The TAVI National Committee project is a collaboration between the Working Group on Cardiac Catheterization and Interventional Cardiology of the Spanish Society of Cardiology and the Spanish Society of Thoracic-Cardiovascular Surgery. The project has the following objectives: to determine TAVI activity and ascertain the results of these treatments in Spain, develop scientific studies based on the data accumulated in the registry, establish treatment recommendations related to TAVI procedures, and collaborate with similar international registries. The ex officio members of this committee are 3 interventional cardiologists, 3 cardiac surgeons, 1 clinical cardiologist, and a member of the Board of the Working Group on Cardiac Catheterization and Interventional Cardiology, who is in charge of the registry. This committee is renewed every 3 years. There is no specific funding for this registry, although the companies Edwards, Medtronic, and Boston Scientific have donated an unrestricted grant to the Spanish Society of Cardiology. Registry design and data publication are the exclusive responsibility of the members of the TAVI National Committee, which is free from industry body influence. The registry complies with Spanish data protection laws and has been approved by a central ethics board. Center participation in this registry is voluntary.

The individualized data of patients who underwent TAVI procedures from 2010 were collected in this registry through an electronic case report form. A total of 131 preoperative, 31 perioperative, and 76 clinical follow-up variables were collected. Of these, approximately 75% were compulsory variables.

In accordance with the recommendations of the Valve Academic Research Consortium,9 the immediate implant success rate and the combined 30-day safety endpoint (all-cause mortality, major stroke, life-threatening bleeding, myocardial infarction, RIFLE grade 3 acute kidney injury, and repeat procedure for valve-related dysfunction) were evaluated. In addition, all-cause mortality was evaluated at 30 days and at maximum follow-up.

The registry is open to any type of commercial device. To date, only 2 types of valves have been included: the Medtronic CoreValve™ and Edwards SAPIEN™. The latter can be implanted via a femoral, transapical, or transaortic approach, whereas the former can be implanted though a femoral, axillary/subclavian, or transaortic approach. The technical specifications of these devices have previously been described.10

Continuous and categorical variables are presented as mean (standard deviation) and as frequencies and percentages, respectively. Continuous variables were compared with Student's t-test or analysis of variance, whereas categorical variables were compared with a chi-square or Fisher's exact test, as appropriate. A logistic regression model was used to identify predictors of hospital mortality by introducing variables with P<.05 in the univariate analysis (age, peripheral vascular disease, creatinine clearance, EuroSCORE, and peak gradient) and clinically relevant variables (type of valve used and apical approach or not). The backward stepwise method was used. The results are expressed as odds ratios and their 95% confidence intervals (95%CIs). Multivariate analysis for the prediction of mortality during follow-up was performed through Cox regression by including all variables that were significant in the univariate analysis (sex, left ventricular ejection fraction, EuroSCORE, implant device failure, and conversion to surgery or aortic regurgitation after implantation) and those considered clinically relevant (type of valve used and apical approach or not). The results are expressed in hazard ratios and their 95%CIs. The survival curve was obtained with the Kaplan-Meier method. A Breslow test was used to compare survival between groups with the mortality variable during follow-up. The statistical analysis was performed with SPSS 15.0 software for Windows (SPSS Inc.; Chicago, Illinois, United States). To determine the representativeness of the TAVI national registry, the ratio between the numbers of patients included in the registry was compared with the number of valves sold during the recruitment period.

From January 2010 to the end of December 2011, 1416 patients were included in the TAVI registry, representing 80% of the total number of valves (1767) implanted in Spain during that year (according to data obtained from the 2 transcatheter prosthesis companies: Medtronic Ibérica SA, Madrid, and Edwards Lifesciences España SL, Paterna, Valencia). Of these 1416 patients, 806 received an Edwards valve (80% of the total number of valves sold) and 610 received a Medtronic CoreValve (81% of the total sold). By approach and number of implanted valves, a higher percentage of valves were inserted through the transfemoral approach than through the transapical approach: 82% (1114/1359) vs 74% (302/408) (P<.0001). The baseline characteristics of the population as a whole and stratified by approach and type of valve are shown in Table 1. The patients who underwent transapical prosthesis implantation were significantly younger, with a lower percentage of women, and showed a higher frequency of peripheral vascular disease, previous myocardial infarction, and previous percutaneous or surgical revascularization. In general, the EuroSCORE in this group of patients was significantly higher than that in the groups who underwent transfemoral implantation of Edwards or CoreValve devices. Surgical contraindication1 and high risk (defined by a EuroSCORE>152) were clinical indications in about 90% of patients treated with an Edwards valve and in 70% of those treated with a CoreValve (Table 1). The baseline ecocardiograph data are presented in Table 2. There were no clinically relevant differences between the types of valves and approaches, except for a significantly larger annulus size in patients treated with a CoreValve.

Procedure data are presented in Table 3. The procedure was mostly performed in the catheterization laboratory in patients undergoing the transfemoral approach, and was largely equally divided between the cardiovascular operating room and the catheterization laboratory in patients undergoing the transapical approach. A hybrid operating room was only available for 3% of the procedures. For procedures with a transfemoral approach, the puncture method—rather than femoral dissection—was more common (76% of the overall population; 91% of the CoreValve patients vs 59% of Edwards patients; P<.0001). Among patients receiving a CoreValve, the size of the device was 26mm in 61%, 29mm in 39%, and 31mm in 0.3%. For patients receiving an Edwards device, the size was 23 mm in 54%, 26 mm in 43%, 29mm in 3.4%, and 20mm in 0.3%. The procedure time, fluoroscopy time, and contrast volume were significantly higher in the CoreValve procedures.

The in-hospital results are shown in Table 4. The overall immediate success rate of prosthesis implantation was 94%. By approach, the transapical success rate tended to be lower than that of the transfemoral approach (91% vs 94%; P=.09). We found a significantly greater requirement for cardiopulmonary bypass support and conversion to conventional surgery in patients receiving valves implanted via the transapical approach, a higher rate of tamponade for Edwards valves implanted via the transfemoral approach, and a greater need for postdilatation with CoreValve valves. All conversions to surgery (n=20) were to conventional valve replacement surgery, although 3 required additional aortic dissection surgery. The reasons for conversion to surgery were severe residual aortic regurgitation (n=11), poor implantation (n=4), valve migration (n=2), and aortic dissection (n=3). During hospitalization, 114 patients (8%) died, without differences due to types of valve or approach. In general, the complication rate was low. The incidences of stroke, renal failure, and myocardial infarction were similar between the valve and approach types. Notably, there was a significantly greater need for pacemakers with CoreValves, a higher rate of major and minor vascular complications with Edwards valves implanted via the transfemoral approach, and a greater overall presence of residual aortic regurgitation with valves implanted via the transfemoral approach. In addition, the prevalence of life-threatening bleeding was higher with the transapical approach. Overall, the mortality and 30-day combined safety endpoint6 rates, at 9% and 14%, respectively, were similar between the 2 types of valves and approaches.

The mean clinical follow-up duration was 244 days. The reported follow-up time was shorter for valves implanted via the transapical approach than via the transfemoral approach (206 [200] vs 254 [223] days; P=.0001). The overall mortality rate following hospital discharge was 9%, which resulted in an accumulated mortality rate of 16% and which tended to be higher in patients treated via the transapical approach (Figs. 1A and B). From the clinical point of view, the percentages of patients with class III/IV dyspnea and class III/IV angina decreased from 76% and 11% before the procedure to 9% and 0.1%, respectively, during follow-up, without differences between valve or approach type. The changes in the transvalvular gradients on echocardiography during follow-up are shown in Figure 2. There were no significant differences between the types of valve or approach (Table 4).

Figure 1.

A: cumulative survival curves by type of valve. B: cumulative survival curves by type of approach.

(0.23MB).

Figure 2.

Mean aortic gradient (mmHg) at baseline, after the procedure (post), and at the 1-month and 1-year follow-up points for CoreValve and Edwards valves. There were no significant differences between the two types of valves or significant changes in the gradients obtained after the procedure and at the 1-year follow-up. TAVI, transcatheter aortic valve implantation.

(0.05MB).

Age, peripheral vascular disease (according to the definition of the logistic EuroSCORE), conversion to surgery, and aortic stenosis severity were 30-day predictors of mortality (Table 5). At maximum follow-up, the predictors of mortality were found to be peripheral vascular disease, renal failure, atrial fibrillation, ejection fraction, conversion to surgery, and residual aortic regurgitation (Table 6). Preoperative mitral regurgitation was associated with in-hospital (odds ratio=4.12; 95%CI, 1.99-8.5; P=.001) and maximum follow-up (hazard ratio=1.67; 95%CI, 0.94-2.96; P=.09) mortality. However, preoperative mitral regurgitation could not be included in the multivariate analysis because it was recorded in only 63% of patients.

National registry participation is voluntary and is thus subject to inclusion bias. However, this registry can be considered to be representative of the current situation in Spain, because it has obtained data on 80% of the valves implanted during the inclusion period. In addition, data on the success and complications rates are similar to those of other national European registries.11–14,25

We were unable to collect clinically relevant variables (such as mitral regurgitation or chronic obstructive pulmonary disease) from all patients. Accordingly, these variables were omitted from the final multivariate analysis.

The inclusion rate and follow-up duration of patients implanted with valves via the transapical approach were significantly lower than those treated by the transfemoral approach, which could indicate an inclusion bias for these types of valves and approaches. Similarly, a few centers did not include all valves implanted during the registration period, which could also have cause some type of inclusion bias.

The events were reported by the investigators dbut were not independently reviewed. In addition, there is currently no way to perform independent audits of a certain percentage of the patients and consequently we cannot rule out the possibility of a certain degree of under- or over-registration of events. Nonetheless, this study involves the first national registry with individualized data of patients treated with this recent technique.

Finally, the different baseline characteristics could have influenced the clinical course of the patients who were treated with different valves or approaches (Figs. 1A and B). To minimize this potential bias, we included the types of valves and approaches in the multivariate analysis that did not appear in the model as predictors of events.