed before TAVI, at a median of 3 days after TAVI (but before discharge), and yearly thereafter according to the guidelines (10). Left ventricular ejection fraction was calculated using the biplane Simpson method. Severity of regurgitation was graded as none/trivial, mild, moderate, or severe, integrating structural, Doppler, and quantitative parameters according to guidelines (11). Pulmonary hypertension was defined as pulmonary artery systolic pressure >60 mm Hg as estimated by Doppler echocardiography or measured by cardiac catheterization (6). Clinical endpoints and prosthetic valve dysfunction were defined according to the Valve Academic Research Consortium ((12),13). Five-year follow-up was available for 84 of 88 patients (96%).

Continuous variables are described using mean ± SD or medians with first (Q1) and third (Q3) quartiles in cases of skewed distributions. Categorical variables are described by frequencies and percentages. Differences between independent groups were tested using the Wilcoxon rank-sum test and Student t test for continuous variables. In cases in which the samples were paired, the Wilcoxon signed rank or paired Student t test were used. Categorical variables were compared with the chi-square test. Survival rates at 1 to 5 years were estimated and graphed using the Kaplan-Meier method. A Cox regression model was used to estimate hazard ratios (HRs) and 95% CIs of baseline and procedural characteristics on outcome. Variables were included in the multivariable model if they were univariately significant at 0.25. Linear mixed-effects models were fit to assess any change in either the aortic valve area or mean gradient from post-TAVI to 5-year follow-up. Individual growth curves were fit with random effects for both intercept and yearly rate of change and assuming an unstructured covariance structure. Analyses were conducted with SAS version 9.2 (SAS Institute, Inc., Cary, North Carolina) and tested using 2-sided tests at a significance level of 0.05.

A total of 88 patients with a mean age of 83 ± 7 years were analyzed. Baseline characteristics are summarized in (Table 1). Valves implanted were the Cribier-Edwards in 49 (56%) and the SAPIEN (Edwards Lifesciences) in 39 (44%). Access was transfemoral (64 patients [73%]) and later also transapical (24 patients [27%]). Procedural characteristics and outcomes are shown in (Table 2).

Median survival time after TAVI was 3.4 years (95% CI: 2.5 to 4.4), and survival rates at 1 to 5 years were 83%, 74%, 53%, 42%, and 35%, respectively (Figure 10_gr2).(Table 3) shows univariate and multivariable predictors of mortality. In multivariable analysis, the presence of chronic obstructive pulmonary disease (COPD) (HR: 2.17; 95% CI: 1.18 to 3.70) and at least moderate paravalvular regurgitation (PAR) post-TAVI (HR: 2.98; 95% CI: 1.44 to 6.17) were associated with increased risk of death. Median survival time for patients with and without COPD was 2.3 and 3.9 years, respectively. Median survival time for patients with at least moderate PAR and mild/trivial/no PAR was 1.7 and 3.4 years, respectively (Figure 10_gr3).

Overall, the mean aortic valve area increased from 0.62 ± 0.17 cm² before TAVI to 1.67 ± 0.41 cm² after TAVI and was 1.40 ± 0.25 cm² at 5 years. The area decreased on average by 0.06 cm²/year (SE 0.01; p for trend <0.01). Mean aortic valve gradient decreased from 46 ± 18 mm Hg before TAVI to 10 ± 4.5 mm Hg after TAVI and was 11.8 ± 5.7 mm Hg at 5 years. The gradient increased on average by 0.27 mm Hg/year (SE 0.15; p for trend = 0.06) (Figure 10_gr4). There were no differences in gradients and aortic valve areas between the Cribier-Edwards and Edwards SAPIEN valves.

Changes over time in valvular and paravalvular regurgitation are shown in (Figure 10_gr5). After TAVI, 25% of patients had no PAR, 63% had mild PCR, and 11.4% had at least moderate PAR. PAR was at least moderate at 1 to 5 years in 5.7%, 1.2%, 0%, 0%, and 0% of patients, respectively.

Mild transvalvular regurgitation was present in 10.3% of patients after TAVI and 9.1%, 14.3%, 11.9%, 9.5%, and 7.1% at 1 to 5 years, respectively. Up to 4 years, no patient had prosthetic valve failure. At 5 years, 3 patients (3.4% of the total cohort) showed signs of prosthetic valve failure. One patient had moderate transvalvular regurgitation and moderate stenosis (aortic valve area 1.2 cm²; mean gradient 26 mm Hg) 5 years after implantation of a Cribier-Edwards valve. Another patient had moderate regurgitation 5 years after implantation of a Cribier-Edwards valve. The third patient had moderate stenosis (aortic valve area 1.1 cm²; mean gradient 23 mm Hg) 5 years after implantation of an Edwards SAPIEN valve. No patient developed severe stenosis or severe regurgitation.

One patient developed infective endocarditis 5 months after implantation of an Edwards SAPIEN valve, underwent surgical open valve replacement, and remained alive 5 years following TAVI.

At baseline, moderate/severe mitral regurgitation (MR) was present in 42 patients (48%). After TAVI, moderate/severe MR had improved to no/mild MR in 24 patients (57%). At 5 years, moderate/severe MR remained improved in 26% and was unchanged in 16%; 58% of patients had died (Figure 10_gr5). No/mild MR worsened to moderate/severe in 4% of patients after TAVI and did not improve in 2% at the 5-year follow-up.

The rate of patients in New York Heart Association (NYHA) class 3 or 4 was 7%, 8%, 6%, 2%, and 5% at 1 to 5 years, respectively (Figure 10_gr5).

Median ejection fraction was 60% (Q1, Q3: 50%, 65%) before TAVI, 60% (Q1, Q3: 55%, 65%) after TAVI, and 56% (Q1, Q3: 45%, 65%) at the 5-year follow-up. Left ventricular end-diastolic diameter was 51.1 ± 6.8 mm before TAVI, 51.5 ± 7.2 mm after TAVI, and 49.2 ± 7.0 mm at 5 years.

During the observation period, 6 ischemic and 4 hemorrhagic major or fatal strokes occurred ((Figure 10_gr6),Table 4). Cumulative major ischemic and hemorrhagic stroke rates at 5 years were 9.7% and 7.3%, respectively, and 50% were fatal.

Our study demonstrated favorable long-term outcomes after successful TAVI with a 5-year survival rate of 35%. Mean aortic valve area decreased on average by 0.06 cm²/year, and mean transvalvular gradient increased by 0.27 mm Hg/year throughout the 5-year period. Moderate prosthetic THV failure was observed in 3 patients at 5 years, but no patient developed severe transvalvular regurgitation or restenosis. Concomitant MR improved in the majority of patients and remained improved at the 5-year follow-up. More than 90% of patients were in NYHA class 1 or 2 during the whole observation period. Cumulative major ischemic stroke rate at 5 years was 9.7%.

In this patient population with a mean age of 83 years at the time of the TAVI procedure, 5-year survival was 35% and median survival after TAVI was 3.4 years excluding patients with procedural mortality (death during the first 30 days). Median survival in patients without COPD was 3.9 years but only 2.3 years in those with COPD. The other predictor of decreased survival was paravalvular leak. Even in these early cases representing the beginning of the learning curve, PAR was mild or less in 90% of patients. The remaining 10% had moderate PAR, which was associated with reduced survival (HR: 2.98). Moderate PAR has previously been shown to be associated with reduced survival rates with HRs/odds ratios ranging from 1.7 to 3.8 ((8),(14),(15),16). Our study suggested that PAR is an independent risk factor not only for peri-procedural but also for long-term mortality.

Up to the 4-year follow-up, no patient demonstrated more than mild transvalvular regurgitation and no patient had clinically relevant THV restenosis. At 5 years, we observed 3 patients (3.4% of the initial group; 9.7% of patients alive) with moderate prosthetic valve failure. One patient had moderate transvalvular regurgitation, one had moderate stenosis, and one had moderate mixed disease. None of the patients had severe regurgitation or stenosis, and no patient required reoperation or reintervention due to structural valve failure. In the PARTNER (Placement of Aortic Transcatheter Valves) trial, moderate or severe transvalvular aortic regurgitation was present in 0.9% (PARTNER 1A) and 4% (PARTNER 1B) of patients after TAVI and in no patients after surgical aortic valve replacement (SAVR) at the 1-year follow-up. Mild transvalvular regurgitation was frequently present in 62% and 63% of patients after TAVI and 45% of patients after SAVR ((4),5). Reported rates of structural valve deterioration in surgically implanted aortic valve bioprostheses requiring reoperation range widely from 6% to 47% by 12 to 20 years after surgical implantation ((17),(18),(19),20). Larger studies are needed to determine the rate of structural deterioration of THVs.

Previously, our group reported that concomitant moderate or severe MR was associated with increased early but not late mortality in patients undergoing TAVI (21). We also reported that moderate or severe MR improved in 61% of patients after TAVI and remained improved in 55% at the 1-year follow-up. In the present cohort, concomitant moderate or severe MR was present in 48% of patients before TAVI and subsequently improved in 55% after TAVI. At 5 years, MR was improved in 26% of patients and remained unchanged in 16%; 58% of patients had died.

Neurological events have appeared as one of the most controversial complications in TAVI. In the PARTNER trial, the rate of strokes was higher after TAVI than after SAVR, although the rate of major strokes did not differ ((4),(5),22). Cumulative major ischemic stroke rate at 5 years was 9.7%, indicating an annual risk of ischemic strokes of approximately 2% that appeared to persist up to 5 years. At the same time, the cumulative major hemorrhagic stroke rate at 5 years was 7.3%.

Our data represent the experience of a single center rather than outcomes of a clinical trial, and they are self-reported with no external independent data adjudication. The number of patients was small, and multivariable predictors of mortality should be confirmed in a larger sample.

Our study demonstrated favorable outcomes 5 years after successful TAVI with excellent hemodynamics and signs of moderate prosthetic valve failure observed in only 3.4% of patients. No patient developed severe valvular regurgitation or stenosis.