Transcatheter aortic valve implantation (TAVI) has emerged as the treatment of choice in nonoperable patients with severe symptomatic aortic stenosis and a good alternative to surgery in those considered to be at high surgical risk ((1),(2),3). However, most data on TAVI are limited to acute and 1-year follow-up, and very few data exist on the outcomes beyond 1-year follow-up (1). Several factors have been identified as predictors of poorer early and midterm (1 year) outcomes after TAVI ((4),(5),(6),7), but very few data exist on the factors determining worse outcomes at longer-term follow-up. Furthermore, the proper evaluation of transcatheter valve durability at long-term follow-up is one of the most important factors determining the potential expansion of this technology toward younger and lower-risk patients. Although TAVI is usually associated with excellent hemodynamic results, few data exist on the long-term durability of transcatheter valves ((1),(8),(9),(10),(11),12). Importantly, the echocardiographic data available to date have been mainly obtained from single center studies or multicenter registries without central echocardiography core laboratory evaluation of the echocardiographic findings. The lack of uniformity in echocardiographic measurements might have introduced a bias in the evaluation of transcatheter valve hemodynamic status, including the degree and type of residual aortic regurgitation (AR). Also, the number of patients evaluated at different times over the follow-up period has tended to vary considerably, and this precludes a real paired evaluation of the echocardiographic exams, which in turn might lead to somewhat misleading results on valve durability. The objectives of this study were to evaluate the long-term outcomes after TAVI in the Multicenter Canadian experience study, with special focus on the causes and predictors of late mortality and valve durability.

The Canadian TAVI Multicenter experience study included 339 consecutive patients who underwent TAVI between January 2005 and June 2009 in 6 Canadian centers with the Cribier-Edwards (n = 57), Edwards SAPIEN (n = 275), or SAPIEN XT (n = 7) valve (Edwards Lifesciences, Inc., Irvine, California) in the setting of the Canadian compassionate clinical use program approved by the Department of Health and Welfare (Ottawa, Ontario, Canada). This represented 61% of the patients evaluated by the Heart Teams of the participating centers for a TAVI procedure (the other 21% had surgical aortic valve replacement, and 18% had medical treatment). Patient and approach selection as well as the main procedural characteristics and results were detailed in a previous publication (4). Briefly, patients considered eligible for TAVI underwent a systematic workup protocol that included Doppler echocardiography, coronary angiography, aorto-iliofemoral angiography, and computed tomography. The patients were selected for transfemoral (TF) or transapical (TA) approach, depending on the size, disease, and degree of calcification of iliofemoral arteries. Starting in May 2007, all cases were presented, discussed, and finally approved for TAVI (TF or TA approach) in a weekly conference call including interventional cardiologists and cardiac surgeons of participating centers. Patient comorbidities were defined with the Society of Thoracic Surgeons risk score definitions. Chronic pulmonary obstructive disease (COPD) was defined as the presence of forced expired volume of air in 1 s <75% of predicted and/or the use of chronic inhaled or oral bronchodilator therapy. Chronic kidney disease (CKD) was defined as an estimated glomerular filtration rate (eGFR) <60 ml/min. Pulmonary hypertension (PH) was defined as a pulmonary systolic pressure >60 mm Hg as estimated by Doppler echocardiography or measured by cardiac catheterization. Frailty was defined as a syndrome of decreased reserve and resistance to stressors, resulting from multiple declines across multiple physiologic systems leading to vulnerability to adverse outcomes. No systematic tests were performed for the evaluation of frailty, and patients were considered nonoperable because of frailty mainly on the basis of the criteria of the multidisciplinary team evaluating them. Indeed, at least 2 cardiac surgeons had to agree when frailty was the main criterion determining inoperability, and the specific reasons for the decision had to be detailed during a weekly conference call involving all centers participating in the study. The study population was divided in 2 groups (i.e., early and late) depending on whether patients had been treated in the first- or second-half of the TAVI program experience in each participating center, to further evaluate the potential influence of the learning curve on clinical outcomes.

Clinical follow-up was carried out in clinical visits and/or through phone contact. The timing and frequency of the clinical follow-up was determined by each participating center. Most patients were followed at 1 year after the procedure and annually thereafter. Follow-up data were prospectively gathered in each participating center and were available in all but 3 patients (99.1% of the study population). Death and re-intervention at any time during the follow-up period were recorded. If death occurred it was further classified as of cardiac or noncardiac origin. The complete medical file of the center taking care of the patient was reviewed, and the primary care physician and cardiologist of the patient were consulted if any further information was needed. Any sudden death was considered cardiac unless proven otherwise. Any death occurring within the 30 days after the TAVI procedure was considered as cardiac. Stroke was defined as a neurological deficit lasting >24 h or associated with brain lesions as determined by cerebral imaging exams. Myocardial infarction was defined on the basis of the guidelines for universal definition of myocardial infarction (13).

Doppler echocardiography was systematically performed at baseline and at hospital discharge in all patients. The timing and frequency of echocardiographic exams at follow-up was determined by each center, but most patients had an echocardiography at 1-year follow-up and then yearly. The echocardiographic exams were analyzed at the Echocardiography Core Laboratory of the Quebec Heart and Lung Institute (co-directed by Drs. Philippe Pibarot and Jean G. Dumesnil) by experienced technicians blinded to clinical data and reviewed by a cardiologist (6). Transvalvular gradients and valve effective orifice area measurements were performed following the methods previously described ((14),15). The severity of AR was evaluated with the multiparametric approach proposed in the American Society of Echocardiography/European Association of Echocardiography guidelines ((16),17).

Qualitative variables were expressed as percentages and compared with the chi-square or Fischer exact test. Quantitative variables were expressed as mean ± SD or median (25th to 75th interquartile range) and compared with the Student t test or Wilcoxon rank sum test, depending on variable distribution. A Cox multivariate analysis including all variables with p value <0.10 in the univariate analysis was used to determine the predictive factors of cumulative late mortality and late (>30 days) mortality. Univariate and multivariate competing risks regression analysis were used to determine the predictive factors of cumulative cardiac mortality to account for noncardiac mortality. Survival curves up to 4 years were presented as Kaplan-Meier curves, and the log-rank test was used for comparison between groups. An analysis of variance for repeated measures was performed to test for equal means at different times (baseline, discharge, 12, 24, 36, and 48 months) for mean gradient and valve area values, and post hoc comparisons were performed by Tukey-Kramer procedure. A 2-way analysis of variance for repeated measures with interaction was used to compare the changes in left ventricular (LV) diameters and LV ejection fraction at different time points between groups (mild AR vs. no or trace AR). Mixed effects ordinal regression was performed to test for significant changes in AR over time. The results were considered significant with p values <0.05. All analyses were conducted with the statistical package SAS (version 9.2, SAS Institute, Inc., Cary, North Carolina).

The main baseline and procedural characteristics of the study population are shown in (Table 1).

At a mean follow-up of 42 ± 15 months, a total of 188 patients (55.5%) had died, 36 patients (10.4%) died within the 30 days after the TAVI procedure, and 152 patients (44.8%) died during the follow-up period. The causes of death during the follow-up period are detailed in (Table 2). Of the 152 patients who died during the follow-up period, 90 (59.2%) died because of noncardiac causes, mostly because of pulmonary causes (45.6%) and end-stage kidney disease (13.3%), and 35 patients (23.0%) died of cardiac causes (cardiac failure in 15.1%). Re-intervention was required in 2 patients who had valve endocarditis at 7 and 13 months after TAVI requiring transcatheter valve explantation.

The predictors of mortality throughout the study period on univariate and multivariate analyses are shown in (Table 3). In the multivariate analysis, the independent predictors of cumulative late mortality were COPD (hazard ratio [HR]: 1.84, 95% confidence interval [CI]: 1.35 to 2.51, p = 0.0001), CKD (HR: 1.12 for each decrease of 10 ml/min in eGFR, 95% CI: 1.02 to 1.23, p = 0.024), chronic atrial fibrillation (AF) (HR: 1.39, 95% CI: 1.03 to 1.89, p = 0.031), and frailty (HR: 1.41, 95% CI: 1.02 to 1.96, p = 0.034) . The predictors of cardiac death on univariate and multivariate analyses are shown in (Table 4). In the multivariate analysis, the predictors of cardiac mortality were CKD (HR: 1.17 for each decrease of 10 ml/min in eGFR, 95% CI: 1.05 to 1.33, p = 0.040) and PH (HR: 1.98, 95% CI: 1.19 to 3.27, p = 0.008). There were no differences in overall mortality or cardiac mortality between patients treated in the early (first-half) and late (second-half) phase of the study period (HR: 1.04, 95% CI: 0.77 to 1.40, p = 0.82 for overall mortality; HR: 1.25, 95% CI: 0.78 to 2.07, p = 0.36 for cardiac mortality). Also, the type of valve (Cribier Edwards vs. Edwards SAPIEN, HR: 0.95, 95% CI: 0.66 to 1.38, p = 0.78 for overall mortality; HR: 1.48, 95% CI: 0.86 to 2.54, p = 0.16 for cardiac mortality) had no influence on clinical outcomes. To further analyze the factors associated with late mortality (beyond the peri-procedural period), the predictors of late mortality in those patients who survived the TAVI procedure (no mortality within the first 30 days) were analyzed on univariate and multivariate analyses (Table 5). In the multivariate analysis, the independent predictors of late mortality were COPD (HR: 2.18, 95% CI: 1.53 to 3.11, p < 0.0001), CKD (HR: 1.08 for each decrease of 10 ml/min in eGFR, 95% CI: 1.01 to 1.19, p = 0.009), AF (HR: 1.44, 95% CI: 1.02 to 2.03, p = 0.044) and frailty (HR: 1.52, 95% CI: 1.07 to 2.17, p = 0.021).

The predictive factors of overall mortality in the TF and TA groups are shown in Tables (Table 6) and (Table 7), respectively. A total of 23 patients (6.8%) had a stroke during the study period, 8 patients within the first 30 days (2.4%), and 15 patients (4.4%) during the follow-up period. All strokes were of ischemic origin, and the antithrombotic treatment at the time of stroke was aspirin + clopidogrel (n = 9), aspirin (n = 6), and warfarin (n = 8).

The Kaplan-Meier survival curves up to 4-year follow-up are shown in (Figure 60_gr1).

After TAVI, the mean transvalvular gradient decreased to 10.5 ± 4.1 mm Hg, and aortic valve area increased up to 1.53 ± 0.34 cm² (p < 0.0001 vs. baseline for both). The changes in valve hemodynamic status at 1-, 2-, 3-, and 4-year follow-up are shown in (Figure 60_gr2). A mild nonclinically significant decrease in valve area was observed at 2-year follow-up (p = 0.007), but no further changes in valve hemodynamic status were observed at 3- and 4-year follow-up (p > 0.2). No cases of structural valve failure were observed throughout the study period. Similar results were obtained when the analyses were limited to the Edwards SAPIEN valve (n = 180, n = 107, n = 58, and n = 20 at 1-, 2-, 3-, and 4-year follow-up, respectively) (p = 0.006 for changes in mean gradient and valve area at 2-year follow-up; p > 0.3 for changes in mean gradient and valve area at 3- and 4-year follow-up) (6). Some degree of AR after TAVI (hospital discharge) occurred in 70% of the patients and was mild or moderate in 60% and 10% of the patients, respectively. Residual AR was paravalvular, transvalvular, and both paravalvular and transvalvular in 38%, 14%, and 48% of patients, respectively. The changes in residual AR over time are shown in (Figure 60_gr3). No significant changes (p > 0.05) in the degree of residual AR were observed during the follow-up period and up to 3- and 4-year follow-up. Similar results were obtained when the analysis was limited to the Edwards SAPIEN valve (p > 0.05 for changes in residual AR over time) (6). The changes in LV end-diastolic diameters and LV ejection fraction over time were grouped according to the presence of mild AR (vs. no or trivial AR) at hospital discharge (Figure 60_gr4), to further evaluate the potential deleterious effects of mild residual AR. No significant changes in LV diameters and LV ejection fraction were observed over time in patients with mild AR, and no differences were detected between these patients and those without AR or with trace AR.

A significant improvement (p < 0.0001) in the functional status of the patients as evaluated by New York Heart Association functional class was observed after the procedure, with 49.8%, 39.1%, 10.3%, and 0.8% of the patients in New York Heart Association functional class I, II, III, and IV at last follow-up, respectively (6).

The present study shows that approximately one-half of the nonoperable or high-surgical-risk patients who underwent TAVI with a balloon-expandable valve for the treatment of severe symptomatic aortic stenosis had died at a median follow-up of 3.5 years. Most (4 of 5) patients died during the follow-up period and, especially within the few months after the procedure, of noncardiac comorbidities. Chronic pulmonary obstructive disease, CKD, frailty, and chronic AF were the main predictors of late mortality, whereas PH and CKD were the main factors associated with cardiovascular mortality. A mild nonclinically significant decrease in valve area and increase in gradient was observed 2 years after TAVI, but no further changes were observed up to 4-year follow-up as evaluated in a central echocardiography core laboratory. The occurrence of mild residual AR, mostly paravalvular, was frequent after TAVI but remained stable throughout the study period. Also, the presence of mild AR had no deleterious consequences in LV diameter and function up to 3-year follow-up. Transcatheter aortic valve implantation was associated with a significant improvement in patient functional status at follow-up.

Consistent with the results of this study, large multicenter registries and the PARTNER (Placement of AoRTic TraNscathetER Valve) trial have shown survival rates at 1-year follow-up after TAVI between 70% and 85%, with most deaths occurring within the months after the procedure ((2),(3),(5),6). The recent U.K. multicenter registry showed survival rates of 76% and 73% at 1- and 2-year follow-up, respectively. However, the very low event rate after the first year (3%) might be explained by the fact that only a minority (one-third) of the study population was followed up to 2 years (6). The mortality rate was 43.3% and 33.9% at 2-year follow-up in the nonoperable and high-risk cohorts of the PARTNER trial, respectively ((10),11). Buellesfeld et al. (8) reported a mortality rate of approximately 38% at 2-year follow-up after TAVI with the self-expandable CoreValve system (Medtronic, Minneapolis, Minnesota) in 125 patients, with a 2- to 3-fold increase in all-cause mortality after the first 30 days, mostly in those patients exhibiting a higher risk profile at baseline. Ussia et al. (12) reported a mortality rate of 35% at 3-year follow-up among 181 patients who had TAVI with the CoreValve system. Gurvitch et al. (9) provided single-center data in 70 patients who underwent TAVI and had a minimum follow-up of 3 years (median: 3.7 years). The survival rate at 3-year follow-up was 61%, and most patients died during the follow-up period due to noncardiac comorbidities. However, the relatively low number of patients included in that study precluded identifying the variables associated with poorer long-term outcomes.

Several studies have identified baseline and procedural factors associated with acute and 1- to 2-year follow-up mortality ((4),(5),(6),(7),(),). The present study identified COPD, CKD, frailty and chronic AF as the main predictors of late mortality (beyond 1 year and up to 6 years, mean of 3.5 years) after TAVI. Respiratory failure was the first cause of death during the follow-up period, mainly due to respiratory infections leading to a respiratory decompensation in patients with impaired pulmonary function at baseline. Therefore, it is not surprising that COPD appeared as the most important factor determining late mortality in this study. COPD is a well-known risk factor of late mortality after surgical aortic valve replacement (18), and previous TAVI studies had already identified this factor as a predictor of acute and midterm mortality ((4),(5),19). These results strongly suggest that an accurate evaluation of these patients by pulmonary specialists before accepting them for a TAVI procedure might be highly clinically relevant. Also, future studies will have to better determine the factors associated with worse outcomes among COPD patients undergoing TAVI and to implement measures to reduce acute exacerbations in this high-risk group (20). Several studies have already identified CKD as a major predictor of mortality after TAVI ((4),(5),(6),(12),(),). CKD is a marker of a higher cardiovascular and all-cause mortality in older people ((21),22). The presence of reduced eGFR levels determined a higher mortality rate very early after the TAVI procedure, and this continued throughout the entire study period. A closer follow-up of such patients after the TAVI procedure might be associated with an improvement in TAVI results. Frailty is a very prevalent comorbidity among patients undergoing TAVI today. This factor has been associated with poorer outcomes in cardiovascular and noncardiovascular procedures ((23),24). The reasons why frailty was not found as a marker of worse outcomes after TAVI in previous studies might be multifactorial. First, the diagnosis of frailty has been mainly subjective and not based on objective parameters, and this could have introduced significant variability in the evaluation of this factor between studies. Also, the limited duration of the follow-up might have contributed to the absence of relationship between frailty and poorer outcomes in previous studies. Future studies will have to establish objective parameters to obtain a more accurate evaluation of this important factor and further determine its prognostic value after TAVI. Chronic AF has been identified as a marker of poorer prognosis after cardiac surgery (25). To the best of our knowledge, this is the first study identifying chronic AF as a prognostic factor after TAVI. Interestingly, most patients with chronic AF included in the present study died of noncardiac causes. No increase in stroke rate but a higher incidence of fatal bleeding was observed in this group, partially explaining the greater mortality among chronic AF patients. This highlights the importance of an accurate control of the anticoagulation therapy in these very old patients as well as of the avoidance of combining multiple antithrombotic therapies. Late cardiovascular mortality was mostly determined by PH and CKD. PH has been associated with a higher rate of mortality after cardiac surgery (26), and previous TAVI studies have identified PH as a marker of acute and midterm mortality ((4),6). PH makes the patient both more prone to and more vulnerable in case of peri-procedural hemodynamic complications. Also, the development of PH in the setting of severe AS is a marker of an advanced stage of the disease, and some of these patients will experience either slight or no improvement after the procedure. In such patients, especially those with very severe PH, a staged procedure with balloon aortic valvuloplasty before TAVI for both testing the hemodynamic response and improving patient hemodynamic status might be useful to improve patient selection and procedural results (27).

The analysis of the prognostic factors in each TAVI approach group (TF or TA) showed that the variables associated with poorer outcomes might vary according to the approach. While PH was the main predictor of overall mortality in the TF group, COPD and frailty were the most important factors associated with late mortality in the TA group. Future studies with a larger number of patients will have to determine whether these differences are related to the approach per se or to the different baseline characteristics of the patients selected for each approach. Importantly, the approach had no prognostic influence on acute or late outcomes in this study.

Several studies have shown a significant improvement in functional status and quality of life after TAVI ((2),(3),(4),(5),(6),(7),(8),(9),(10),(11),(12),). The present study shows that the improvement in functional class occurred in most patients after the procedure and was maintained up to a median follow-up of >3 years.

Several studies have evaluated transcatheter valve hemodynamic status up to 1-year follow-up, showing valve function stability and no cases of structural valve failure at midterm follow-up ((2),(3),(4),(5),(6),(7),(8),(9),(10),11). Buellesfeld et al. (8) and Ussia et al. (12) showed the absence of significant changes in valve hemodynamic status up to 2- and 3-year follow-up, respectively, after TAVI with the self-expandable CoreValve system. The 2-year follow-up results of the PARTNER trial showed the stability in valve hemodynamic status after TAVI with the Edwards SAPIEN valve ((10),11). Gurvitch et al. (9) evaluated the valve hemodynamic status after TAVI with a balloon-expandable Cribier-Edwards valve up to 3-year follow-up, showing a mild nonclinically significant increase and decrease in mean transvalvular gradient and aortic valve area, respectively. However, none of these studies except PARTNER evaluated the echocardiographic exams in a central echocardiography core laboratory with standardized protocols, and in all studies the number of patients being evaluated at each time period changed (significantly decreased) over time, precluding a real paired comparison of the echocardiographic measurements at different times. These factors might have led to a difficult or even misleading interpretation of the results with regard to valve durability. The present study exclusively evaluated patients with serial echocardiographic exams up to 4-year follow-up, therefore including the same patients and number of exams at each time. The results showed good stability of the Edwards balloon-expandable valve function up to 4-year follow-up, with only mild nonclinically significant changes in transvalvular gradient and valve area at 2-year follow-up. No significant changes were observed over time in the presence and degree of residual AR. Importantly, mild residual AR (mostly paravalvular) was frequent after TAVI but had no impact on LV diameters and function. These findings are in accordance with the results of previous surgical aortic valve replacement studies ((29),30) and support the lack of clinical relevance of this frequent feature after TAVI.

Although the data were prospectively collected in each of the participating centers, there was no pre-specified case report form designated for the purpose of this study. However, that the cases were presented with a similar presentation format in a weekly conference call ensured data uniformity and partially compensated for this limitation. There was no independent event adjudication committee for this study, but this might have been less relevant, because the primary clinical outcome was the occurrence of mortality (yes/no). Also, this was a multicenter study, but the TAVI experience was very variable among centers. The diagnosis of frailty was based on the subjective evaluation of 2 cardiac surgeons. Although this represents the most usual way of evaluating this comorbidity, it might have introduced some variability in the diagnosis, and further research is needed to obtain a more objective evaluation of this important factor. The strict criteria for echocardiographic evaluation (only echocardiographic exams from patients with complete serial echocardiographic follow-ups were analyzed) led to a decrease in the number of echocardiographic exams available for analysis. However, serial echocardiographic analysis was available in up to 90%, 67%, and 66% of the patients at risk at 1- , 2-, and 3-year follow-up, respectively. Also, the main purpose of this study with respect to the echocardiographic analysis was to evaluate in a central echocardiography core laboratory the changes in valve hemodynamic status over time to provide reliable data on valve durability. The number of patients with serial echocardiographic exams up to 4-year follow-up was limited, and future studies with a larger number of patients are needed to confirm these results.

Although TAVI in patients considered to be nonoperable or at very high surgical risk was associated with good hemodynamic results translating into a significant functional improvement in most patients, the mortality rate at a mean follow-up 3.5 years after the procedure remained relatively high (>50%). Patients with COPD, CKD, frailty, and chronic AF were at higher risk of death within the few years after TAVI, suggesting that a more careful evaluation and follow-up of patients with these comorbidities might translate into better mid- to long-term outcomes. Also, cardiac mortality was mostly related to the presence of PH and CKD at the time of the TAVI procedure. The presence of these entities, especially of PH, might reflect an overly advanced cardiovascular disease condition, and the potential role of staged procedures (balloon aortic valvuloplasty before TAVI) in these patients might merit further evaluation. Finally, the present results showing valve function stability with only minor changes in valve hemodynamic status up to 4-year follow-up as evaluated in a central echocardiography core laboratory provide important insights into the long-term durability of balloon-expandable transcatheter valves. Continuous monitoring of these patients beyond 4-year follow-up is essential to provide definite data on long-term transcatheter valve durability in the near future.