Patient Risk Assessment for Transcatheter Aortic Valve Replacement at Veterans Health Administration Hospitals

Abstract: Objective. To compare patient-level risk assessment at Veterans Affairs (VA) hospitals in patients undergoing transcatheter aortic valve replacement (TAVR) with patients included in the Society for Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy (STS/ACC TVT) registry. Methods. We retrospectively analyzed the outcomes of veterans with severe aortic stenosis (AS) receiving TAVR from 2012-2016 at eight VA hospitals and compared them with TVT registry outcomes from 2012-2015. Patients were identified via administrative data. Univariable and multivariable Cox proportional hazards models were used to examine 30-day and 1-year all-cause mortality, 30-day and 1-year transient ischemic attack/stroke rates, and permanent pacemaker (PPM) implantation rates. Results. During the study period, a total of 726 veterans underwent TAVR including valve-in-valve procedures (n = 50). Patients were predominantly male (98.2%), with mean age of 78.5 ± 9.3 years; 49.1% were at prohibitive risk and 12.1% were at high risk for surgical aortic valve replacement; 30-day and 1-year all-cause mortality rates were 2.5% and 14.7%, respectively; 30-day and 1-year combined TIA/stroke rates were 6.5% and 13.5%, respectively. In the TVT registry, 15.8% and 37.8% of patients were at prohibitive and high risk, respectively; 30-day and 1-year mortality rates were 5.7% and 22.7%, respectively, and stroke rates were 2.1% and 4.0%, respectively. Conclusions. This report on TAVR risk assessment within the VA system demonstrates that despite a large proportion of patients classified as prohibitive risk, TAVR was associated with favorable 30-day and 1-year all-cause mortality rates when compared with published outcomes from the STS/ACC TVT registry. 

J INVASIVE CARDIOL 2020;32(8):302-309. 

Key words: mortality, risk assessment, transcatheter aortic valve replacement

In patients with severe aortic stenosis (AS), transcatheter aortic valve replacement (TAVR) has been shown to be a safe and effective alternative to surgical aortic valve replacement (SAVR) in patients at low,^1,2 intermediate,^3,4 and high risk^5,6 for SAVR. Patients who undergo TAVR at Veterans Health Administration (VA) hospitals are not included in the National Society of Thoracic Surgeons (STS)/American College of Cardiology (ACC) Transcatheter Valve Therapy (TVT) registry and VA patients with severe AS are underrepresented in clinical trials. In the CoreValve high-risk United States pivotal⁵ and intermediate-risk SURTAVI³ trials, the VA Palo Alto Health Care System was the only VA hospital to be included among the 45 study sites. Furthermore, no VA sites were included in the PARTNER series of clinical trials. Since gaining Food and Drug Administration (FDA)approval in 2011, a small number of VA hospitals have been performing TAVR, and until recently outcomes data were limited to single-center reports.^7-14 Recently, a system-wide analysis reported VA-TAVR outcomes from 2012-2017 and demonstrated favorable short- and long-term mortality and rehospitalization outcomes.¹⁵ Importantly, patient-level risk assessment was not available via the VA Clinical Assessment, Reporting, and Tracking System (CART) data used for that analysis, and indeed, STS predicted risk for operative mortality (PROM) scores are not required to be reported prior to TAVR for reimbursement purposes in the VA system. Since differences in patient risk as well as differences in risk may exist between civilian and VA hospitals, our aim was to describe patient risk among veterans undergoing TAVR from patient-level risk-assessment data, and compare those with what has been reported in the civilian population. 

Methods

After institutional review board approval at Hines VA Medical Center, patients with severe AS who underwent TAVR within the VA system between January 1, 2012 and December 31, 2016 were identified via administrative data. Baseline sociodemographic data, medical history, admission, procedure and discharge dates, and date of death were collected. Manual chart review was performed for all patients in order to extract additional patient and procedural characteristics, including risk assessment and classification of patients as low, intermediate, high, or prohibitive risk by the multidisciplinary structural heart team at each individual site. STS-PROM scores were obtained; if no risk score was documented by the heart team, the clinical characteristics that led to patients being considered prohibitive risk, more contemporarily referred to as extreme risk, were collected.

Outcomes evaluated were 30-day and 1-year all-cause mortality and transient ischemic attack (TIA)/stroke, and were compared with the TVT registry report from 2012-2015.­ Patient risk assessment, baseline patient characteristics, and procedural characteristics were compared primarily with a separate TVT registry publication from 2011-2015,¹⁷ as that particular report provided patient and procedural data more consistent with the VA-TAVR cohort data for comparison purposes; however, data on STS-PROM scores were obtained from the TVT registry outcomes report from 2012-2015¹⁶ for reasons described below. Some baseline patient characteristics collected for the VA cohort were not available in these TVT registry reports. Of note, the VA-TAVR data extracted TIA and stroke as a composite endpoint, while the TVT registry reported stroke only. While short-term mortality and stroke outcomes were reported in the TVT registry for 2012-2015, the 1-year mortality rates by the United States Centers for Medicare & Medicaid Services (CMS) linkage were only available from 2012-2014. Vascular complications and bleeding events were not readily available via administrative data in the VA cohort and were thus not included in this analysis, but have been reported recently.¹⁵

Statistical analysis. Descriptive statistics are presented for patients who underwent TAVR. Unadjusted hazard ratios from univariable Cox proportional hazards models were calculated for demographics and clinical characteristics predicting 1-year survival. A multivariable Cox proportional hazards model was specified to include predictors of 30-day and 1-year outcomes. Patient and procedural data were used as covariates. Analyses were performed using SAS 9.4 (SAS Institute). 

Results

A total of 726 patients were identified who had TAVR procedures performed at eight VA sites from 2012-2016. This included all TAVRs for native AS and valve-in valve (ViV), and elective, emergent, and salvage procedures. Patient risk assessment, baseline patient demographic characteristics, and clinical characteristics are provided in Table 1 for both the VA cohort and for TVT registry data from 2011-2015.¹⁷ 

Mean patient age was 78.5 ± 9.3 years and 51.7% were >80 years old in the VA group vs mean age of 83 ± 5 years in the TVT group. Almost half of the patients (49.2%) were deemed to be prohibitive risk for SAVR and 12.1% were high risk (based on an STS-PROM score of >8%) in the VA group vs 15.8% and 38.7% in the TVT group, respectively. Prior coronary artery bypass graft (CABG) surgery occurred in 36.8% of the VA group vs 30.9% of the TVT group, with peripheral arterial disease (PAD) present in 23.4% of the VA group vs 31.0% of the TVT group and COPD on supplemental oxygen in 7.4% of the VA group vs 12.5% of the TVT group. Prior TIA/stroke occurred in 14.5% of the VA group vs 12.2% of the TVT group, while chronic kidney disease (CKD) on dialysis was present in 1.9% of the VA group vs 4.0% of the TVT group. Patient factors most frequently cited to classify patients as prohibitive risk were age, frailty, COPD, redo sternotomy, aortic calcification, pulmonary hypertension, and active malignancy.  

The mean STS PROM in 2012/2013 based on 69 procedures was 7.1%; this remained constant through 2016 (313 procedures). The median STS-PROM score in the TVT registry from 2012-2015¹⁶ was 6.5%, and was higher in 2012 (7.1%) than in 2015 (6.5%). A balloon-expandable valve was used in 64% of the cases in the VA group vs 76.9% in the TVT group, a self-expanding valve was used in 33% of the VA group vs 21.1% of the TVT group, and a mechanically expanded valve was use in 2.8% the VA group vs 0% of the TVT group. Femoral artery access was used in 90.6% of the VA group vs 71.1% of the TVT group. 

The 30-day and 1-year all-cause mortality rates were 2.5% and 14.7%, respectively, in the VA group vs 5.7% and 22.7%, respectively, in the TVT group (Table 2).¹⁶ In the VA group, the 30-day mortality rates remained similar over the study period; however, the 1-year mortality rate decreased from 21.7% to 13.9% over time (Table 3). In comparison, the TVT registry reported improved 30-day mortality (from 7.5% in 2012 to 4.6% in 2015) and 1-year mortality (from 25.8% in 2012 to 21.6% in 2014, by CMS linkage).¹⁶ After multivariable analysis was performed, age (hazard ratio [HR], 1.34 per 10-year increase; 95% confidence interval [CI], 1.07-1.68; P<.05), ejection fraction (HR, 1.19 per 10% decrease; 95% CI, 1.02-1.39; P<.05), prior TIA/stroke (HR, 1.72; 95% CI, 1.08-2.76; P<.05), and CKD (HR, 1.72; 95% CI, 1.15-2.5; P<.01) were associated with 1-year mortality (Figure 1). Use of non-femoral vascular access sites was not associated with increased mortality. All 37 patients who underwent transapical TAVR were alive at 30 days, and there were only 2 mortalities at 1 year (Table 2). Within the cohort of the 50 ViV-TAVR patients, the 30-day and 1-year mortality rates were 4.1% and 10.2%, respectively, while these rates were 2.9% and 11.7%, respectively, in the TVT group (Table 2).¹⁸ 

The 30-day and 1-year TIA/stroke rates were 6.5% and 13.5%, respectively, in the VA group vs stroke rates of 2.1% (overall from 2012-2015) and 4.0% (overall, from 2012-2014 via CMS linkage), respectively, in the TVT group¹⁶ (Table 2). Multivariable analysis demonstrated that age (HR, 1.28 per 10-year increase; 95% CI, 1.01-1.64; P<.05) and prior TIA/stroke (HR, 2.55; 95% CI, 1.64-3.97; P<.01) were associated with post-TAVR TIA/stroke at 1 year (Figure 2). The overall 30-day permanent pacemaker implantation rate was 20.4% (balloon-expandable in 12.9% and self-expanding in 33.5%). In the TVT group, the overall 30-day permanent pacemaker implantation rate was 11.8%.¹⁶ The use of non-femoral access declined in the VA cohort over time, from 11.6% in 2012 to 8.4% in 2016 (Table 3).

Discussion

In this retrospective, patient-level analysis of predicted risk and outcomes within the VA healthcare system, veterans who underwent TAVR represented a very high-risk patient cohort, yet had very favorable outcomes at 30 days and 1 year compared with a contemporaneous civilian population in the TVT registry. 

Risk-assessment reporting differed from the TVT registry, in that STS-PROM scoring was not documented for all patients; we retrieved patient-level data to quantify risk for this report. Veteran patients without STS-PROM scores who were designated as prohibitive risk for SAVR were documented to have risk factors that have been shown to lead to poor outcomes after cardiac surgery. The most prevalent characteristics that led VA patients to be considered prohibitive risk for SAVR included age, frailty, COPD, prior sternotomy, aortic calcification, pulmonary hypertension, and active malignancy. Despite the high proportion of prohibitive-risk patients, all-cause mortality was low compared with TVT registry reports, while 1-year mortality rates decreased significantly over time — a trend that was observed in the TVT registry as well. Mortality was largely influenced by patient age and baseline comorbidities and not by procedural characteristics. Age and prior TIA/stroke were associated with post-TAVR cerebrovascular events. The need for permanent pacemaker implantation post TAVR was higher than reported in the TVT registry; this difference appeared to be related to a higher proportion of patients receiving self-expanding transcatheter heart valves. While the absolute rates of pacemaker implantation after TAVR for self-expanding transcatheter heart valves have decreased, they still tend to be higher than observed with balloon-expandable valves.^1,2,19 

Both the VA and TVT registry cohorts demonstrated an improvement in 1-year mortality over time. This trend is likely multifactorial, involving enhanced patient selection, increasing operator experience, lower-profile delivery systems, and improved valve platforms. In an analysis of operator volume outcomes data from the TVT registry, after controlling for various patient and procedural characteristics, there was a statistically significant lower in-hospital mortality rate after the first 100 cases performed at any given hospital.¹⁷ This observation is in accord with a report demonstrating that low-volume centers (<50 TAVRs/year) had an 8.8% 30-day mortality vs 3.9% in high-volume centers (>100 TAVRs/year).²⁰ The sites included in the TVT registry include a significant number of lower-volume centers. A recent TVT registry report of all non-ViV transfemoral TAVR for severe AS from 2015-2017 noted that approximately 50% of TAVR sites performed <54 TAVRs per year.²¹ Within the two lower-volume quartiles that comprise this group of TAVR sites, the adjusted 30-day mortality rate was approximately 3.16%, which is higher than the 2.5% rate in the VA-TAVR cohort of all comers. 

The four VA sites with the highest volumes comprised 68% of the procedures and performed >40 TAVRs each in 2016, with two of these sites performing >100 TAVRs from 2015 to 2016. In addition, many of the VA-TAVR operators who performed these procedures had academic appointments at their respective University hospital affiliates, and thus may have performed additional TAVRs, which supplemented their operator volumes and facilitated the creation of TAVR programs at VA sites.²² Finally, TAVR site approval is more stringent within the VA system, including rigorous onsite inspection and significant oversight of interventional cardiology and cardiac surgery programs.²³ These findings may suggest that the concentrated procedural volume of the few VA-TAVR sites that predominated in our study may partly explain the lower 30-day and 1-year mortality rates when compared with TVT registry data. 

The overall 30-day and 1-year all-cause mortality rates among VA-TAVR patients were significantly lower than those reported in the TVT registry from 2012-2015 despite almost 50% of VA-TAVR patients being deemed prohibitive risk for SAVR. While this may reflect different populations between the two cohorts, it is also possible that the difference reflects differing approaches and perspectives on assessing, measuring, and classifying operative risk at VA centers. Since the FDA approval for TAVR in intermediate-risk patients, VA heart teams no longer documented STS-PROM scores on a consistent basis.¹⁵ While this might streamline the pre-TAVR assessment process, ensuring a standardized approach for assessing patient risk and accordingly counseling patients to better understand their personal risk for valve intervention may facilitate shared decision making as recommended by current guidelines.²⁴ Conversely, with the recent approval of TAVR for low-risk patients based on randomized trial data (PARTNER 3 and Medtronic Low Risk),^1,2 some have argued that calculation of STS-PROM is moot, as it is no longer a discriminator of eligibility for TAVR as a treatment option. We continue to believe that calculation of STS-PROM is a useful exercise in counseling patients regarding procedural risk, and in ensuring that the heart team continues to be aware of the spectrum of risk and anticipated complications likely to be encountered in the periprocedural period.

The VA group demonstrated relatively constant STS-PROM and 30-day mortality over the study period, whereas the TVT group demonstrated a decrease in 30-day mortality with relatively constant STS-PROM. It is unclear as to whether changing populations selected for treatment over time in either or both cohorts, or improving technical and procedural variables, contributed in varying proportions or were offsetting in these observations. While baseline characteristics were obtained via administrative data for our study, recently reported procedural outcomes in VA-TAVR patients from 2012-2017 using the VA-CART program were notable in that the prevalences of PAD, chronic lung disease, and CKD on dialysis were in fact higher than in the TVT registry¹⁵ (Table 1), thus likely underestimating the true prevalence of these comorbidities by our data-extraction process and further highlighting the high ambient baseline risk in these patients. 

Mortality outcomes data from the TVT registry recently noted that PAD, home oxygen use, worse renal function, and non-femoral access were independent predictors of mortality at 30 days.²⁵ Despite the high prevalence of these particular risk factors for poor outcomes in the VA-TAVR cohort, the short- and long-term outcomes in VA patients were favorable when compared with the TVT registry. Despite being younger as a cohort when compared with TVT, patient age was cited (in combination with other patient characteristics) as a reason to classify patients as prohibitive risk for SAVR in approximately 18% of cases that did not have STS-PROM scores documented. Advanced age is associated with poor long-term survival after SAVR; however, as age increases, there is also competing risk from other comorbid conditions. Some recent registry studies have demonstrated comparable outcomes between TAVR and SAVR with regard to advanced age; however, registry studies have inherent biases toward higher-risk patients receiving TAVR.²⁶ In our study, after multivariate analysis, advanced age was in fact found to be associated with increased 1-year mortality. Among VA-TAVR patients, the mean age was 88.4 ± 5 years for patients who had their age documented as part of the reasoning for classifying them as prohibitive risk, vs a mean age of 78.5 ± 9.4 years in the rest of the patient cohort. 

The assessment of frailty remains an important aspect of patient evaluation for TAVR and has been associated with increased mortality after cardiac surgery independent of age.²⁷ Detailed, standardized frailty assessments were documented in <50% of cases and were site specific. Frailty was listed as a reason for inoperability in ~10% of VA patients who underwent TAVR and did not have STS-PROM scores reported in our patient cohort. Frailty was not associated with short- or long-term mortality among VA-TAVR patients; however, patient frailty was noted as a reason for patients not being SAVR candidates only when no STS-PROM score was documented, and thus, the true prevalence of frailty in our cohort is likely underestimated. More patients in the VA cohort had prior sternotomy than in the TVT registry when combining those with prior CABG and prior SAVR. Redo sternotomy is associated with a number of complications that are associated with high mortality risk;²⁸ thus, it is frequently viewed as an important factor to consider when determining risk for SAVR. While prior sternotomy is not prohibitive for redo cardiac surgery, in combination with other patient factors and comorbidities, it frequently led to a determination of being prohibitive risk for SAVR in the VA-TAVR cohort.

While severe or oxygen-dependent COPD is a common reason for patients not being considered candidates for cardiac surgery, even mild COPD has been shown to be associated with higher long-term mortality after surgical AVR with or without CABG.²⁹ Severe COPD has been associated with high mortality in TAVR patients;³⁰ however, when compared with SAVR, TAVR has been demonstrated to offer a convincing advantage as a treatment modality in patients with COPD not only with regard to mortality, but also kidney injury, respiratory complications, bleeding, and myocardial infarction.³¹ Thus, in the VA patients without documented STS-PROM scores who had COPD, it is reasonable that this combination with other risk factors for cardiac surgery would deem them to be prohibitive-risk candidates for SAVR. 

It has been shown that pulmonary hypertension (PH) is a risk factor for poor outcomes after SAVR.³² PH was infrequently cited as a contributing factor to patients being deemed prohibitive risk for SAVR; however, a significant proportion of patients had PH by invasive hemodynamic assessment. Of the 726 VA patients studied, right heart catheterization data were available for 301 patients. Of these, a total of 176 (58%) had PH, as defined by mean pulmonary artery pressure (MPAP) of ≥25 mm Hg, and 94 of those patients (31%) had at least moderate PH (MPAP ≥35 mm Hg). While there is conflicting evidence with regard to mortality outcomes after TAVR in high-risk patients with PH as measured by echocardiography,^33,34 within the VA-TAVR cohort, the 30-day mortality rates for patients without PH was 0% vs 6.25% in veteran patients with PH measured invasively. In comparison, from TVT registry data evaluating 3724 patients with PH (assessed by either echocardiography or catheterization) undergoing TAVR from 2010-2015, PH was associated with increasing 30-day and 1-year mortality rates with a HR for moderate PH of 1.3 and a HR for severe PH (PASP >60 mm Hg) of 1.635. 

Determination of surgical risk among veteran patients has been previously evaluated in a single-center study of 537 patients undergoing AVR with or without CABG using STS-PROM, EuroScore and VA risk scores. The receiver operator characteristic (ROC) curves revealed that all three scores predicted mortality with c-indexes of 0.73, 0.68, and 0.66, respectively.³⁶ These c-index values indicate that these particular scoring systems alone are lacking in their ability to accurately discriminate procedural risk. Recently, a single-center VA study applied both STS-PROM and STS/ACC-TVT TAVR risk scores to male patients undergoing TAVR, and while both scores overestimated in-hospital mortality, the STS-PROM score significantly overestimated mortality, with an observed/expected ratio of 0.57 (vs 0.83 for the STS/ACC-TVT TAVR risk score). Furthermore, while long-term mortality was similar between STS-PROM quartiles, the highest-risk quartile of patients when using the STS/ACC-TVT TAVR risk score had significantly worse mortality, which suggests that this score might be superior to STS-PROM in identifying the highest-risk veteran patients being considered for TAVR.¹² 

Risk assessment in TAVR is continuing to evolve and shift away from solely using STS-PROM scoring and more toward assessing both cardiovascular and non-cardiovascular risk factors in determining patient risk for TAVR. A recent study evaluated the effect of global comorbidity burden as determined by the 30-parameter Elixhauser Comorbidity Score (ECS) on mortality, complications, and length of stay in 40,000 patients undergoing TAVR. While the ECS performed similarly to the EuroScore and STS score with regard to predicting mortality, it was an independent predictor of vascular complications, acute kidney injury, myocardial infarction, major bleeding, as well as stroke and TIA.³⁷ Incorporating such global comorbidity assessments into existing risk stratification models may better predict outcomes in TAVR patients, and such an approach may be especially advantageous in light of recent FDA approval for TAVR in low-risk patients with severe AS. 

Study limitations. First, this study includes an almost exclusively male cohort. Second, the TAVR procedures and some baseline characteristics were identified from administrative records, which have inherent limitations. Third, differentiating cardiovascular from non-cardiovascular death was not possible because only all-cause mortality data were available. Fourth, since there is inherent subjectivity based on individual surgeon/center classification of prohibitive risk, differences in a point of reference of risk between VA and civilian surgeons/centers could exist.

Conclusion

This study is a unique report on patient-level risk assessment for TAVR performed within the nationwide VA healthcare system. The overall VA patient cohort evaluated in this study represents a very high-risk patient demographic given the prevalence of frailty, prior sternotomy, COPD, and PH. Despite this, TAVR within the VA system achieved favorable 30-day and 1-year all-cause mortality outcomes when compared with the TVT registry. 

From the ¹Division of Cardiology, St. Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada; ²Division of Cardiology, Loyola University Chicago Stritch School of Medicine, and Loyola University Health System, Maywood, Illinois; ³Biostatistics Collaborative Core, Loyola University Chicago, Maywood, Illinois; ⁴Center of Innovation for Complex Chronic Healthcare, Department of Veterans Affairs, Hines, Illinois; ⁵North Florida/South Georgia Veterans Health System, Gainesville, Florida; ⁶Minneapolis Heart Institute at Abbott Northwestern Hospital, Minneapolis, Minnesota; ⁷University of Missouri-Kansas City, Kansas City, Missouri; ⁸University of California, San Francisco, and VA Medical Center, San Francisco, California; and ⁹the Worldwide Network of Innovation and Clinical Education and Research (WNICER), New York, New York.

Funding: A grant was provided by the Cardiovascular Research Institute at Loyola University Medical Center, Maywood, Illinois, for data extraction.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Cohen reports research grant support to his institution and consulting income from Edwards Lifesciences, Medtronic, Abbott Vascular, and Boston Scientific. Dr Garcia reports personal fees and grant support from Edwards Lifesciences. Dr Shunk reports grant support from Siemens Medical Systems, Svelte Medical, and Cardiovascular Systems, Inc; consulting income from TransAortic Medical, Syntactx, and PercAssist; non-financial support from the American College of Cardiology NCDR, outside the submitted work. Dr Weaver reports consultant fees from PCORI. The remaining authors report no conflicts of interest regarding the content herein. 

Manuscript accepted February 4, 2020.

Address for correspondence: Verghese Mathew, MD, Worldwide Network of Innovation In Clinical Education and Research (WNICER) Institute, One Rockefeller Plaza, 11th Floor, New York, NY 10020. Email: verghese.mathew@lumc.edu

1. Popma JJ, Deeb GM, Yakubov SJ, et al. Transcatheter aortic-valve replacement with a self-expanding valve in low-risk patients. N Engl J Med. 2019;380:1706-1715.
2. Mack MJ, Leon MB, Thourani VH, et al. Transcatheter aortic-valve replacement with a balloon-expandable valve in low-risk patients. N Engl J Med. 2019;380:1695-1705.
3. Reardon MJ, Van Mieghem NM, Popma JJ, et al. Surgical or transcatheter aortic-valve replacement in intermediate-risk patients. N Engl J Med. 2017;376:1321-1331. 
4. Leon MB, Smith CR, Mack MJ, et al. Transcatheter or surgical aortic-valve replacement in intermediate-risk patients. N Engl J Med. 2016;374:1609-1620.
5. Adams DH, Popma JJ, Reardon MJ, et al. Transcatheter aortic-valve replacement with a self-expanding prosthesis. N Engl J Med. 2014;370:1790-1798.
6. 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. 
7. Bavry AA, Aalaei-Andabili SH, Park KE, Choi CY, Manning III EW, Stinson WW. Trend and outcomes of direct transcatheter aortic valve replacement from a single-center experience. Cardiol Ther. 2018;7:191-196. 
8. Omer S, Kar B, Cornwell LD, et al. Early experience of a transcatheter aortic valve program at a Veterans Affairs facility. JAMA Surg. 2013;148:1087-1093. 
9. Gurevich S, Reiff C, Bertog S, et al. Transcatheter aortic valve replacement improves health status in elderly veterans. J Invasive Cardiol. 2018;30:207-211.
10. Garcia S, Kelly R, Mbai M, et al. Outcomes of intermediate-risk patients treated with transcatheter and surgical aortic valve replacement in the Veterans Affairs Healthcare System: a single center 20-year experience. Catheter Cardiovasc Interv. 2018;92:390-398.
11. Oakley L, Pritchard W, Colletta J, et al. Development and early experience of the First Joint Military Health System-Veterans Affairs Transcatheter Aortic Valve Replacement Program. Mil Med. 2017;182:e2036-e2040.
12. Reiff C, Gurevich S, Bertog S, Sorajja P, Kelly R, Garcia S. Validation of STS/ACC TVT-TAVR score in veterans undergoing transcatheter aortic valve replacement. J Invasive Cardiol. 2018;30:447-451.
13. Yang J, Zimmet J, Ponna V, et al. Evolution of Veterans Affairs Transcatheter Aortic Valve Replacement Program: the first 100 patients. J Heart Valve Dis. 2018;27:24-31.
14. Liang N, Wisneski A, Wang S, et al. Veterans Affairs heart team experience with transcatheter aortic valve replacement and minimally invasive surgical aortic valve replacement. J Invasive Cardiol. 2019;31:217-222.
15. PS Hall, CI O’Donnell, V Mathew, et al. Outcomes of veterans undergoing TAVR within Veterans Affairs Medical Centers. JACC Cardiovasc Interv. 2019;12:2186-2194. 
16. Grover FL, Vemulapalli S, Carroll JD, et al. 2016 annual report of the Society of Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy Registry. J Am Coll Cardiol. 2017;69:1215-1230.
17. Carroll J, Vemulapalli S, Dai D, et al. Procedural experience for transcatheter aortic valve replacement and relation to outcomes: the STS/ACC TVT registry. J Am Coll Cardiol. 2017;70:29-41. 
18. Tuzcu EM, Kapadia SR, Vemulapalli S, et al. Transcatheter aortic valve replacement of failed surgically implanted bioprostheses: the STS/ACC Registry. J Am Coll Cardiol. 2018;72:370-382.
19. Siontis GCM, Jüni P, Pilgrim T, et al. Predictors of permanent pacemaker implantation in patients with severe aortic stenosis undergoing TAVR: a meta-analysis. J Am Coll Cardiol. 2014;64:129-140. 
20. Wassef AWA, Rodes-Cabau J, Liu Y, et al. The learning curve and annual procedure volume standards for optimum outcomes of transcatheter aortic valve replacement: findings from an international registry. JACC Cardiovasc Interv. 2018;11:1669-1679.
21. Vemulapalli S, Carroll JD, Mack MJ, et al. Procedural volume and outcomes for transcatheter aortic-valve replacement. N Engl J Med. 2019;380:2541-2550.
22. Gurevich S, John R, Kelly RF, et al. Avoiding the learning curve for transcatheter aortic valve replacement. Cardiol Res Pract. 2017;2017:1-5. 
23. Messenger JC. The Veterans Affairs transcatheter aortic valve experience. JACC Cardiovasc Interv. 2019;12:2195-2197. 
24. Otto CM, Kumbhani DJ, Alexander KP, et al. 2017 ACC expert consensus decision pathway for transcatheter aortic valve replacement in the management of adults with aortic stenosis. J Am Coll Cardiol. 2017;69:1313-1346. 
25. Arnold SV, O’Brien SM, Vemulapalli S, et al. Inclusion of functional status measures in the risk adjustment of 30-day mortality after transcatheter aortic valve replacement. JACC Cardiovasc Interv. 2018;11:581-589.
26. Hirji SA, Ramirez-Del Val F, Kolkailah AA, et al. Outcomes of surgical and transcatheter aortic valve replacement in the octogenarians—surgery still the gold standard? Ann Cardiothorac Surg. 2017;6:453-462.
27. Sündermann SH, Dademasch A, Seifert B, et al. Frailty is a predictor of short- and mid-term mortality after elective cardiac surgery independently of age. Interact Cardiovasc Thorac Surg. 2014;18:580-585. 
28. Imran Hamid U, Digney R, Soo L, Leung S, Graham ANJ. Incidence and outcome of re-entry injury in redo cardiac surgery: benefits of preoperative planning. Eur J Cardiothorac Surg. 2015;47:819-823. 
29. Gunter RL, Kilgo P, Guyton RA, et al. Impact of preoperative chronic lung disease on survival after surgical aortic valve replacement. Ann Thorac Surg. 2013;96:1322-1328.
30. Thourani VH, Forcillo J, Beohar N, et al. Impact of preoperative chronic kidney disease in 2,531 high-risk and inoperable patients undergoing transcatheter aortic valve replacement in the PARTNER trial. Ann Thorac Surg. 2016;102:1172-1180.
31. Ando T, Adegbala O, Akintoye E, et al. Is transcatheter aortic valve replacement better than surgical aortic valve replacement in patients with chronic obstructive pulmonary disease? A nationwide inpatient sample analysis. J Am Heart Assoc. 2018;7:e008408.
32. Gutmann A, Zirlik A, Bothe W, et al. Impact of pulmonary hypertension on in-hospital outcome after surgical or transcatheter aortic valve replacement. EuroIntervention. 2017;13:804-810. 
33. Alushi B, Beckhoff F, Leistner D, et al. Pulmonary hypertension in patients with severe aortic stenosis: prognostic impact after transcatheter aortic valve replacement: pulmonary hypertension in patients undergoing TAVR. JACC Cardiovasc Imaging. 2019;12:591-601.
34. Barbash IM, Escarcega RO, Minha S, et al. Prevalence and impact of pulmonary hypertension on patients with aortic stenosis who underwent transcatheter aortic valve replacement. Am J Cardiol. 2015;115:1435-1442. 
35. Don CW, Johnson A, Stebbins A, et al. Pulmonary hypertension Is associated with worse TAVR outcomes independent of COPD: report from the TVT registry. J Am Coll Cardiol. 2016;67:2.
36. Basraon J, Chandrashekhar YS, John R, et al. Comparison of risk scores to estimate perioperative mortality in aortic valve replacement surgery. Ann Thorac Surg. 2011;92:535-540.
37. Nagaraja V, Cohen MG, Suh W, et al. Non-cardiovascular comorbidities as evaluated by Elixhauser comorbidity score in individuals undergoing TAVR. Struct Heart. 2019;3:406-414.