Transcatheter aortic valve implantation (TAVI) has become an alternative to surgery for the treatment of degenerative aortic valve disease.1,2 Transfemoral access (TFA) is the preferred route for TAVI and has been used in approximately 70% to 80% of cases since the first trials demonstrated the safety and efficacy of this procedure in high-risk patients.3,4 Despite the appearance of new-generation valves with smaller delivery devices, alternative access routes were still needed when TFA was not feasible. While the PARTNER I trial (Placement of Aortic Transcatheter Valve Trial) used an exclusively transapical approach in its first cohort (Cohort A),5 the CoreValve trial used both a transaxillary (TFA) and a transaortic approach.6 TFA remains unfeasible in approximately 7% to 10% of patients treated with TAVI,7,8 and TXA would appear to be a good alternative.9 One meta-analysis of outcomes in 10 528 patients treated with TAVI reported comparable in-hospital and 30-day mortality rates for TFA and TXA,8 although the latter route was associated with higher mortality in the mid- and long-term. The authors also detected higher mortality in patients treated via TXA vs other routes (eg, transapical and transaortic). Another meta-analysis comparing TXA and TFA in 4500 patients detected no differences in mortality after 30 days or 12 months.10 The main aim of this study was to compare early mortality rates in unselected patients treated with TAVI via TFA or TXA (figure 1 and figure 2).

Figure 1.

Curves showing 30-day rates for survival free of death for the total population of unmatched patients from the Spanish TAVI registry treated via transaxillary access (TXA) or transfemoral access (TFA). TAVI, transcatheter aortic valve implantation.

(0.1MB).

Figure 2.

Curves showing 30-day rates for survival free of acute myocardial infarction in the total population of unmatched patients from the Spanish TAVI registry treated via transaxillary access (TXA) or transfemoral access (TFA). TAVI, transcatheter aortic valve implantation.

(0.11MB).

The Spanish TAVI registry, an initiative of the Interventional Cardiology Association (ICA) of the Spanish Society of Cardiology, is a prospective registry of data from 45 Spanish centers that perform TAVI.11 Participation in the registry is voluntary. The data are recorded in a dedicated, centralized online database accessible through the ICA website.12 To ensure compliance with rigorous security standards and guarantee the validity of the data, each researcher is assigned a unique username and password to enter patient data. Centers with their own database send data in a specific format through a secure channel. All data are anonymized and each patient is assigned a code in accordance with current data protection regulations. There are no formal audits, but the database is systematically and periodically checked for discrepancies and missing data.

All consecutive patients treated with TAVI added to the registry between 2009 and 2019 were included in this study. Patients who underwent transaortic (n = 56) or transapical TAVI (n = 568) were excluded. Clinical, procedural, and echocardiographic data were collected prospectively from the cardiology and cardiac surgery departments.

Events were defined according to the recommendations of the Valve Academic Research Consortium13 for patients included from 2009 to 2013 and the recommendations of the Valve Academic Research Consortium II14 for patients included from 2014 to 2019. Clinical follow-up data were obtained from all participating hospitals during clinical visits or by telephone. All patients provided signed informed consent before TAVI. The study was approved by the lead ethics committee (21/385-E).

The main findings of this study are that TXA was associated with higher in-hospital and 30-day mortality than TFA in a propensity score-matched population treated with TAVI; TXA was also associated with higher rates of acute myocardial infarction, general kidney complications, and pacemaker implantation.

Our findings largely have implications for patients deemed candidates for TXA and indicate that TXA should only be considered when TFA is not feasible, not as a general alternative. Indeed, the benefits of TAVI in low- and intermediate-risk patients have only been demonstrated for TFA.1,2,15

There are controversial data on the safety of TXA vs TFA. A meta-analysis of 4504 patients found no differences between the approaches for either events or mortality at 12 months.8 Similarly, a multicenter study of 21 611 patients from the French TAVI registry that used prespecified propensity score matching to compare clinical outcomes between patients treated via TFA (19 995, 92.5%) and those treated via a nonfemoral peripheral access (1616, 7.5%) found no differences in mortality and the only difference in events was a higher rate of major vascular complications in the TFA group.16 The differences could be due to aspects of study design and population makeup. The French study, unlike ours, combined results for TXA and transcarotid access and also had a shorter inclusion period.

The authors of a meta-analysis of 10 528 patients who underwent TAVI via different access routes (TXA, TFA, transaortic, and transapical) reported similar early (in-hospital and 30-day) all-cause mortality rates among patients treated via TXA and TFA. There were, however, differences in mid- and long-term outcomes (at 6 months and 3 years), with higher mortality, infarction, and stroke rates observed in the TXA group. These findings are consistent with ours, as in our series, patients treated with TXA-TAVI showed significantly higher myocardial infarction rates and a tendency toward a higher incidence of stroke. Although we matched all the variables that differed between the 2 populations, including coronary artery disease, and used a statistical test to minimize selection bias, we consider that the groups cannot be matched by PVD, as this is related to access route. Nevertheless, 41% of the patients in the TXA group did not have PVD and could therefore have undergone TAVI via TXA or TFA. Accordingly, our results are not invalidated by severe multicollinearity, although PVD was more common in the TXA group (59% vs 10% in the TFA group). This greater prevalence of PVD among patients treated via TXA may explain the higher mortality rates observed, as PVD has been linked to a higher incidence of death in the setting of TAVI.17 Patients with PVD also have greater atherosclerotic burden, which is the main predictor of risk for cardiovascular disease and death in patients with either obstructive or nonobstructive coronary artery disease.18

Greater atherosclerotic burden might also be associated with the higher rates of kidney complications observed in the TXA group. Atherosclerotic burden in the thoracic aorta has been linked to higher kidney failure rates in patients treated with TAVI, and kidney failure is also a predictor of mortality in this setting.19,20 The amount of contrast medium used to position the valve might also be related to the higher rates of kidney failure observed with TXA, as positioning of the valve through this route can be complicated by the angle between the subclavian artery and the aortic arch, particularly in the horizontal aortas and especially when using a right approach. In addition, most valves used for TXA are self-expanding and their correct placement requires more time and control. This could explain the longer procedures in this group and the greater quantity of contrast medium used.

A higher percentage of patients in the TXA group required pacemaker implantation, supporting findings by Takagi et al.8 in their meta-analysis. We believe that the greater need for pacemakers observed in our study and others21–23 is linked more to the type of valve than the access route used. In our series, a self-expanding valve was used in 85% of TXA-TAVI procedures compared with just 55% of TFA procedures.

There is growing a body of evidence supporting the safety and efficacy of fully percutaneous TXA for TAVI. In a study of 100 consecutive patients treated with percutaneous TXA-TAVI, 85% of the procedures were performed via the left subclavian artery, while 15% were performed via the right subclavian artery.24 Device implantation was successful in 95% of cases and there were no major access site complications. The authors highlighted the presence of a clear learning curve attributable to the greater complexity of the access route. This curve, combined with the low use of TXA in Spain (it is used in just 3.3% of TAVI procedures),25 may have influenced the clinical outcomes in our series.

The percutaneous TXA approach is less invasive than surgery and hence is more comparable to TFA. No studies, however, have directly compared the 2 approaches and as such there is no evidence to support the use of TXA when TFA is feasible.

This is the first study to show that TXA was associated with higher in-hospital and 30-day mortality rates than TFA in a propensity score-matched population treated with TAVI before and after adjustment for potential confounders. Our findings also show that TXA is associated with higher rates of acute myocardial infarction, general kidney complications, and pacemaker implantation.

The Spanish TAVI registry is managed and maintained with funding from the Interventional Cardiology Association of the Spanish Cardiology Society.

P. Jiménez-Quevedo contributed to conceiving and designing the study; collecting, analyzing, and interpreting the data; and writing the article and critically reviewing its content. L. Nombela-Franco, E. Muñoz-García, R. del Valle-Fernández, R. Trillo, J.M. de la Torre Hernández, L. Salido, J. Elizaga, S. Ojeda, J. Sánchez Gila, B. García del Blanco, A. Berenguer, G. Lasa-Larraya, C. Urbano Carrillo, A. Albarrán, R. Ruiz-Salmeron, J. Moreu, L. Gheorghe, D. Arzamendi, G. Yanes-Bowden, J. Diaz, I. Pérez-Moreiras, M. Artaiz, B. Vaquerizo, I. Cruz-Gonzalez, V. Ruiz-Quevedo, R. Blanco-Mata, J.A. Baz, M. Villa, Á. Ortiz de Salazar, V. Tascón-Quevedo, S. Casellas, and R. Moreno contributed to collecting the data and critically reviewing the content of this article.

R. Trillo is a proctor for transcatheter valves for Medtronic and Boston Scientific. R. Moreno is a proctor for Boston, Abbott, Biosensors/NVT and has received speaker and consultancy fees from Edwards, Boston, Abbott, and Medtronic and research fees from Edwards, Abbott, Boston, and Medtronic. L. Nombela-Franco is a proctor for Abbott and has participated in talks for and received speaker fees from Edwards Lifesciences. J.M. de la Torre Hernández has received consultancy and presentation fees from Medtronic, Abbott, and Boston Scientific. I. Cruz-Gonzalez is a proctor for Medtronic. D. Arzamendi is a proctor for Abbott. The rest of the authors declare no conflicts of interest.