Transcatheter aortic valve implantation (TAVI) has emerged as an alternative to surgical aortic valve replacement (SAVR) for patients with symptomatic severe aortic stenosis considered to be at very high or prohibitive operative risk (1- 6). Both the transfemoral (TF) and transapical (TA) approaches have been used for TAVI with the Edwards valve system (Edwards Lifesciences, Inc., Irvine, California) (1- 5). Patients have usually been evaluated for TAVI in the setting of a program including interventional cardiology and cardiac surgery teams, with the selection of 1 approach over the other on the basis of the size and/or disease of iliofemoral arteries. However, most studies have evaluated the results of TAVI separately according to the approach (TF vs. TA) used, and very few single-center series have determined the results of a global TAVI program including both approaches for the treatment of symptomatic severe aortic stenosis (7- 9).

The TAVI technology has been mostly applied in very old patients with a high number of comorbidities (1- 9). Operative risk score calculators (i.e., logistic EuroSCORE, Parsonett score, Society of Thoracic Surgeons Predicted Risk of Mortality [STS-PROM] score) that take into consideration the most important patient comorbidities have been used to determine which patients are at very high or prohibitive surgical risk. However, many elderly patients are deemed inoperable on the basis of comorbidities not included in surgical risk calculators (10). Therefore, a high proportion of patients have been refused for SAVR and have undergone TAVI on the basis of risk factors such as porcelain aorta or frailty (1- 9), neither of which is included in the surgical risk calculators. To date, the safety and efficacy of TAVI procedures have been measured by comparing the procedural and 30-day results with predicted operative mortality as calculated by surgical risk calculators, but no studies have determined the results of TAVI in patients with porcelain aorta or frailty, 2 of the most frequent comorbidities in elderly patients with severe aortic stenosis not included in the surgical risk scores. The objectives of this multicenter study were: 1) to evaluate the acute and midterm follow-up results and prognostic factors of a comprehensive TAVI program including both the TF and TA approaches for the treatment of severe aortic stenosis in patients at very high or prohibitive surgical risk; and 2) to determine the results of this TAVI program in patients deemed inoperable because of either porcelain aorta or frailty.

In 2005, the Canadian TAVI program was approved by the Department of Health and Welfare (Ottawa, Ontario, Canada) for compassionate clinical use in patients with symptomatic severe aortic stenosis considered nonoperable or very high surgical risk candidates. All consecutive patients who underwent TAVI between January 2005 and June 2009 in 6 Canadian centers with the Cribier-Edwards, Edwards-SAPIEN or SAPIEN XT valve (Edwards Lifesciences) in the setting of the Canadian compassionate clinical use program were included. All potential candidates for TAVI were evaluated by a multidisciplinary team composed of interventional cardiologists and cardiac surgeons who determined the eligibility of the patient for TAVI. Patients considered eligible for TAVI underwent a systematic workup protocol that included Doppler echocardiography, coronary angiography, aorto-iliofemoral angiography, and computed tomography. Depending on the size, disease, and degree of calcification of iliofemoral arteries the patients were selected for TF or TA approach. 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. Patients' comorbidities were defined with the STS risk score definitions. 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 (11). 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 medical 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 the weekly conference call. Porcelain aorta was defined as an extensive circumferential calcification of the thoracic aorta as assessed by computed tomography and/or fluoroscopy. Baseline clinical and echocardiography data were prospectively gathered in each participating center. All patients provided written informed consent for the procedures.

The TF and TA procedures were performed as previously described (1- 5). The Edwards valve (Cribier-Edwards, Edwards-SAPIEN, SAPIEN XT) was used in all cases. The 23-mm valve was implanted if the transesophageal echocardiographic measurement of the aortic annulus was between 17 and 21 mm, and the 26-mm valve was implanted if the aortic annulus measured between 22 and 25 mm.

Procedural success was defined as the implantation of a functioning valve within the aortic annulus, without intraprocedural mortality. Major procedural complications included valve embolization, need for a second valve, need for hemodynamic support with balloon counterpulsation or femoral-femoral extracorporeal circulation, conversion to open heart surgery, major access site complications, and life-threatening arrhythmias. Major access site complications were defined as those leading to either severe bleeding requiring blood transfusion, fatal bleeding, or need for surgical or transcatheter repair. In the TA approach, both the occurrence of myocardial tears requiring further surgical repair and accidental damage of a coronary artery during apical repair were also considered major access complications. Major post-procedural (30-day) complications included stroke, myocardial infarction, sepsis, need for hemodialysis, and need for a permanent pacemaker. Doppler-echocardiography was performed at hospital discharge in all patients who survived the procedure. Procedural and 30-day events and echocardiographic data were prospectively recorded in each of the participating centers. After the procedure patients received aspirin (80 mg/day) indefinitely and clopidogrel (75 mg/day) for 3 to 6 months.

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 6 months to 1 year after the procedure and annually thereafter. Death and re-intervention at any time during the follow-up period were recorded.

Qualitative variables were expressed as percentages, and quantitative variables were expressed as mean (SD) or median (25th to 75th interquartile range). Comparison of numerical variables was performed with the Student t test or Wilcoxon rank sum test, depending on variable distribution. The chi-square test or Fischer's exact test was used to compare qualitative variables. Procedural and outcome results of TF and TA approaches are provided but not compared, due to the differences in baseline characteristics between groups. A stepwise logistic regression analysis including all variables with p value <0.2 in the univariate analysis was used to determine the predictive factors of 30-day mortality. A Cox multivariate analysis including all variables with p value <0.2 in the Cox univariate analysis was used to determine the predictive factors of cumulative late mortality. Survival rates up to 2 years were presented as Kaplan-Meier curves, and the log-rank test was used for comparison between patients with porcelain aorta and frailty and the rest of the study population. Differences were considered statistically significant at p values <0.05. The data were analyzed with SAS statistical software version 9.1.3 (SAS Institute, Inc., Cary, North Carolina).

A total of 396 patients were considered potential candidates for TAVI by the multidisciplinary team of each center. Of these, 5 patients were turned down during the weekly conference call (of 291 patients presented since May 2007) and 52 patients were included in the PARTNER (Placement of AoRTic TraNscathetER Valve) trial, leading to a final study population of 339 patients who underwent TAVI under the Canadian compassionate clinical use TAVI program (Figure 1). One-hundred sixty-seven patients (49.6%) were selected for TF approach, and 172 patients (50.7%) were selected for TA approach. Six patients who had had unsuccessful TF procedure underwent a second TA (n = 5) or TF (n = 1) procedure later on, leading to a total of 345 TAVI procedures (168 TF, 49%; 177 TA, 51%). Baseline clinical and echocardiographic characteristics of the study population are shown in (Table 1).

Procedural and 30-day outcomes are shown in (Table 2). The procedure was successful in 322 (93.3%) cases. Procedural, post-procedural, and cumulative 30-day mortality were 1.7% (n = 6), 8.7% (n = 30), and 10.4% (n = 36), respectively. Reasons for unsuccessful procedure were at least 1 of the following: balloon instability during aortic balloon valvuloplasty that precluded valve implantation attempt (n = 1, 0.3%), inability to advance the delivery catheter through iliofemoral arteries (n = 5, 1.4%), major vascular complications (n = 2, 0.6%), inability to cross the native aortic valve (n = 5, 1.4%), valve embolization with no implantation of a second valve (n = 6, 1.7%), and procedural death (n = 6, 1.7%). The reasons leading to procedural death were major vascular complications (n = 2, 0.6%), severe left ventricular dysfunction after valve implantation (n = 2, 0.6%), cardiac perforation (n = 1, 0.3%), and acute severe MR after balloon valvuloplasty (n = 1, 0.3%). Major access site complications were the most frequent procedural complication (45 cases, 13%) and occurred in TF (22 cases, 13.1%) and TA (23 cases, 13.0%) cases. Valve embolization occurred in 7 procedures (2%), and a second valve was implanted in 9 procedures (2.6%) because of valve embolization, valve malposition, and/or severe transvalvular or peri-valvular prosthetic regurgitation. A total of 14 patients (4.1%) needed hemodynamic support with aortic balloon counterpulsation (n = 3, 0.9%) or extracorporeal circulation (n = 10, 2.9%) or both (n = 1, 0.3%), due to severe maintained hypotension or hemodynamic collapse secondary to acute severe left ventricular dysfunction (n = 10, 2.9%), ventricular apical bleeding (n = 3, 0.9%), or cardiac perforation (n = 1, 0.3%).

A total of 30 patients (8.7%) died within the 30 days after TAVI. The causes of death for these patients were: multiorgan failure (n = 6, 1.7%), major bleeding (n = 5, 1.4%), pneumonia/septicemia (n = 4, 1.2%), stroke (n = 2, 0.6%), ventricular arrhythmia (n = 2, 0.6%), congestive heart failure (n = 2, 0.6%), cardiogenic shock (n = 1, 0.3%), sudden unexplained death (n = 2, 0.6%), myocardial infarction (n = 1, 0.3%), late (48 h) ventricular embolization of the valve leading to cardiogenic shock (n = 1, 0.3%), pulmonary embolism (n = 1, 0.3%), peripheral embolism (n = 1, 0.3%), aortic rupture (n = 1, 0.3%), and severe mitral regurgitation (MR) likely secondary to mitral leaflet perforation (n = 1, 0.3%). The clinical and procedural characteristics of the patients who died within 30 days after TAVI compared with those who survived are shown in (Table 3). The predictive factors of cumulative 30-day mortality identified by multivariate analysis were PH (odds ratio [OR]: 2.09, 95% confidence interval [CI]: 1.02 to 4.43, p = 0.048), severe MR (OR: 3.01, 95% CI: 1.09 to 8.24, p = 0.033), and the need for peri-procedural hemodynamic support (OR: 6.84, 95% CI: 2.04 to 22.93, p = 0.002).

Mean aortic gradient and aortic valve area decreased and increased, respectively, from 46 ± 17 mm Hg and 0.63 ± 0.17 cm² at baseline to 10 ± 4 mm Hg and 1.55 ± 0.41 cm² at discharge (p < 0.0001 for both). Most patients (84%) had some degree of residual aortic regurgitation at hospital discharge (trivial or mild: 78%; moderate: 5%; severe: 1%).

Porcelain aorta was present in 61 patients (18%). Baseline and procedural characteristics of patients with porcelain aorta are shown in (Table 4). Patients with porcelain aorta were younger; were more frequently female; exhibited a lower STS-PROM score and creatinine values; and had a lower prevalence of cerebrovascular disease, PH, and severe MR. Approximately one-half of the patients with porcelain aorta underwent TAVI by TF approach. The procedure was successful in 98.4% of the patients, but valve malposition requiring the implantation of a second valve tended to be more frequent in these patients. The stroke and 30-day mortality rate were 1.6% and 11.5%, respectively, with no differences compared with patients without porcelain aorta. Frailty was a comorbidity in 85 patients (25%). Baseline and procedural characteristics of frail patients are shown in (Table 4). Frail patients were older and more frequently women, exhibited a lower weight and body mass index, and were associated with a higher STS-PROM score and a lower rate of prior CABG. Procedural and 30-day mortality rates (2.4% and 8.2%, respectively) were similar to that of the rest of the study population, but frail patients more frequently developed acute renal failure requiring hemodialysis in the post-operative period.

Clinical follow-up was available in all patients at a median of 8 months (25th to 75th interquartile range: 3 to 14 months) after TAVI. A total of 39 patients (11.5%) died during the follow-up period, at a median of 162 days (25th to 75th interquartile range: 72 to 390 days) after the TAVI procedure. Of these, 26 patients (7.7%) died of noncardiac causes (respiratory = 13, hemorrhagic or ischemic stroke = 3, renal failure = 3, cancer = 2, cachexy = 2, amyotrophic lateral sclerosis = 1, pancreatitis = 1, subdural hematoma = 1). The baseline and procedural characteristics of the patients who died during the follow-up period compared with those who survived are shown in (Table 5). The predictors of cumulative late mortality were post-procedural sepsis (hazard ratio [HR]: 3.49, 95% CI: 1.48 to 8.28), need for peri-procedural hemodynamic support (HR: 2.58, 95% CI: 1.11 to 6), PH (HR: 1.88, 95% CI: 1.17 to 3), chronic kidney disease [CKD] (HR: 2.30, 95% CI: 1.38 to 3.84), and chronic obstructive pulmonary disease [COPD] (HR: 1.75, 95% CI: 1.09 to 2.83). (Figure 2) shows the Kaplan-Meier survival curves for the entire study population and for the TF and TA groups. Survival rates at 1- and 2-year follow-up were, respectively, 76% (95% CI: 71% to 82%) and 64% (95% CI: 56% to 74%) for the entire study population; 75% (95% CI: 68% to 82%) and 65% (95% CI: 53% to 75%) for the TF group; and 78% (95% CI: 71% to 85%) and 64% (95% Cl: 52% to 80%) for the TA group. The percentage of patients free of death, MI, or stroke at 1- and 2- year follow-up were, respectively, 72% (95% CI: 67% to 78%) and 60% (95% CI: 53% to 71%) for the entire study population; 73% (95% CI: 65% to 81%) and 63% (95% CI: 52% to 75%) for the TF group, and 72% (95% CI: 65% to 81%) and 59% (95% CI: 47% to 75%) for the TA group. (Figure 3) shows the Kaplan-Meier survival curves for patients with porcelain aorta or frailty as comorbidities. Patients with porcelain aorta tended to have a better survival rate at 1-year follow-up (86%, 95% CI: 77% to 95% vs. 74%, 95% CI: 68% to 80%, p = 0.14). There were no cases of structural valve dysfunction at follow-up, and only 1 patient required SAVR at 8-month follow-up due to endocarditis.

This multicenter study including a large series of patients diagnosed with severe symptomatic aortic stenosis deemed inoperable or at very high surgical risk showed that a global TAVI program including both TF and TA approaches was associated with a 30-day mortality of 10.4% and a cumulative mortality rate of 22.1% after a mean follow-up of nearly 1 year. The characteristics of the population and the size of the catheters determined the greater use of the TA approach, which was undertaken in more than one-half of the patients. The presence of severe PH and severe MR but not the STS-PROM score were predictive factors of 30-day mortality. The procedural variable associated with 30-day mortality was any complication leading to the need for hemodynamic support. Predictive factors of cumulative late mortality were peri-procedural need for hemodynamic support, post-procedural sepsis, PH, COPD, and CKD. The TAVI approach (TF vs. TA) had no prognostic value in acute and late outcomes. Patients diagnosed with porcelain aorta or frailty had similar 30-day and late results compared with the rest of the study population. Porcelain aorta patients were associated with a relatively low risk of peri-procedural stroke (1.6%), although with a wide CI, and those who survived the procedure also tended to have a lower risk of death at 1-year follow-up.

The retrograde TF approach has become the approach of choice for TAVI. The TA approach appeared as a complementary option for those who were noncandidates for the TF approach (3- 5), but only a few single-center studies have evaluated its role in a global TAVI program (7- 9). The present study showed that the TA approach made it possible to treat more than one-half of the potential TAVI candidates, highlighting the relevance of this approach in a TAVI program. According to our results, previous reports from single centers with the 2 approaches showed that 32% to 52% of the patients were treated by TA approach (7- 9). Interestingly, we have previously shown that a global TAVI program allowed the treatment of up to 76% of the patients refused for SAVR (7). More recently, Himbert et al. (8) showed that having the TF and TA program allowed approximately one-half of the patients refused for SAVR to be treated. Reducing the catheter delivery size in the near future will increase the number of TF cases, but a high proportion of patients will still exhibit inappropriate iliofemoral arteries for accommodating ≥18-F catheters.

The 30-day mortality rate of 10.4% observed in the present study was similar to the 8.6% to 11.3% mortality rate reported in previous smaller single-center studies that included both TF and TA approaches (7- 9). The predictive factors of 30-day mortality included baseline characteristics such as severe PH and MR and peri-procedural variables such as the occurrence of any complication leading to the use of hemodynamic support. Malouf et al. (12) identified the presence of severe PH as an important predictive factor of 30-day mortality in patients undergoing SAVR. Our results showed that severe PH also had a prognostic value in patients undergoing TAVI. The presence of PH makes the patient both more prone to and vulnerable during any procedural situation of hemodynamic instability and could also increase the risk of post-operative complications (13). Inoperable patients with symptomatic severe AS associated with severe MR were accepted for TAVI with the objective of improving symptoms and the expectation of reducing the degree of MR by reducing pressure overload and improving left ventricular remodeling (14- 15). However, patients with severe MR had 3 times greater risk of dying within the 30 days after TAVI. The presence of severe MR might increase patient's vulnerability during peri-procedural hemodynamic changes and post-procedural complications, as with severe PH. Further research is needed to optimize the acute results of TAVI in patients with severe PH and/or MR and to determine both the reversibility of these 2 conditions and the improvement of symptoms and quality of life of these patients after TAVI. The need for peri-procedural hemodynamic support increased the risk of death at 30 days approximately 7-fold. The hemodynamic instability leading to the need for hemodynamic support was mainly secondary to either severe ventricular dysfunction after valvuloplasty and/or valve implantation or life-threatening bleeding. Importantly, although the 30-day mortality rate was high in these patients, approximately two-thirds of them survived, highlighting the potential benefits of having an extracorporeal circulation machine and a surgical backup when performing these procedures. Interestingly, the STS-PROM score was unable to identify those patients who would die within the first 30 days, and this was consistent with previous TAVI studies (8- 9). In fact, the vast majority of patients who underwent TAVI would not have been operated on in the past, and new predictive risk scores including specific variables for this particular subset of patients should be developed in the future.

The survival rate of 76% at 1-year follow-up was consistent with the 74% to 78% reported in previous studies including TF and TA approaches (8- 9). Importantly, most deaths occurring during the follow-up period were of noncardiac origin. The predictive factors of cumulative late death in the present study included peri-procedural variables such as the need for hemodynamic support and post-procedural sepsis, a cardiovascular condition such as PH, and noncardiac comorbidities such as COPD and CKD. Chronic obstructive pulmonary disease is a frequent morbidity in elderly patients and 1 of the leading causes of death in this population (16). Grossi et al. (17) already showed the long-term prognostic value of COPD in patients undergoing SAVR. Our findings highlight the relevance of both an accurate diagnostic and prognostic evaluation of the patients' respiratory status before performing TAVI and a close follow-up of COPD patients so that acute exacerbations can be reduced (18). Several studies have shown that CKD is an independent predictor of late mortality after SAVR in elderly patients (17,19- 20), and Webb et al. (9) recently reported that this comorbidity was also a predictor of worse outcomes in patients undergoing TAVI. The mechanisms linking CKD and late mortality in our study population were probably multifactorial (21). Importantly, CKD has been associated not only with higher cardiovascular but also with higher all-cause mortality in older people (22), and this prognostic factor should probably be taken into account when evaluating the potential benefits of the TAVI procedure at midterm follow-up, especially in very old patients with moderate to severe CKD. Finally, the TAVI approach had no prognostic value in the present study. In fact, TF and TA approaches were associated with very similar survival rates at 1-year (TF: 75%, TA: 78%) and 2-year (TF: 65%, TA: 64%) follow-up. Some previous studies suggested worse clinical outcomes associated with the TA approach, but patients undergoing TA-TAVI have been systematically associated with a higher risk profile (3- 5). Himbert et al. (8) showed no prognostic value of either the TF or TA approach, consistent with our results, whereas Webb et al. (9) showed that TA approach was an independent predictor of late mortality. Future studies should further investigate the potential independent role of approach selection on TAVI outcomes.

Surgical aortic valve replacement becomes a high-risk or even a prohibitive procedure in patients in whom the ascending aorta cannot be clamped, because of extensive calcification due to the risk of both cerebral embolism and the impossibility of safe aortic dissection and clamping (23- 24). Approximately 20% (5% to 33%) of the patients undergoing TAVI are diagnosed with porcelain aorta (1- 9), and the present study is the first to evaluate the procedural and late outcomes in this specific high-risk group of patients. Of high clinical relevance, the stroke rate was relatively low (1.6%), with no differences compared with the rest of the study population. However, the need for a second valve tended to be more frequent in this group of patients, probably reflecting a higher incidence of valve malposition due to either difficulty in valve positioning or valve displacement during balloon inflation in the presence of a highly calcified aorta. Although STS-PROM scores were lower, 30-day mortality remained similar to that of the rest of the study population, pointing out the high risk of this group of patients. However, those patients who survived the procedure had a relatively low late mortality, leading to a survival rate of 86% at 1-year follow-up. These results support further research into the role of TAVI for the treatment of patients with porcelain aorta.

Older age and frailty have been among the main reasons for considering a high (up to 33%) proportion of elderly patients diagnosed with severe symptomatic aortic stenosis as inoperable (10). Approximately one-half of the frail patients also had other comorbidities leading to a high STS-PROM score, but those patients with a lower STS-PROM score refused for SAVR because of frailty exhibited similar 30-day mortality rates as those of the rest of the study population, suggesting that frailty remains per se an important risk factor in patients undergoing TAVI. Further research will be of major importance in improving identification and management of frail patients undergoing TAVI as well as in determining the benefits of TAVI in this particular subset of patients.

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, the fact that the cases were presented with a similar presentation format in a weekly conference call ensured data uniformity and partially compensated for this limitation. The diagnosis of porcelain aorta and frailty was subjective and based on the judgment of the physician in charge of each patient. This was partially compensated, because patients were evaluated by at least 2 cardiac surgeons agreeing on the diagnosis and were presented in a weekly conference call in which investigators of other centers also had to agree on the diagnosis. However, further research is needed to provide an objective and reproducible evaluation of these 2 important criteria for inoperability.

The results of this large multicenter series of consecutive patients with severe aortic stenosis at very high or prohibitive surgical risk showed that a TAVI program including both TF and TA approaches was associated with comparable mortality as predicted by surgical risk calculators despite a very high-risk patient profile. Cardiac comorbidities (severe PH, MR), procedural factors (need for hemodynamic support, sepsis), and noncardiac comorbidities (COPD, CKD) determined worse outcomes. Finally, the results of the study suggest that TAVI might be a good alternative for the treatment of patients with porcelain aorta and/or frailty, the 2 most common comorbidities not included in surgical risk score calculators that determined patient's inoperability. The prospective multicenter PARTNER trial, which is a randomized trial and has both arms comparing TAVI with surgical therapy and medical therapy, will further determine the safety and efficacy of TAVI for the treatment of the challenging group of patients with symptomatic severe aortic stenosis at very high or prohibitive surgical risk.

The authors thank Mélanie Côté, MSc, from the Quebec Heart and Lung Institute, for his great work on database management and technical support, and Serge Simard, MSc, for statistical analysis.