Trends and Outcomes of Alternative-Access Transcatheter Aortic Valve Replacement

Abstract: Background. Alternative access (AA) is still required for a significant proportion of patients undergoing transcatheter aortic valve replacement (TAVR). We sought to compare the clinical outcomes of patients undergoing AA vs transfemoral (TF) access. Methods. We retrospectively evaluated the outcomes of patients undergoing AA-TAVR between April 2011 and November 2016, and compared them with those who had TF-TAVR. Chi-square and Mann-Whitney U-tests were used to compare the groups and Kaplan-Meier analysis was performed to estimate long-term survival. Results. TAVR was performed in a total of 600 patients, of which 78 (13%) had AA and 522 (87%) had TF access. Patients undergoing AA were younger, and had higher prevalence of chronic obstructive pulmonary disease, peripheral vascular disease, prior myocardial infarction, and prior sternotomy. Greater than mild paravalvular regurgitation (4.2% vs 0.0%; P=.04) and unplanned vascular surgery (5.4% vs 1.3%; P=.09) were more frequent in the TF group. However, patients who underwent AA had longer hospital stay (median 4 days [interquartile range, 3-7 days] vs 3 days [interquartile range, 3-4 days]; P<.001) and an increased incidence of prolonged ventilation (5.1% vs 1.3%; P=.06), 30-day all-cause (5.1% vs 1.7%; P=.08), and cardiovascular mortality (5.1% vs 1.3%; P=.04). The 6-month (15.7% vs 5.7%; P<.01) and 12-month (16.7% vs 10.2%; P=.07) mortality rates were higher for patients undergoing AA. The usage of AA significantly decreased over time (P=.01), primarily driven by a decrease in transapical (P<.001) and direct aortic access (P=.02). Conclusions. AA-TAVR is associated with an increased incidence of postoperative adverse events, including mortality, when compared with those undergoing TF access.

J INVASIVE CARDIOL 2019;31(7):E184-E191.

Key words: alternative access, paravalvular regurgitation, transcatheter aortic valve replacement

Transcatheter aortic valve replacement (TAVR) has proven to be superior to medical therapy for inoperable patients with severe symptomatic aortic stenosis (AS), as well as to be similar or superior to surgical aortic valve replacement for those at intermediate and high surgical risk.^1-3 Transfemoral (TF) access remains the most utilized and preferred route for TAVR. However, severe peripheral vascular disease and other patient characteristics precluding TF access are present in a significant proportion of patients.² Although, alternative access (AA) was needed in 20%-30% of the patients in the clinical trials with the self-expanding and balloon-expandable valves,³ data from the Transcatheter Valve Therapy (TVT) registry have shown AA rates as high as 43.6%.² With improvements in device technology, including smaller delivery sheath size, it is possible that the need for AA has decreased over time and its outcomes have improved. Although there are recent reports showing TF access use in close to 90% of cases, recent data from the TVT registry revealed thoracotomy or sternotomy access for TAVR was still used in 30% of the cases in 2014, with only an 8.7% reduction from 2012 to 2013.⁴

The available data suggest a challenging learning curve with AA-TAVR, with significant improvements over time in the procedural time, as well as a reduction in adverse outcomes, including improved survival.^5,6 Several AA routes exist, including transapical, direct aortic, axillary-subclavian, transcarotid, and transcaval. Transapical access, which is the second most utilized route after TF, has been associated with an increased risk for major adverse events and death.^7-9 The direct aortic approach was found to have similar results when compared with transapical access in a large study from the TVT registry.⁹ The largest study on axillary-subclavian access included 141 patients and showed comparable outcomes with TF access after 2 years of follow-up.¹⁰ Studies with small patient cohorts using the transcarotid and transcaval approaches without comparison with TF or other AA routes have been reported with acceptable results.^11-13 Despite the available data reporting on the feasibility of various AA routes, contemporary data still show AA may have worse outcomes, including mortality, when compared with the TF approach.^8,9,14-16 Therefore, the goal of the present study was to evaluate the clinical outcomes of patients undergoing AA-TAVR, as well as the trend in the usage across quartiles over time from a single academic medical center.

Methods

Study population and follow-up. We retrospectively reviewed the Duke Aortic Valve Disease database to identify all consecutive patients who underwent AA-TAVR between April 2011 and November 2016, and compared them with patients who underwent TF-TAVR. The study was approved by the Duke University Hospital institutional review board. Only patients who underwent TAVR using commercially approved transcatheter aortic prostheses were included. The baseline characteristics, echocardiograms, operative variables, and postoperative data of TF-TAVR patients were compared with AA-TAVR patients. All patients were evaluated in the outpatient setting 30 days after TAVR. Vital status for all patients was obtained from local electronic medical records and clinical follow-up. The definitions and variables selected were based on the American College of Cardiology (ACC)/Society of Thoracic Surgeons (STS) TVT Registry data collection form, version 2.0.

Patient selection and interventions. All patients with severe symptomatic AS were evaluated clinically and for TAVR in a multidisciplinary review conference composed of interventional cardiologists, cardiothoracic surgeons, imaging specialists, and program and research coordinators. All patients underwent preoperative right and left cardiac catheterization, transthoracic echocardiography, and electrocardiography-gated, contrast-enhanced computed tomography (CT) angiography. A single-injection contrast bolus protocol in combination with two-part multidetector CT was the technique used for CT angiography image acquisition.¹⁷ Access route for TAVR was primarily decided based on the operator’s review of the CT angiography images. Operative risk was estimated using the STS PROM score, frailty, and anatomic considerations, with the final determination of surgical risk based on the Duke heart team’s assessment. The antiplatelet regimen after TAVR, irrespectively of access route, consisted of aspirin 81 mg daily indefinitely and clopidogrel 75 mg daily for 3-6 months. For patients undergoing direct aortic or transapical approach, the initiation of clopidogrel was delayed for 3-5 days. For patients with new-onset or history of atrial fibrillation requiring oral anticoagulation, clopidogrel was not prescribed.

Statistical analysis. Continuous variables, including patient demographics and operative data, were expressed as the median and interquartile range (IQR, 25%-75%) and compared with a Mann-Whitney U-test. All dichotomous variables were compared using Chi-square analysis. Logistic regression analysis was carried out to identify variables associated with the 30-day all-cause and cardiovascular mortality. Variables with a P-value ≤.10 in the univariable analysis were entered into the multivariable model. The trend over time in the usage of various types of access for TAVR was also assessed using 1-way analysis of variance. A P-value of <.05 was considered statistically significant. A Kaplan-Meier analysis was performed to estimate long-term survival. The statistical analyses were conducted using Statistical Package for Social Sciences, version 21 (SPSS, Inc).

Results

From April 2011 to November 2016, a total of 600 patients underwent TAVR, from which 522 (87.0%) had TF approach and 78 (13.0%) had AA approach. Patients undergoing AA were younger (median age, 76 years [IQR, 70-82 years] vs 81 years [IQR, 73-86]; P<.001) and had a higher prevalence of COPD (50.0% vs 29.5%; P<.001), peripheral vascular disease (65.4% vs 29.5%; P<.001), prior myocardial infarction (26.9% vs 17.0%; P=.03), prior sternotomy (57.7% vs 41.0%; P<.01), and were active smokers (56.4% vs 24.3%; P<.001) when compared with TF access patients. However, the STS score was similar in both groups (median 6.5% in TF patients [IQR, 4.1%-9.8%] vs 7.2% in AA patients [IQR, 4.2%-9.4%]; P=.38) (Table 1).

Overall, self-expandable valves were implanted in 345 patients (57.5%), balloon-expandable valves were implanted in 253 patients (42.2%), and TAVR for bioprosthetic valve dysfunction (valve-in-valve) was performed in 33 patients (5.5%) (Table 2). Intraprocedural complications were infrequent and included valve malposition requiring intervention (0.5%), ventricular perforation or tamponade (0.7%), annular rupture (0.5%), and aortic dissection (0.2%), leading to 1 (0.2%) conversion to open heart surgery. Procedural success, defined as successful TAVR implantation, was achieved in 98.7% of AA patients and 97.1% of TF patients (P=.52). Transaxillary access, which was performed in 30 patients (38.5%), was the most common AA, followed by direct aortic in 24 patients (30.8%), transapical in 19 (patients 24.4%), and transcarotid in 5 patients (6.4%). A small surgical cutdown was performed for all transaxillary and transcarotid cases. The type of implanted valve was significantly different between the access types, with self-expanding valves more frequently used during transcarotid and transaxillary access and balloon-expandable valves used more often during transapical access (Table 3).

Mild-or-less postoperative paravalvular regurgitation, as assessed by transthoracic Doppler echocardiography before discharge, was achieved more frequently in patients undergoing AA (100% vs 96% in TF patients; P=.05) (Table 4). However, several postoperative adverse events were more frequent in the AA group, including prolonged ventilation (5.1% vs 1.3%, P=.06), Acute Kidney Injury Network stage >1 acute kidney injury (21.8% vs 14.4%; P=.07), and new atrial fibrillation (10.3% vs 5.4%; P=.08), leading to a longer median hospital stay in the AA group (4 days [IQR, 3-7 days] vs 3 days [IQR, 3-4 days]; P<.001) (Table 5). Overall, the incidence of major postoperative adverse events was low and included ischemic stroke (1.5%), myocardial infarction (0.2%), life-threatening or disabling bleeding (0.3%), unplanned cardiac surgery (0.8%), and new need for permanent pacemaker implantation (13.0%), with no statistically significant differences between the two groups. Only 1 stroke occurred in the AA group in a patient who underwent direct aortic access. However, the 30-day cardiovascular mortality rates (5.1% vs 1.3%; P=.04) and all-cause mortality rates (5.1% vs 1.7%; P=.08) were higher for patients who underwent AA-TAVR. The actual 30-day all-cause mortality rate was lower than the expected postoperative mortality rate based on the STS score for both the TF and AA groups (Figure 1). Follow-up (median, 13 months [IQR, 5-25 months]) was available for 99.1% of patients, with no differences between the groups. Compared with patients undergoing TF, those who received AA had higher mortality rates at 6 months (15.4% vs 5.7%; P<.01) and 12 months (16.7% vs 10.2%; P=.07) after TAVR (Table 5). However, 2 years after TAVR, the survival rate was similar between both groups. There was no difference in the 30-day or 1-year all-cause mortality rates between the different AA routes (Table 6). TF access was associated with a decreased risk of both the 30-day all-cause and cardiovascular death rates in the multivariable analysis (Table 7).

The trend in the usage of AA over time was assessed by dividing the group in quartiles of equal number (n = 150) of consecutive patients (1st quartile, April 2011 to June 2013; 2nd quartile, June 2013 to August 2014; 3rd quartile, August 2014 to August 2015; 4th quartile, August 2015 to November 2016). There was a significant decrease in the use of AA over time (P=.01), with the highest rate during the 2nd quartile (19.3%) and the lowest rate in the 4th quartile (6.7%) (Table 8 and Figure 3). This was driven by a significant decrease in the use of transapical (P<.001) and direct aortic access (P=.02), which were almost nonexistent during the latest quartile.

Discussion

In this analysis of our experience with AA-TAVR, we compared patient characteristics, rates of procedural complications, and rates of short-term and long-term survival among 600 patients undergoing either AA- or TF-TAVR. Patients undergoing AA-TAVR were more likely smokers and had an increased incidence of chronic obstructive pulmonary disease (COPD), peripheral vascular disease, and a prior history of myocardial infarction. Patients undergoing TF-TAVR had increased rate of unplanned vascular surgery, yet had decreased median length of stay. The rates of 30-day all-cause and cardiovascular mortality were higher in patients undergoing AA-TAVR, and 6-month and 12-month survival rates were significantly worse for those patients undergoing AA- vs TF-TAVR. Finally, we demonstrated that our utilization of AA has significantly decreased over the study period, primarily driven by decreased use of transapical and direct aortic access approaches.

In an analysis of postprocedural complications and outcomes of data from the Italian OBSERVANT study,⁸ AA (transapical) TAVR was significantly associated with lower rates of major vascular damage (7.2% vs 1.0%; P<.01) and paravalvular regurgitation (50.2% vs 34.6%; P<.01). The decreased rates of postprocedural aortic regurgitation in patients undergoing AA-TAVR are thought to be secondary to an easier, more direct positioning of the valve, especially when using the transaortic or transapical approaches avoiding vascular tortuosity.

In our center’s experience, patients undergoing AA-TAVR had a statistically longer length of stay than those undergoing TF-TAVR. In a substudy of PARTNER-1,⁷ the median postprocedural length of stay was 8 days for trans- apical TAVR vs 5 days for TF-TAVR (P<.001). Additionally, recovery (as assessed by New York Heart Association class) after transapical TAVR was significantly slower at 30 days than after TF-TAVR. Furthermore, in a subanalysis of TVT data comparing postoperative outcomes of transapical, transaortic, and transcarotid TAVR techniques, the median length of stay ranged between 9-11 days depending on access site, which is significantly longer than the median length of stay of 3 days for TF-TAVR in our study.¹⁸ Additionally, another multi-institutional study of three high-volume TAVR centers demonstrated that transapical access was associated with a significantly longer length of stay than TF access (hazard ratio, -0.49; 95% confidence interval, 0.41-0.58; P<.001).¹⁹ AA-TAVR has consistently been shown to be associated with prolonged lengths of stay.

As with our analysis, several studies have demonstrated increased short-term mortality rates for patients undergoing AA-TAVR with similar long-term survival between AA- and TF-TAVR. Results from the PARTNER-I study⁷ demonstrated increased 30-day mortality for AA (transapical) TAVR when compared with a TF approach (9.1% vs 3.7%; P<.001). However, at around 4 months after TAVR, the hazard functions converged so that the mortality risk of AA- vs TF-TAVR was similar after this point. Data from the United Kingdom TAVR registry²⁰ of 1620 patients undergoing either AA (transapical) or TF-TAVR demonstrated increasing mortality in the transapical group at 30 days (11.2% vs 4.4%; P=.01), 1 year (28.7% vs 18.1%; P=.01), and 2 years (56% vs 43.5%; P=.01). While increased mortality in the transapical TAVR population may also be a function of the comorbidities of the patients undergoing this approach, analyses from the United Kingdom TAVR registry suggested a transapical approach was an independent predictor of mortality at both 30 days and 2 years. The transapical approach was also demonstrated to be a significant independent predictor of mortality at 1 year in the FRANCE 2 registry.²¹ Our results demonstrated a similar significant increase in mortality at both 6 months and 1 year. As we have noted, patients with increased comorbidities, including prior myocardial infarction, COPD, and worsened lung function, were significantly more likely to undergo AA-TAVR. It is likely that the baseline comorbidities and risk factors of these patients may contribute to higher postprocedural mortality, in addition to a lengthier recovery after AA-TAVR. Despite the significantly increased incidence of preoperative comorbidities, including COPD, peripheral vascular disease, and prior sternotomy, among others, the median STS PROM score was similar between the groups, which was likely offset by the significant difference in the median age between cohorts. Furthermore, the actual postoperative all-cause mortality rate was lower than the expected rates based on the STS score for both the TF and AA groups.

The present study also examined rates of non-fatal complications, such as acute kidney injury and atrial fibrillation, and demonstrated higher rates of these complications among patients undergoing AA-TAVR. When examining data from the Italian OBSERVANT study,⁸ patients undergoing AA (transapical) TAVR were significantly more likely to experience postprocedural acute kidney injury (44.4% vs 21.9%; P<.001). Koifman et al demonstrated significantly higher rates of postprocedural acute kidney injury (8.0% vs 1.4%) and atrial fibrillation (32% vs 9%) among 648 patients undergoing transapical TAVR vs TF-TAVR.²² However, these findings have not been consistent among studies of post-TAVR outcomes. In a subanalysis of the Italian CoreValve registry, there were increased rates of acute kidney injury among those patients undergoing TF-TAVR compared with those undergoing TAVR with subclavian access.¹⁰ Additionally, this study found no significant difference in postprocedural, in-hospital stroke, or major vascular complications between the two approaches.¹⁰

In the current analysis, there was a significant decrease in the utilization of AA-TAVR over time, with a notable decrease in the use of transapical and direct aortic approaches. With continually decreasing sheath sizes, TF access has become increasingly an option in patients with peripheral arterial disease and difficult femoral access.²³ For example, in the PARTNER trial using first-generation Edwards Sapien valves, the introducer sheaths needed to be 22 or 24 Fr, so that only 60% of cases could be performed with TF access.²⁴ However, nearly 83% of TAVR candidates underwent TF access in the CoreValve trials, when the valves were inserted using an 18 Fr sheath. Given data to support prolonged length of stay and recovery and increased mortality, our more contemporary data demonstrated decreased use of AA-TAVR as TAVR technology improved, introducer sheath size decreased, and preprocedure CT planning allowed more patients to undergo TF-TAVR. If TF access cannot be utilized, we now favor transaxillary access first, with transcarotid access as our second choice. Although percutaneous access of the axillary artery can be done, we prefer to obtain arterial access through a small surgical cut-down for both axillary and carotid access, as it allows complete control of the arteriotomy and its closure. In our current practice, thoracic AA is now reserved for patients without femoral, axillary, and carotid access options. Transcaval access has not been adopted at our center.

Study limitations. There are several limitations to the current analysis; this study is retrospective and limited to the experience of a single, large academic medical center. Additionally, procedural characteristics and approaches utilized over the study period may have differed at our facility when compared with other high-volume centers. Additionally, evolving TAVR technology over the study period could have influenced the access route. Finally, as operator experience with AA- and TF-TAVR likely improved over the course of the study, this could have impacted postprocedural outcomes.

Conclusion

AA-TAVR was associated with worse clinical outcomes, including a higher 30-day all-cause and cardiovascular mortality rate and lower survival rate after 6 and 12 months. In addition to procedural morbidity, the observed differences in clinical outcomes may be the result of the higher prevalence of certain comorbidities in patients requiring AA-TAVR. The use of AA has decreased over time, which was driven by a marked decrease in transapical and direct aortic access.

References

1. Bax JJ, Delgado V, Bapat V, et al. Open issues in transcatheter aortic valve implantation. Part 1: patients selection and treatment strategy for transcatheter aortic valve implantation. Eur Heart J. 2014;35:2627-2638.

2. Holmes DR, Brennan JM, Rumsfeld JS, et al. Clinical outcomes at 1 year following transcatheter aortic valve replacement. JAMA. 2015;313:1019-1028.

3. Reardon MJ, Adams DH, Coselli JS, et al. Self-expanding transcatheter aortic valve replacement using AA sites in symptomatic patients with severe aortic stenosis deemed extreme risk of surgery. J Thorac Cardiovasc Surg. 2014;148:2869-2876.

4. Holmes DR, Nishimura RA, Grover FL, et al. Annual outcomes with transcatheter valve therapy: from the STS/ACC TVT registry. Ann Thorac Surg. 2016;101:789-800.

5. Henn MC, Percival T, Zajarias A, et al. Learning AA approaches for transcatheter aortic valve replacement: implications for new transcatheter aortic valve replacement centers. Ann Thorac Surg. 2017;103:1399-1405. Epub 2016 Oct 17.

6. Suri RM, Minha S, Alli O, et al. Learning curves for transapical transcatheter aortic valve replacement in the PARTNER-1 trial: technical performance, success, and safety. J Thorac Cardiovasc Surg. 2016;152:773-780.

7. Blackstone EH, Suri RM, Rajaswaran J, et al. Propensity-matched comparisons of clinical outcomes after transapical or TF transcatheter aortic valve replacement: a placement of aortic transcatheter valves (PARTNER)-1 trial substudy. Circulation. 2015;131:1989-2000.

8. Biancari R, Rosato S, D’Errigo P, et al. Immediate and intermediate outcome after transapical versus TF transcatheter aortic valve replacement. Am J Cardiol. 2016;117:245-251.

9. Thourani VH, Jensen HA, Babaliaros V, et al. Transapical and transaortic transcatheter aortic valve replacement in the United States. Ann Thorac Surg. 2015;100:1718-1726.

10. Petronio AS, De Carlo M, Bedogni F, et al. 2-year results of Core-Valve implantations through the subclavian access. J Am Coll Cardiol. 2012;60:502-507.

11. Debry N, Delhaye C, Azmoun A, et al. Transcarotid transcatheter aortic valve replacement: general or local anesthesia. JACC Cardiovasc Interv. 2016;9:2113-2120.

12. Mylotte D, Sudre A, Teiger E, et al. Transcarotid transcatheter aortic valve replacement: feasibility and safety. JACC Cardiovasc Interv. 2016;9:472-480.

13. Greenbaum AB, Babaliaros VC, Chen MY, et al. Transcaval access and closure for transcatheter aortic valve replacement: a prospective investigation. J Am Coll Cardiol. 2017;69:511-521.

14. Arai T, Romano M, Lefevre T, et al. Direct comparison of feasibility and safety of TF versus transaortic versus transapical transcatheter aortic valve replacement. JACC Cardiovasc Interv. 2016;9:2320-2325.

15. Schymik G, Wurth A, Bramlage P, et al. Long-term results of transapical versus TF TAVI in a real world population of 1000 patients with severe symptomatic aortic stenosis. Circ Cardiovasc Interv. 2015;8:e000761.

16. Panchal HB, Ladia V, Amin P, et al. A meta-analysis of mortality and major adverse cardiovascular and cerebrovascular events in patients undergoing TF versus transapical transcatheter aortic valve inplantation using Edwards valve for severe aortic stenosis. Am J Cardiol. 2014;114:1882-1890.

17. Walker WL, Boll DT, Bueno JM, et al. Assessment of single-bolus contrast administration technique using hybrid dual-source ECG-gated thoracic and dual-source non-ECG-gated high-pitch abdominopelvic CT acquisitions for procedural planning before transcatheter aortic valve replacement. J Comput Assist Tomogr. 2015;39:207-212.

18. Thourani V, Li C, Devireddy C, et al. High-risk patients with inoperative aortic stenosis: use of transapical, transaortic, and transcarotid techniques. Ann Thorac Surg. 2015:99:817-825.

19. Arbel Y, Zivkovic N, Mehta D, et al. Factors associated with length of stay following trans-catheter aortic valve replacement – a multicenter study. BMC Cardiovasc Disord. 2017;17:137.

20. Blackman DJ, Baxter PD, Gale CP, et al. Do outcomes from transcatheter aortic valve implantation vary according to access route and valve type? The UK TAVI registry. J Interv Cardiol. 2014;27:86-95.

21. Gilard M, Eltchaninoff H, Lung B, et al. Registry of transcatheter aortic-valve implantation in high-risk patients. N Engl J Med. 2012;366:1705-1715.

22. Koifman E, Magalhaes M, Kiramijyan S, et al. Impact of TF versus transapical access on mortality among patients with severe aortic stenosis undergoing transcatheter aortic valve replacement. Cardiovasc Revasc Med. 2016;17:318-321.

23. Barker C, Reardon M. AA and closure options for TAVR. Cardiac Interv Today 2015;9:30-34.

24. Smith CR, Leon MB, Mack MJ, et al. Transcatheter versus surgical aortic-valve replacement in high-risk patients. N Engl J Med. 2011;364:2187-2198.

From the ¹Department of Cardiology; ²Department of Surgery; ³Department of Radiology, Division of Cardiovascular Imaging; and ⁴Department of Cardiothoracic Surgery, Duke University Medical Center, Durham, North Carolina.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Harrison reports grant support from Edwards Lifesciences and Medtronic (unrelated to this manuscript). The authors report no conflicts of interest regarding the content herein.

Manuscript submitted January 10, 2019, provisional acceptance given January 21, 2019, final version accepted January 25, 2019.

Address for correspondence: J. Kevin Harrison, MD, Cardiac Catheterization Laboratory, Duke University Medical Center, 2301 Erwin Rd, Durham, NC 27710. Email: john.harrison@duke.edu