Transcatheter Aortic Valve Replacement

Comparative Intermediate-Term Outcomes of Subclavian and Transcaval Access for Transcatheter Aortic Valve Replacement

Ashleigh Long, MD, PhD and Paul Mahoney, MD

Ashleigh Long, MD, PhD and Paul Mahoney, MD

Abstract: Background. Multiple alternative access routes have been employed for patients with contraindications to standard transfemoral transcatheter aortic valve replacement (TAVR); however, the optimal route for alternative access approaches is not established. In order to better understand possible differences in alternative access routes, we compared the procedural efficacy and outcomes at 30 days and 1 year in patients who underwent TAVR via subclavian (SC) or transcaval (TC) access route at a single, tertiary-care center. Methods. This retrospective analysis included all TAVR procedures performed via SC or TC approaches between December 2011 to January 2020, with outcomes reported to 1 year post procedure. Additional safety and feasibility studies, including successful device deployment, procedural time, blood loss, and total hospitalization length, are included as part of this study. Results. A total of 41 patients underwent SC access and 22 patients underwent TC access for TAVR. Between both cohorts, SC patients were older at the time of TAVR (83.2 ± 3.7 years for SC vs 80.7 ± 3.9 years for TC; P=.03) and all patients were previously deemed high or prohibitive surgical risk (Society of Thoracic Surgeons score, aortic valve replacement only: 10.4 ± 2.6% for SC vs 9.0 ± 1.9% for TC; P=.12), with similar preoperative hemodynamic profiles. Procedural safety. Device deployment was successful in all patients in both groups, with longer procedural times noted in the SC cohort (62.1 ± 12.1 minutes for SC vs 39.8 ± 12.5 minutes for TC; P<.05). There were no in-hospital deaths in the SC group and 1 intraoperative death in the TC group that was unrelated to access route. Average length of hospital stay was consistent between the two groups (3.8 ± 1.4 days for SC vs 3.4 ± 1.1 days for TC; P=.06). More cerebrovascular accidents were noted in the SC group at 30 days (6 for SC vs 1 for TC), 6 months (3 for SC vs 0 for TC), and 1 year (2 for SC vs 0 for TC), with more postprocedural permanent pacemakers implanted in the SC group at 30 days (9 for SC vs 3 for TC; P<.05, but with fewer at 6 months (2 for SC vs 3 for TC) and 1 year (1 for SC vs 2 for TC). Mortality rate was not statistically different between the two groups at 30 days, 6 months, and 1 year (P>.05 for all). Conclusion. Both SC and TC access routes can be safe and feasible options for TAVR in patients at increased or prohibitive surgical risk with contraindications to standard transfemoral access.

J INVASIVE CARDIOL 2020;32(12):463-469. Epub 2020 September 10.  

Key words: alternative access TAVR, subclavian access, transcaval access

Transcatheter aortic valve replacement (TAVR) is a widely implemented treatment for patients with hemodynamically significant aortic stenosis and increased surgical risk. Typically, access is obtained percutaneously via the common femoral arteries, an approach that holds a number of procedural advantages.1 Recent studies1,2 have demonstrated that as many as 5% of TAVR candidates have contraindications for transfemoral access, and while recent findings from Transcatheter Valve Therapy trial data show that a transaxillary (TA)/subclavian (SC) approach has become a frequently employed alternative access route for balloon expandable valves,3 little consensus exists on a preferred alternative access route. For the cohort of patients at high or prohibitive surgical risk with contraindications to femoral access, alternative access strategies have been demonstrated in proof-of-concept studies and in several case series, and from smaller studies with multicenter trial data.3 

In this single-center study, we review the demographics, procedural safety and feasibility metrics, and outcomes to 1 year in patients who underwent either SC or transcaval (TC) access TAVR between December 2011 and January 2020.


Study design and purpose. This was a retrospective, non-randomized, single-center study designed to assess safety, feasibility, and efficacy of TAVR procedures when performed using SC and TC access routes.

TAVR patient database. A secure, in-house, Health Insurance Portability and Accountability Act (HIPAA)-compliant database was created containing patient demographics, procedural data, and postoperative follow-up for every patient who underwent TAVR at our tertiary referral center since 2011. This database was then retrospectively reviewed with attention to the following parameters: (1) type of access employed for the TAVR procedure, with selection of patients who underwent SC or TC access TAVR for further review; (2) in the SC and TC patient cohorts, preoperative factors were assessed that included demographics (age at time of procedure, gender, comorbidities, previous cardiac bypass or valve replacement, preoperative risk scoring [standardized frailty assessments and New York Heart Association symptom scoring]; (3) preoperative hemodynamics (ejection fraction, aortic valve area, valvular gradients, pulmonary artery pressure); (4) procedural details (intraoperative blood loss, procedural duration, anesthesia used, and complications encountered); (5) in-hospital outcomes; and (6) outpatient follow-up at 30 days, 6 months, and annually thereafter. Review of patient electronic medical records (Epic EHR; Rightfax Software) was additionally employed where needed to retrospectively analyze patient data according to the aforementioned parameters. From 2011 to 2020, a total of 41 patients underwent SC access TAVR and 22 underwent TC access TAVR (out of 1948 TAVRs performed at our institution during the study period), and form the basis of the study detailed herein. 

Selection of patients for TAVR and determination of access strategy. All patients included in this study were previously referred to the structural heart team at our institution for treatment of hemodynamically significant aortic valve stenosis. They were determined to be at increased or prohibitive risk for surgical aortic valve replacement (SAVR) by two independent evaluations by cardiothoracic surgeons at this institution. Determination of TAVR eligibility was performed through comprehensive review of medical comorbidities, preoperative risk stratification, frailty assessments, determination of structural and valvular disease through echocardiography (transthoracic or transesophageal echocardiography), angiography, and through contrast-enhanced computed tomography scans to determine access suitability. All patients included within this study were determined to have contraindications to transfemoral arterial access. In the early experience, TC access was reserved for those without SC arterial access; over the second half of our experience, TC access was often the first choice for alternative access. 

TAVR procedural details. All SC and TC access TAVR procedures were performed within a hybrid operating room equipped with bypass capabilities and cardiothoracic team on standby.

SC access TAVR. Access was obtained either by cutdown in the SC artery (n = 36) or by percutaneous axillary access (n = 8). In the cases where percutaneous access was employed, conscious sedation was used. Predominant access route was via left SC (n = 32), but right SC access was also used (n = 9). A sheath was placed into the SC artery, the valve was crossed in standard fashion,4,5 and TAVR was then performed with balloon-expandable stent valve (n = 14) or self-expanding stent valve (n = 27). Postoperative echocardiography (either transesophageal or transthoracic) was used for assessment of aortic insufficiency, paravalvular leak, and overall valve function. Mean gradients were recorded, with angiography to reassess aortic insufficiency. The stent valve delivery system was removed over the wire, the large sheath pulled back, and surgical or percutaneous closure of the SC access site was performed (Figure 1). 

TC access TAVR. Procedural details have been previously described.6 Briefly, access was gained from the right femoral vein. A gooseneck snare sized to the aortic crossing site was placed via the left femoral artery and positioned at a predetermined site for aortic crossing. A right-facing guide was placed in the inferior vena cava across from the gooseneck. A 0.014˝ coronary wire was placed inside a 0.14-0.35 Piggyback converting catheter (Teleflex Systems), and then placed inside a 0.35 mm microcatheter, with this system placed into the guide. The wire was aimed at the center of the gooseneck, and an electrocautery system was attached to the distal wire. The wire was advanced across the inferior vena cava using electrocautery and into the aorta, snared into the aorta, and advanced to the thoracic descending aorta. With the 0.014˝ wire fixed in the thoracic aorta, the Piggyback catheter was advanced, and the microcatheter was then advanced over the wire to the thoracic aorta. A stiffened 0.035˝ wire (Lunderquist) was advanced to the descending aorta. A large-bore 14/16 Fr Esheath (Edwards Lifesciences) or 18 Fr DrySeal sheath (Gore Medical) was advanced from the right femoral vein to the abdominal aorta. TAVR was then performed with balloon-expandable stent valve (n = 10) or self-expanding stent valve (n = 12) per standard transfemoral techniques. The larger sheath was withdrawn to inside the inferior vena cava, and a 10/8 Amplatzer ductal occluder (Abbott Vascular) was used to close the aortic puncture site. Angiography was again performed to confirm adequate closure (Figure 2).

Procedural safety assessment. As part of this study, an analysis of procedural safety factors was undertaken, which included access success (defined as successful placement and removal of sheath), procedural success (defined as successful deployment of the TAVR valve), the valve size and type employed, procedural duration (for TC access TAVR, this assessment was divided into “crossing time,” and “closure time” as previously described),7 intraprocedural transfusion need, total blood loss, type of device closure employed, need for covered stent use as part of a bail-out procedure, and need for procedural reintervention at any time point assessed. The duration of postprocedural recovery in the Intensive Care Unit and total hospitalization time were also assessed.

Postoperative outcomes. Outcomes following TC access and SC access TAVRs were divided into those occurring during the immediate postprocedure inpatient stay, the first 30 days following the procedure, and at annual assessments at 6 months and 1 year post procedure. Parameters that were analyzed at these time points included mortality rate, postprocedure transfusion, acute kidney injury, new dialysis requirement, myocardial infarction, cerebral vascular accident, and need for pacemaker implantation.

Data analysis. Data are reported as mean ± standard deviation, percentage of patient cohort, or median value, where appropriate. Continuous and integer data were compared using Student’s t-test or Fisher’s exact test. A P-value <.05 was considered statistically significant. 


Preoperative demographics and hemodynamics. While the SC patients were older at the time of procedure, groups were similar in terms of surgical risk and comorbidities, and were highly symptomatic (Table 1). Ejection fraction was slightly worse in the SC group, while other hemodynamic parameters were similar (Table 1).

 Procedural safety and feasibility. Computed tomography angiography was used to find optimal targets along the aortic wall and for ex vivo valve sizing determination (Figures 1A-1C) prior to undergoing subclavian TAVR (Figures 1D-1F), with similar preoperative assessments (Figures 2A and 2B) made prior to undergoing TC access TAVR (Figures 2C-2I). Following this, specific intraprocedural assessments were obtained to determine safety and feasibility (Table 2). 

Intraoperative assessments of SC and TC access TAVR. Successful access and deployment of the TAVR valve were performed in all patients (Table 2). General anesthesia was used in the 34 SC cases undergoing cutdown, with conscious sedation used in the 7 percutaneous access SC cases and in 14 of 22 TC access cases (Table 2). Blood loss was similar between groups; only 1 patient in the TC group required transfusion out of the entire cohort (Table 2). 

In the SC group (n = 41), procedural duration was defined as the average time from first percutaneous stick to completion of final angiography, and was assessed as 72.1 ± 12.6 minutes. In the TC group (n = 22), procedural duration was 39.8 ± 12.5 minutes (with 22.8 ± 6.3 minutes the average crossing time of snare to sheath, and 17.0 ± 6.2 minutes the average time of closure to completion of final angiography). No difference in the rate of paravalvular leak was seen (Table 2). 

TC outcomes. All patients with TC access had endoleak scores of 0 or 1 at the time of procedure conclusion. There was 1 intraprocedural death in the cohort, which was considered to be associated with the TAVR procedure, but not the TC access site. In this patient, TC access was uncomplicated, and the sheath passed into the aorta without incident. Immediately after balloon inflation for the TAVR valve, catastrophic hemodynamic collapse occurred, requiring emergent sternotomy with intraoperative findings of annular rupture and left main coronary occlusion. After prolonged surgical efforts to repair the rupture, replace the aortic valve, and graft the coronaries, the TC access sheath was removed while on bypass. The patient failed to separate from bypass with persistent hypotension. Given that the patient was in extremis, an endograft was placed at the time of sheath removal; however, hypotension was persistent and ultimately the patient died. In retrospect, the endograft was considered unnecessary and death was clearly due to annular rupture and not access route. 

Postprocedural outcomes. Immediate postprocedure period. There were no other deaths that occurred during the inpatient period, and no patients required surgical reintervention in either cohort. Rates of permanent pacemaker implantation, cerebrovascular accident, and vascular complications were similar between both cohorts (Table 3). No patient required long-term hemodialysis after index TAVR procedure. Average postprocedure Intensive Care Unit stay and average total duration of admission stay were not statistically different between the two groups. Additionally, postprocedure hemodynamics were reassessed prior to discharge and were consistent in both groups (Table 3). 

Long-term procedural outcomes. Comparison of outcomes was performed using echocardiograms collected at outpatient follow-up, with clinical and hemodynamic data collected at 30 days (Table 4), 6 months (Table 5), and 1 year (Table 6). 

At 30 days following hospital discharge, postoperative outcomes data was available for 41 patients from the SC group and 22 patients from the TC group. There was 1 mortality from the SC cohort (2.4%) who died of non-cardiogenic (septic) shock at postoperative day 28, and no additional mortalities in the TC group after the single intraoperative mortality following annular rupture. There were 5 cerebrovascular accidents in the SC group and 2 in the TC group (P>.05). No differences were seen between the two groups for readmission or permanent pacemaker rate.

On assessment of hemodynamics by transthoracic echocardiography at 30 days, there were no statistical differences between groups for ejection fraction, aortic valve area, mean gradient, peak gradient, maximal velocity, or pulmonary artery pressure (Table 4), with improvements at each postprocedure hemodynamic assessment made compared with baseline preoperative measurements (Table 4).

Outcomes at 6 months and 1 year. On follow-up at 6 months post procedure, data were available from 41 patients in the SC group and 16 patients in the TC group. There was 1 death in the TC group (6.3% of the 16 patients) (Table 5) related to a 92-year-old patient with chronic respiratory failure from chronic obstructive pulmonary disease and readmission with influenza infection. There was 1 readmission in the SC group (2.4% of the 41 patients) and 1 readmission  in the TC cohort (6.3% of the 16 patients; P=.50) (Table 5); both readmissions were for congestive heart failure with 6 months of the index procedure. Two patients from the SC group (4.9% of the 41 patients) required permanent pacemaker implantation compared with 3 patients in the TC group (18.7% of the 16 patients; P=.09) (Table 5). There were no significant differences between either group on reassessment of postoperative hemodynamics by transthoracic echocardiography (Table 5). 

At 1 year, data were available for 40 patients in the SC group and 16 patients in the TC group (Table 6). As with data at 30 days and 6 months, there were no statistically significant differences in clinical outcomes or in hemodynamic data (P>.05) (Table 6). 


There is little consensus in committee guidelines on preferred alternative access route for TAVR in patients without transfemoral access. Data from the Transcatheter Valve Therapy registry show that the axillary/SC access route is by far the preferred default route in United States centers by a factor of 10 or more.3 However, few data exist to support any specific approach as the preferred route. 

We sought to compare clinical characteristics and outcomes out to 1 year in a single-center experience with both SC and TC access routes. Potential advantages of TC access over other alternative approaches include preservation of transfemoral set-up, avoidance of need for surgical cutdown or thoracic incisions, use of moderate sedation, and rapid postoperative recovery given only large-bore venous access. However, adoption of TC has been slowed by lack of consensus over safety and efficacy, perceived technical difficulties with the access route, and lack of additional data after an initial pivotal trial of 100 patients. The Transcatheter Valve Therapy registry did not track TC access in the database until recently; therefore, TC-specific registry data are lacking. Our center adopted TC as a preferred alternative access route in July of 2016, and we have a relatively high volume of patients with TC access; we report our results as compared with our SC access cases.

As expected, our patients were all high risk or inoperable surgical risk. However, TC access was able to be achieved uniformly and quickly in all patients. Average procedural time skin to skin was under an hour, with conscious sedation in >80% of patients. No vascular complications were seen. One adverse outcome (intraprocedural death due to annular rupture) was seen, but was determined to be unrelated to TC access; the remainder of the group had 1 stroke and 1 transfusion of 1 unit. Overall length of stay, Intensive Care Unit length of stay, need for transfusion, stroke, and vascular complication rates were acceptable in this high-risk/inoperable group when compared with the more widely used SC approach. Follow-up outcomes to 1 year post procedure demonstrated a stable hemodynamic and clinical profile. 

Study limitations. This study has a number of limitations. The data were non-randomized; therefore, direct comparison of outcomes is subject to operator selection bias. TC was initially used as a second option for alternative access, but has become our first-line choice over time, and the relatively small number of patients may limit the ability to clearly discern differences between the two groups. A single operative team performed all SC and TC accesses, and less-experienced sites or teams might not consistently reproduce the same results. Additionally, in the TC group, access was obtained in two different approaches, with either (1) the placement of purse-string sutures by the surgeon; or (2) with percutaneous transaxillary access, which may provide an additional source of variation in the analysis of postprocedure outcomes that was not taken into consideration. Finally, the use of balloon-expandable stent valves and self-expanding stent valves in both cohorts may be a confounder. 


We demonstrate in this single-center study that both SC and TC approaches for TAVR are feasible and safe strategies for patients who have contraindications to traditional transfemoral arterial access. We report acceptable periprocedural, 30-day, 6-month, and 1-year outcomes in a cohort of patients at high or prohibitive surgical risk. Larger-scale comparisons, especially within the national Transcatheter Valve Therapy registry, should be performed for greater clarity on the comparative safety and efficacy of these two approaches to alternative access for TAVR.

From the Sentara Heart Valve and Structural Disease Center of Sentara Heart Hospital, Norfolk, Virginia.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

The authors report that patient consent was provided for publication of the images used herein.

Manuscript accepted April 15, 2020.

Address for correspondence: Paul Mahoney, MD, Sentara Heart Valve and Structural Disease Center, Sentara Heart Hospital, 600 Gresham Drive, Norfolk, VA 23507. Email:


1. Overtchouk P, Modine T. A comparison of alternative access routes for transcatheter aortic valve implantation. Expert Rev Cardiovasc Ther. 2018;16:749-756. Epub 2018 Sep 28. 

2. Madigan M, Atoui R. Non-transfemoral access sites for transcatheter aortic valve replacement. J Thorac Dis. 2018;10:4505-4515.

3. Dahle TG, Kaneko T, McCabe JM. Outcomes following subclavian and axillary artery access for transcatheter aortic valve replacement: Society of Thoracic Surgeons/American College of Cardiology TVT registry report. JACC Cardiovasc Interv. 2019;12:662-669.

4. Muensterer A, Mazzitelli D, Ruge H, et al. Safety and efficacy of the subclavian access route for TAVI in cases of missing transfemoral access. Clin Res Cardiol. 2013;102:627-636. Epub 2013 May 11.

5. Petronio AS, De Carlo M, Giannini C, et al. Subclavian TAVI: more than an alternative access route. EuroIntervention. 2013;9(Suppl):S33-S37.

6. Greenbaum AB, O’Neill WW, Paone G, et al. Caval-aortic access to allow transcatheter aortic valve replacement in otherwise ineligible patients: initial human experience. J Am Coll Cardiol. 2014;63(25 Pt A):2795-2804. Epub 2014 May 7. 

7. Lederman RJ, Babaliaros VC, Rogers T, et al. The fate of transcaval access tracts: 12-month results of the prospective NHLBI transcaval transcatheter aortic valve replacement study. JACC Cardiovasc Interv. 2019;12:448-456.