Omer S, Kar B, Cornwell LD, Blaustein A, Levine GN, Ali N, Jneid H, Paniagua D, Atluri PV, Bechara CF, Kougias P, Ruma M, Preventza O, Bozkurt B, Carabello BA, Bakaeen FG. Early Experience of a Transcatheter Aortic Valve Program at a Veterans Affairs Facility. JAMA Surg. 2013;148(12):1087-1093. doi:10.1001/jamasurg.2013.3743
The US Food and Drug Administration recently approved the use of a transcatheter aortic valve in patients for whom traditional valve replacement surgery poses a high or prohibitive risk. Our hospital was one of the first Veterans Affairs facilities to launch a transcatheter aortic valve replacement (TAVR) program.
To evaluate our early experience with transfemoral TAVR.
Design and Setting
We retrospectively reviewed the records of all patients who underwent TAVR during the first year of our program at the Department of Cardiothoracic Surgery, Michael E. DeBakey Veterans Affairs Medical Center.
The mean (SD) age of the patients was 77 (9) years, and their mean (SD) Society of Thoracic Surgeons predicted risk of mortality score was 8.8 (10.7).
All patients underwent TAVR with the SAPIEN transcutaneous valve.
Main Outcome Measures
We evaluated operative mortality and major operative morbidity (stroke, myocardial infarction, renal failure necessitating dialysis, and requirement for mechanical circulatory support, as well as vascular complications and requirement for permanent pacemaker), in addition to length of hospital stay and discharge status.
Between December 21, 2011, and December 13, 2012, a total of 19 transfemoral TAVR procedures were performed at our center. Implantation was successful in all cases. There were no reports of operative (30-day) mortality, prosthetic valve endocarditis, renal failure necessitating dialysis, perioperative myocardial infarction or stroke, or conversion to surgical aortic valve replacement. Seven patients (37%) had mild paravalvular leak, 3 patients (16%) had moderate paravalvular leak, 2 patients (11%) had groin wound complications, 2 patients (11%) required a permanent pacemaker, 1 patient (5%) had a vascular access complication requiring endovascular repair, and 1 patient (5%) required temporary circulatory support (with extracorporeal membrane oxygenation). The mean (SD) length of hospital stay after TAVR was 8.0 (5.9) days. All patients were discharged home.
Conclusions and Relevance
Transcatheter aortic valve replacement can be performed safely and with good outcomes at a Veterans Affairs facility with a committed multidisciplinary team and substantial experience in heart valve and endovascular therapies.
Aortic valve (AV) disease, especially aortic stenosis (AS), carries high mortality risk once it becomes severe and symptomatic.1 Surgical AV replacement (SAVR) is durable and time tested, with an operative mortality of 1% at most high-volume cardiac surgery centers.1 With refinements in operative care, SAVR has been extended to older, sicker patients and has produced acceptable outcomes. Nevertheless, many patients are refused or not referred for SAVR because of real or perceived risk as assessed by the operating surgeon or cardiologist.2- 5 Notable among the risk factors are frailty, liver disease, porcelain aorta, hostile mediastinum, and severe advanced lung disease. Typically, these conditions are not captured by traditional open surgical risk models.6,7
In recent years, transcatheter AV replacement (TAVR) has emerged as an attractive option for patients for whom SAVR poses an above-average risk. The 2-year results of the Placement of Aortic Transcatheter Valves (PARTNER) trial are encouraging.8 In addition to this landmark trial, various registries and investigations evidence the growing popularity of TAVR in prohibitive-risk and high-risk patients. The guidelines have restricted the use of TAVR to centers with the necessary expertise and infrastructure.1
In keeping with these recent developments in the treatment of AV disease, TAVR was introduced at the Michael E. DeBakey Veterans Affairs Medical Center in December 2011. Ours is the first center in the VA system to offer the commercially available transfemoral heart valve system (Edwards SAPIEN transfemoral heart valve system; Edwards Life Sciences) to military veteran patients. Our group’s experience in establishing and launching a TAVR program was previously reported, focusing on the advantages and challenges of doing so in the VA system.9 The present study delineated the 1-year outcomes of our TAVR program and examined its effect on SAVR volume.
After institutional review board approval and waiver of individual consent were granted, we retrospectively reviewed data from all patients who underwent SAVR or TAVR at the Michael E. DeBakey Veterans Affairs Medical Center between December 21, 2011, and December 13, 2012. Patient medical records were reviewed for data on demographics and risk profiles, symptoms and surgical treatments, and mortality and operative morbidity.
All patients who underwent TAVR received extensive screening, including echocardiography, heart catheterization, and computed tomographic angiography to ascertain the feasibility of femoral access. Patients were selected by a multidisciplinary heart valve team (S.O., B.K., L.D.C., A.B., G.N.L., N.A., H.J., D.P., P.V.A., C.F.B., P.K., B.B., and F.G.B.). To determine whether each patient was a TAVR candidate, we followed the device label and the judgment of 3 experienced cardiothoracic surgeons (S.O., L.D.C., and F.G.B.), as well as the consensus of the heart valve team.
During the study period, the Edwards SAPIEN transfemoral heart valve system was used exclusively at our center. Two sizes (23 and 26) were available.
Sixteen TAVR cases were performed in the hybrid operating room; the other 3 were performed in the catheterization laboratory. All procedures used general anesthesia.
A transfemoral retrograde approach was used with open surgical cutdown and exposure of the common femoral vessels. In patients whose femoral arteries could not accommodate the large sheaths (22F to 24F catheter), retroperitoneal insertion of a graft conduit via the iliac arteries was planned.
During prosthesis deployment, rapid ventricular pacing was used to facilitate proper placement. If significant (>2) paravalvular regurgitation was seen on transesophageal echocardiography, repeat dilatation was performed. All patients were extubated in the operating room. Transthoracic echocardiography was performed before hospital discharge, at the 1-month follow-up visit, and yearly thereafter.
The primary outcome measure was operative mortality, defined as death during the index hospitalization or the first 30 postoperative days. Other outcome measures included survival and operative morbidity, including heart block, endocarditis, major bleeding, paravalvular leak, prolonged ventilation, wound complications, vascular access complications, renal failure necessitating dialysis, and perioperative myocardial infarction (MI) or stroke. Follow-up analysis involved a comprehensive review of electronic medical records at our site and referring VA sites.
In addition, we evaluated the management strategies used for all patients referred to us for surgical AV disease and the effect of TAVR on AV referral to our facility. The risk profiles of patients undergoing SAVR and TAVR were also analyzed and compared.
A χ2 test with the Yates correction and the Fisher exact test (when needed) were used to calculate P values for the categorical variables. The t test was used to calculate P values for the continuous variables. All statistical analyses were conducted using commercially available software (SAS version 9.1; SAS Institute).
Between December 21, 2011, and December 13, 2012, transfemoral TAVR was performed in 19 male patients (mean [SD] age, 77  years; age range, 62-91 years) at the Michael E. DeBakey Veterans Affairs Medical Center. All patients had severe symptomatic AS, with a mean (SD) transvalvular gradient of 41.7 (11.7) mm Hg and a mean (SD) AV area of 0.7 (0.2) cm2 (Table 1). The mean (SD) Society of Thoracic Surgeons predicted risk of mortality score was 8.8 (10.7).
There was no perioperative mortality. On follow-up analysis (mean [SD] duration, 8.8 [3.9] months; range, 4-16 months), which was 100% complete, 2 deaths were recorded, one 4 months after surgery from massive stroke and the other at the patient’s home 6 months after surgery from cardiac arrest due to ventricular fibrillation. Therefore, the 1-year mortality rate was 11% (2 of 19 patients).
No patient required emergent conversion to SAVR because of complications or the inability to complete the TAVR procedure. These results are summarized in Table 2.
One patient needed emergent cardiopulmonary bypass (CPB) and institution of extracorporeal membrane oxygenation for stabilization. One patient required more than 1 valve deployment because of moderate to severe aortic insufficiency due to prosthetic valve dysfunction. Deploying a second valve produced excellent function.
We observed no perioperative MI or stroke, renal failure necessitating dialysis, or readmission because of congestive heart failure (Table 2). Two patients (11%) required permanent pacemaker implantation during the index hospitalization for complete heart block.
On follow-up transthoracic echocardiography, 3 patients (16%) had moderate paravalvular leak (PVL). No patient had severe PVL, new-onset PVL, or PVL progression. Furthermore, no structural valve problems were noted. The mean (SD) length of hospital stay after TAVR was 8.0 (5.0) days (range, 2-26 days).
Of 19 patients who underwent TAVR, 16 had inoperable disease, and 3 were considered very high risk. The high-risk patients were treated toward the end of the study period, corresponding with the updated US Food and Drug Administration label (November 2012). The mean (SD) length of hospital stay after TAVR was shorter in the high-risk group (6.0 [1.7] days; range, 5-8 days) than in the nonoperative group (8.4 [6.3] days, range 2-26 days). Other outcome variables were not significantly different between the 2 groups.
Fifteen additional patients were referred for possible TAVR but were deemed not to be TAVR candidates (Table 3). Their mean (SD) Society of Thoracic Surgeons predicted risk of mortality score was 12.4 (7.3). Optimal medical therapy was continued in all these patients. Balloon aortic valvuloplasty was performed in 3 of them as a palliative procedure.
During the same 1-year period, our group performed 68 SAVRs: 32 (47%) were isolated SAVRs, and 36 (53%) were SAVRs combined with coronary artery bypass grafting or other cardiovascular procedures. Among the numerous differences between the patients undergoing SAVR and TAVR, the former were less likely to have prior cardiac surgery, previous percutaneous coronary intervention, or New York Heart Association class IV heart failure (Table 1).
The SAVR group had no endocarditis, renal failure, mediastinitis, 30-day mortality, or perioperative MI or stroke. One reoperation (2%) was required for mediastinal bleeding. Seven patients (10%) required prolonged (>48 hours) mechanical ventilatory assistance. The 1-year mortality rate was 6% (4 patients). All 4 patients who died had conditions that required combined procedures.
In the 4 years before the TAVR program began, the mean annual volume of AV procedures at our center was 53. In the first year of the TAVR program, the annual SAVR volume increased to 68. Therefore, the SAVR volume increased by 22% since the adoption of TAVR.
Veterans have kept up with the rest of the US population in benefiting from contemporary medical care. This has translated into longer life spans, resulting in an increasing number of older patients for whom to provide care. More than 9.3 million veterans are 65 years or older.10 Therefore, it is unsurprising that the prevalence of AV disease (mainly calcific AS, which is predominantly seen in older patients) is rising in the veteran population. In addition, many veterans with AV disease not only are old and frail but also have multiple comorbidities. In a study11 published by our group of all veterans (N = 7142) who underwent isolated AV replacement in the VA system between 1991 and 2007, a total of 7% were 80 years or older, 13% had peripheral vascular disease, 15% had cerebrovascular disease, and 29% had chronic obstructive airway disease.
Keeping up with the growing incidence of calcific AS, the number of isolated AV replacements performed in VA hospitals has increased in recent years.9 Despite this increase, approximately 40% to 50% of veterans with severe AS do not undergo AV replacement for reasons that include frailty, older age, comorbidities, and, most important, subjective assessments by referring physicians and surgeons.4,12
Nevertheless, despite having lived as long as 80 years, these patients still have almost 9 years of life expectancy.13 Therefore, there is great potential value in popularizing interventions, such as TAVR, that can improve the quality of the last few years of life without undue morbidity.
Consistent with this philosophy, our center began using TAVR in the VA population in 2011. To our knowledge, this is the first study of the early outcomes of this groundbreaking procedure in VA patients.
In the first year of our TAVR program, we had favorable results. Our 30-day operative mortality was zero and compares favorably with that of the PARTNER trial’s high-risk, transfemoral surgical cohort, whose operative mortality was 3.4%.14 In addition, we had no perioperative MI, renal failure necessitating dialysis, or major bleeding requiring blood transfusion. This again compares favorably with the data from the PARTNER trial,14 in which 10.9% of patients had major bleeding, 3.4% of patients required renal replacement therapy, and no patients experienced perioperative MI.
Notably, our 30-day stroke and transient ischemic attack rate has been zero compared with 4.6% in the PARTNER trial’s transfemoral cohort.14 At the 1-year follow-up analysis, our major stroke rate was 5% (1 of 19 patients) compared with 6.1% in the PARTNER trial.14
Major vascular access complications are seen in approximately 11% of patients in various studies1,15- 22 and are in part related to high-profile delivery systems and sheaths. With US Food and Drug Administration approval of lower-profile systems, this complication will become less frequent. In our series, only 1 patient (5%) had a major vascular complication (iliac artery dissection). This was recognized quickly and was treated effectively with immediate intraoperative stenting of the dissected vessel.
During all our cases, CPB and perfusionist support were readily available. Full CPB was needed only once (5%), after which the patient required extracorporeal membrane oxygenation for 24 hours because of decompensation during valve deployment. Most other series have also described a 5% incidence of CPB during TAVR1; reasons for instituting CPB have included tamponade, root rupture, coronary occlusion, and severe torrential aortic insufficiency. It is difficult to predict which patients will need emergent CPB support; therefore, it is safer to always perform the procedure with CPB on standby.
Because of the proximity of the conduction system, the profile and design of available valves, and the heavy calcification of the diseased valve and annulus, heart block necessitating pacemaker placement after TAVR has been common. This can occur immediately after ballooning and valve deployment or days or months later, caused by either inflammation or further expansion and remodeling of the valve stent. Most conduction problems occur before valve implantation (46% during balloon valvuloplasty, 29% during expansion of the prosthesis, and 25% during prosthesis positioning and wire crossing of the AV.1,23
In the PARTNER trial,8 the incidence of complete heart block was 3.8% at 30 days, 6.4% at 1 year, and 7.2% at 2 years. These rates are slightly lower than those of the surgical cohort in that trial, whose rates of complete heart block were 3.6% at 30 days, 5.0% at 1 year, and 6.4% at 2 years. In the CoreValve series,1 the incidence of heart block was even higher (19.2%-42.5%). In our series, only 2 patients (11%) required a permanent pacemaker.
A major pitfall of TAVR in all series and trials published so far has been paravalvular leak.1 This complication is common because patients undergoing TAVR typically have heavy calcification and difficult aortic root anatomy. The risk of paravalvular leak is compounded by the fact that in the United States the SAPIEN valve is available for clinical use in only 2 sizes (23 and 26), making it difficult to obtain an appropriate fit. In the TAVR arm of the PARTNER trial,8 the rates of moderate to severe paravalvular leak were 12.2% at 30 days and 6.8% at 1 year compared with 0.9% at 30 days and 1.9% at 1 year in the open surgical arm. Even mild paravalvular leak was associated with increased 2-year mortality in the PARTNER trial. In our series, we had 3 moderate perivalvular leaks (16%) and 7 trace to mild perivalvular leaks (37%).
Another complication that is unique to the transfemoral delivery system is groin wound problems, which we saw in 2 patients (11%). In the future, the development of better and lower-profile sheaths and delivery systems may make this complication less common by allowing truly percutaneous approaches and obviating surgical cutdown.1,24
The mean (SD) length of hospital stay in our patients undergoing TAVR was longer than expected at 8.0 (5.9) days (range, 2-26 days). Because of advanced age and frailty, many patients required postoperative rehabilitation, and our center implements an aggressive early physical therapy program as soon as patients are clinically ready. Most of our patients completed this process on the surgical floor before a rehabilitation or long-term care bed became available. In addition, our hospital is a referral center for the entire VA system; accordingly, only 4 of our patients (21%) were from Houston, whereas the rest were from other parts of Texas or from other states. This resulted in longer hospital stays because of delays in arranging transportation and social work issues regarding long-distance travel.
Careful patient selection for TAVR is important. In addition to the 19 patients who underwent TAVR, 15 other patients were referred for TAVR, but they were deemed unsuitable by the heart valve team for various physiologic and anatomic reasons. Additional patients were offered SAVR instead. As technology improves, TAVR will doubtless be offered to a larger pool of patients. The biggest hurdles in expanding the applicability of TAVR to other, lower-risk patients will be the issues of paravalvular leak and long-term durability. Two-year data from randomized trials and 5-year data from registries show acceptable durability.1 A study25 conducted in Vancouver, British Columbia, Canada, revealed signs of prosthetic valve failure (ie, moderate AS) in 3.5% of patients at 5 years. Five-year survival was only 35%, which could be related to the generally poor health of the patients to whom TAVR was offered.
In the PARTNER trial,26 the cumulative 1-year costs remained higher for inoperable patients treated with TAVR vs standard medical therapy ($106 076 vs $53 621), although the follow-up costs through 12 months were lower with TAVR ($29 289 vs $53 621) because of reduced rehospitalization rates. The incremental cost-effectiveness ratio for TAVR was estimated at $50 200 per year of life gained, or $61 889 per quality-adjusted life-year gained.
Compared with SAVR, TAVR provided a modest benefit in quality-adjusted life expectancy and only slightly reduced costs compared with AV replacement in the PARTNER trial.27 The mean (SD) total admission costs were $73 219 ($40 596) for transfemoral TVAR vs $74 067 ($47 422) for SAVR.
Before launching TAVR at our facility, we had to fee base TAVR candidates to other non-VA TAVR programs at a cost of $85 000 per procedure.9 Our VA spends $32 000 per device, a cost differential that translates into significant savings.
Although we did not formally evaluate patient retention or referrals as part of our study, we believe that the TAVR program retains patients and boosts referrals within the VA system. In addition to saving the VA some revenue, the halo effect associated with the TAVR program benefited the cardiovascular services and other ancillary services in terms of allocation of resources and personnel.
Some patients have a limited life expectancy regardless of their AS and should never undergo TAVR. Other patients might experience a minor complication during TAVR, such as a ventricular wire perforation or access complication, which could be readily salvaged by an open operation, and there are complications that are drastic and are unlikely to be salvageable. From an ethical standpoint, our philosophy is to handle the decision-making process on a case-by-case basis. The value of rigorous patient selection by the multidisciplinary heart valve team, along with candid informed consent, cannot be overemphasized. Realistic expectations relating to the procedure, including the management of potential complications, are laid out up front. We believe that tests for frailty and the Society of Thoracic Surgeons predicted risk of mortality score and even the surgeon’s assessment are imperfect tools in predicting outcomes. In a recent review article from the Mayo Clinic of 119 patients with transfemoral TAVR, 5 experienced acute catastrophic complications.28 Four of 5 underwent a salvage operation, of whom 1 died and 3 survived. Hence, there is a growing awareness that no definite line exists beyond which a patient becomes inoperable; rather, a continuum of risk is subjective in most cases.
A final word of caution is that outside the United States there is a growing trend of offering TAVR to younger, healthier patients because of unrealistic fears of SAVR. Excellent results are shown with SAVR, including proved durability and, most important, a low risk of PVL.1 Therefore, it is important not to indiscriminately expand the indications for TAVR until at least comparable mid-term and long-term outcomes are achieved. In a recent study29 from Germany, SAVR offered to high-risk TAVR candidates had excellent results, including a hospital mortality of only 1.9%. Likewise, our open AV replacement data from the same year show no 30-day mortality and low morbidity.
This study has the limitations that are typical of retrospective studies. However, the data were collected prospectively, and follow-up analysis was complete. These data are limited to one VA center; future studies should investigate whether our experience is generalizable to other VA facilities. Our study is unique in describing the early experience with TAVR at a VA facility with an in-house TAVR program.
In summary, TAVR can produce excellent results when offered at a VA facility. Thoughtful expansion of TAVR to other qualified VA facilities will have important implications for veterans’ access to care and may improve patient outcomes.
Accepted for Publication: June 25, 2013.
Corresponding Author: Shuab Omer, MD, Michael E. DeBakey Veterans Affairs Medical Center, Mail Code 112/Operative Care Line, 2002 Holcombe Blvd, Houston, TX 77030 (firstname.lastname@example.org).
Published Online: September 18, 2013. doi:10.1001/jamasurg.2013.3743.
Author Contributions:Study concept and design: Omer, Kar, Cornwell, Ali, Paniagua, Atluri, Kougias, Preventza, Carabello, Bakaeen.
Acquisition of data: Omer, Kar, Cornwell, Blaustein, Levine, Ali, Jneid, Paniagua, Bechara, Ruma, Bakaeen.
Analysis and interpretation of data: Omer, Kar, Cornwell. Ali, Paniagua, Bozkurt, Bakaeen.
Drafting of the manuscript: Omer, Cornwell, Ali, Jneid, Atluri, Kougias, Carabello, Bakaeen.
Critical revision of the manuscript for important intellectual content: Omer, Kar, Cornwell, Blaustein, Levine, Jneid, Paniagua, Bechara, Ruma, Preventza, Bozkurt, Bakaeen.
Statistical analysis: Kar, Cornwell, Bozkurt.
Obtained funding: Bakaeen.
Administrative, technical, or material support: Kar, Cornwell, Blaustein, Levine, Ali, Jneid, Paniagua, Atluri, Bechara, Kougias, Ruma, Carabello.
Study supervision: Omer, Cornwell, Blaustein, Ali, Jneid, Atluri, Preventza, Bakaeen.
Conflict of Interest Disclosures: None reported.
Previous Presentations: Presented at the 37th Annual Surgical Symposium of the Association of VA Surgeons; April 23, 2013; Milwaukee, Wisconsin.
Additional Contributions: Stephen N. Palmer, PhD, ELS, contributed to the editing of the manuscript.