Comparison of Transfemoral versus Transapical Access in Transcatheter Aortic Valve Implantation Using the Sapien XT Prosthesis- A Propensity Score Matched Analysis

Objective: Recently, several multicenter reports suggest inferiority of the transapical (TA) access in transcatheter aortic valve implantation typically following a pronounced TF first strategy. In contrast, we favor TA access in case of borderline vascular anatomy leading to a relatively high TA rate. Subsequently, this might lead to less selection bias and higher surgical expertise. We aimed to compare both modalities by restricting analysis to one valve type to further reduce confounding.

Methods: 289 procedures (TA and TF) using the SapienXT prosthesis were included (2011–2014). To adjust for baseline differences propensity score (PS) 1:1 matching was utilized. A multivariable analysis was conducted to identify independent risk factors for 30-day mortality and paravalvular leakage (PVL).

Results: Univariate baseline analysis TA (n = 179) versus TF (n = 110): age 80 ± 9 versus 82 ± 9 years (p = 0.53), STS-Score 8.1 ± 6.0 versus 6.4 ± 4.0 (p = 0.005).

PS adjustment accounts for clinical baseline values, CT anatomy, known risk factors and for the learning curve, and resulted in 81 matched pairs with excellent bias reduction.

30d mortality rate was 7.4% with TA and 8.6% with TF access (OR 0.86, p = 1). 1-year survival (p = 0.83, and stroke rate (2.5 versus 1.2%, p = 1) were comparable. At discharge, only 4.9% of both groups demonstrated a residual 2 + PVL (p = 1); Pmean (10 ± 5 versus 11 ± 7 mmHg, p = 0.49) as well as LVEF (54 ± 11 versus 56 ± 13 mmHg, p = 0.20) were comparable.

Vascular complications were less frequent in the TA group (3.7 versus 23.5%, p = 0.001). Frequency of bleeding greater than the VARC-minor definition was similar (6.2 versus 11.1%, p = 0.42). Post procedural renal failure defined by AKIN stage > I did not differ between groups (6.2 versus 4.9%, p = 1).

Separate from clinical endpoints, antegrade TA access reduced fluoroscopy time (5.8 ± 2.6 versus 14.7 ± 5.6 min, p < 0.001).

In addition we conducted a multivariable analysis including STS, age, gender, BMI, GFR, LVEF and Agatston score as confounders: TA access could not been identified as an independent risk factor for 30d mortality (OR 1.02, p = 0.96). However, TA access was associated with less likelihood for PVL2+ (OR 0.42, p = 0.005).

Conclusion: In contrast to previous reports, our analysis consisting of less biased groups, do not suggest inferiority of the TA access itself. TA access reduces fluoroscopy time, vascular complication rates and was associated with less PVL. A high-volume TA program with a dedicated team might lead to improved outcomes.