Transcatheter Aortic Valve Replacement

Since the first balloon-expandable transcatheter aortic valve replacement (TAVR) in 2002 by Cribier and colleagues¹ and the first self-expanding TAVR by Grube and associates in 2004,² TAVR has grown rapidly to more than 50,000 implantations worldwide. This rapid expansion has occurred despite relatively small feasibility studies, case series, and registries. Few of these early trials had independent clinical-event committees or core labs. The data, however, were encouraging and TAVR for nonoperative and high-risk patients spread rapidly. Approval in the United States, though, required more rigorous data. Currently, there are only 4 prospective randomized trials of TAVR. These are the Placement of Aortic Transcatheter Valve (PARTNER) trial, the Medtronic CoreValve® U.S. Pivotal Investigational Device Exemption (IDE) trial (Medtronic Inc.; Minneapolis, Minn), the PARTNER II trial, and the Safety and Efficacy Study of the Medtronic CoreValve® System in the Treatment of Severe, Symptomatic Aortic Stenosis in Intermediate Risk Subjects Who Need Aortic Valve Replacement (SurTAVI) trial. The PARTNER trial has been completed and the results published.^3,4 The PARTNER trial data have led to U.S. Food and Drug Administration (FDA) approval of the SAPIEN balloon-expandable valve (Edwards Lifesciences Corporation; Irvine, Calif) for nonoperative and high-risk patients. This is the only currently approved TAVR valve in the U.S. The CoreValve U.S. IDE Pivotal Trial was completed in 2012, and initial publication is likely in late 2013, after presentation at the Transcatheter Therapy meeting. The PARTNER II and SurTAVI trials are still actively recruiting at this point.

The PARTNER trial had 2 arms. The first consisted of patients deemed by 2 experienced cardiac surgeons to be inoperable. This group was randomized 1:1 between TAVR with the SAPIEN valve and best medical therapy, which could include balloon aortic valvuloplasty; it was designed as a superiority trial.³ The 1-year all-cause mortality rate was 50.8% for the best-medical-therapy group and 30.7% for the TAVR group, for an absolute survival difference of 20.1% or a number-to-treat (to save a life) of 5. The hemodynamic improvement of the TAVR group appeared to be excellent: the pre-TAVR mean aortic valve gradient of 44.2 mmHg had decreased to 10.2 mmHg at 1 year post procedure. The pre-TAVR mean aortic valve area of 0.6 cm² had increased to a mean of 1.5 cm² at 1 year. The study was powered for a primary endpoint of all-cause death and major stroke at 12 months, but the major question raised by this study was the rate of neurologic events in total. There was no statistical difference between TAVR and best medical therapy in major stroke, minor stroke, or transient ischemic attack, when those categories were considered individually. When they were taken as a combined endpoint, however, a statistically significant difference emerged, both at 30 days and at 1 year: 12 (6.7%) for TAVR vs 3 (1.7%) for best medical therapy; and 19 (10.6%) for TAVR vs 8 (4.5%) for best medical therapy. The issue with neurologic events was thoughtfully highlighted in an editorial by Schaff,⁵ but the large survival difference led to FDA approval of the SAPIEN valve for nonoperative patients. The 2-year data have since been published,⁶ and the 3-year data were presented by Samir Kapadia at the 2012 Transcatheter Therapy meeting. The survival difference at 3 years had grown to 26.8%, with the number-to-treat falling to only 3.7 patients; the hemodynamic results remained excellent. For inoperative or extreme-risk patients, TAVR is now generally accepted. The greatest challenge going forward lies in defining the point at which relief of aortic stenosis does not substantially benefit the patient as a whole, thereby rendering TAVR a futile endeavor.

The 2nd arm of PARTNER was a noninferiority trial for patients who were operable but considered to be at very high risk. This was generally defined as a potential operative mortality rate of >15%, or about the upper 10% of patients who currently undergo operation in the U.S. This mortality risk was determined from the Society of Thoracic Surgeons Predicted Risk of Mortality (STS-PROM) score and site-surgeon determination.

To move TAVR into the area now occupied by surgical AVR, which has a 60-year history and well-defined short- and long-term outcomes, I would want to see better or equivalent results for mortality rates, hemodynamic status, quality of life, patient acceptance, morbidity, and durability. One-year data for the PARTNER Cohort A trial, comparing AVR to TAVR, have been published.⁴ That study was powered for a primary endpoint of all-cause mortality at 12 months. No significant difference was seen between 1-year mortality rates for AVR versus TAVR (26.8% vs 24.3%). Hemodynamic outcomes were good for both, with TAVR being slightly but not significantly better. On the Kansas City Cardiomyopathy Questionnaire, both AVR and TAVR showed a large and equivalent improvement of over 20 points at 1 year. At this point, TAVR has been shown to be noninferior to AVR, and its patient acceptance is clear to anyone in this field, without supporting data. Patients will always prefer a catheter-based therapy to open surgery when that is a possible and reasonable alternative. The major areas of concern for morbidity were stroke and paravalvular leak (PVL). Stroke and transient ischemic attack rates for surgical AVR versus TAVR were not significantly different when considered alone (10 [3.2%] vs 20 [6%] and 4 [1.5%] vs 7 (2.6%]). Although the study was not powered to determine a combined neurologic event rate, a significant difference did emerge for surgery versus TAVR if neurologic events were taken as a combined endpoint (13 [4.3%] vs 28 [8.7%]; P = 0.03).

If one considers that what a patient really wants is to be alive without having experienced a stroke, that combined endpoint yields no difference between surgery and TAVR (95 [28.6%] vs 95 [27.4%]). The largest issue for TAVR appears to be PVL, because over 60% of patients in the PARTNER Cohort A trial had mild, moderate, or severe PVL (52.1%, 11.4%, and 0.7%, respectively) at 30 days.⁴ In the 2-year data, even mild PVL was a marker for increased mortality rates, and this was a cause of great concern.⁷ In the nonrandomized continued-access data for this group, a larger cohort of patients showed that moderate or severe (but not mild or less) PVL was a marker for death. This has also been corroborated by the CoreValve Advance trial and by the 5-year data from Vancouver.⁸ Nonetheless, mild PVL remains a concern and is one of the greatest challenges in extending this therapy to lower-risk patients. For other sequelae, there were tradeoffs: myocardial infarction was greater in AVR, major vascular injury was greater in TAVR, major bleeding was greater in AVR, and there were no differences in endocarditis, renal failure, or need for pacemakers.

Finally, we come to the issue of durability. For the true answer on durability, only time will tell, but we can make several comments. We know that the older a patient is when a biologic valve is implanted, the slower that valve degenerates.^9,10 We also know from the STS database that among almost 146,000 isolated AVRs performed on patients >65 years of age, the average post-AVR survival time for a patient >80 years old is only 6 years.¹¹ Because the average age in both PARTNER groups is over 80 years, the needed durability is somewhat limited. This compendium of data led to FDA approval of the SAPIEN device for TAVR in high-risk patients.

Until the PARTNER trial, no randomized trial for the treatment of aortic stenosis with AVR or TAVR was available. A second randomized trial, the CoreValve U.S. IDE Pivotal trial, has been completed and is currently under analysis. Those results will provide the medical community with unprecedented knowledge in this area. The CoreValve IDE U.S. Pivotal trial was designed in a manner similar to the PARTNER trial, but it differs in the inoperable extreme-risk group. By the time that this trial was ready to launch, the PARTNER Cohort B data showed a substantial survival difference between TAVR and best medical therapy. This unquestionably large survival difference rendered randomization to medical therapy unethical. Therefore, all patients in the extreme-risk arm of this trial received TAVR, and their results will be compared with those of the PARTNER Cohort B TAVR group with a primary endpoint of all-cause mortality and stroke at 1 year. This arm of the trial closed in January 2012, and 1-year data will be presented at the 2013 Transcatheter Therapy meeting. The high-risk arm of the CoreValve trial was substantially the same in design as that of the PARTNER Cohort A trial, in which high-risk patients with an estimated operative mortality rate of >15% had been randomized to AVR or TAVR, with an endpoint of all-cause mortality at 1 year. The CoreValve trial also used the STS-PROM and site-surgeon determination to establish an operative risk of >15%. In addition, multiple comorbidities that are not currently included in the STS-PROM and frailty data were collected. This arm of the trial closed in August 2012 and will be analyzed at the 1-year point. Two randomized trials for inoperable and high-risk patients with symptomatic severe aortic stenosis will provide data of a quality not yet seen.

Transcatheter aortic valve replacement for extreme-risk and high-risk patients with symptomatic severe aortic stenosis is now commercially available in the U.S., if one uses the SAPIEN valve for patients who have the characteristics from the PARTNER trials. Transcatheter aortic valve replacement already has been well accepted worldwide, but does that mean we should begin including intermediate-risk patients? Surgical AVR remains the gold standard for that risk group, but data from Europe suggest that there might be equipoise between AVR and TAVR. The CoreValve Advance trial enrolled 1,015 patients in 44 experienced centers in Europe, Asia, and South America. In this trial, 100% of patients were monitored, all primary endpoints were adjudicated by an independent clinical-events committee, all cerebrovascular events were adjudicated by an independent neurologist, and a core laboratory was used for all electrocardiograms and procedural angiograms. A EuroSCORE of 10 to 20, which corresponds to intermediate risk, was present in 40% of the patients. The 30-day all-cause mortality rate was 4.5% and the stroke rate was 2.9%. The preprocedural mean gradient of 45.6 mmHg decreased to 9.5 mmHg at 6 months. The preprocedural aortic valve area of 0.7 cm² rose to 1.7 cm² at 6 months. These data were presented at the 2012 American College of Cardiology meeting by Alex Linke but have not yet been published.

The Bern–Munich–Rotterdam (BERMUDA) study looked at 510 propensity-matched patients with an STS score of 3 to 8, who had been divided into groups of 255 TAVR and 255 AVR patients. In this study, there was no difference in 30-day or 1-year mortality rates between TAVR and AVR. A similar propensity-matched intermediate-risk study was published by Latib and colleagues.¹² That study matched 111 pairs with an average STS-PROM of 4.57% for TAVR and 4.6% for AVR. The 30-day and 1-year mortality rates for TAVR versus AVR were 2 (1.8%) versus 2 (1.8%) and 7 (6.4%) versus 9 (8.1%). The 30-day and 1-year cerebrovascular-event rates for TAVR versus AVR were 4 (3.7%) versus 9 (8.1%) and 5 (4.6%) versus 10 (9.1%). None of these reached statistical significance, which clearly showed parity between TAVR and AVR.

These preliminary data were the impetus for 2 randomized trials in the intermediate-risk population, the PARTNER II A and the SurTAVI trial. These trials have a similar design. Both look at intermediate-risk patients, defined as those with an STS-PROM of 4% to 10%. Both allow treatment of coronary artery disease along with aortic stenosis. Patients with an STS-PROM of 4% to 10% will be separated into a need-to-treat coronary artery disease arm or not. These 2 arms will then be randomized into catheter-based treatment (TAVR alone or TAVR plus percutaneous coronary intervention) or surgical treatment (AVR or AVR plus coronary artery bypass grafting). Both trials will use independent clinical-event committees, neurologic evaluation of neurologic issues, and core laboratories for echocardiograms. Both will be large trials that require more than 2,000 patients each. Both trials are now actively recruiting. All-cause mortality and stroke rates at 24 months will serve as the primary noninferiority endpoint. Although much of the community outside of the U.S. is already performing TAVR for intermediate risk, it is unlikely that we will ever get a reasonable answer on what should be done in this risk group without such randomized trials. It is my opinion that enough equipoise exists to test TAVR in the intermediate-risk patient for noninferiority with AVR, but this should be done only within the context of a clinical trial, given our current state of knowledge.

The technology of TAVR is still early and refinements are already occurring. The SAPIEN valve has evolved into the SAPIEN XT (Edwards Lifesciences), using a cobalt–chromium rather than a stainless-steel frame. This has enabled transition to an 18F delivery system. The SAPIEN 3 (Edwards Lifesciences), with an inflow skirt or parachute to help control PVL, is also being tested. One of the goals of new devices and device iteration is the ability to recapture, reposition, and redeploy the valve, which is possible only in a self-expanding valve. It is not surprising that almost all new devices are self-expanding in design. The CoreValve has evolved into the Evolute valve (Medtronic), with a re-engineered nitinol frame that leads to more constant radial force over the planned expansion range and a re-engineered shape to aid in PVL sealing.

This valve will also evolve into a fully recapturable, repositionable, and redeployable valve. Edwards Lifesciences has developed and will begin testing its own self-expanding valve, the Centura. The Direct Flow Medical® Transcatheter Aortic Valve System (Direct Flow Medical, Inc.; Santa Rosa, Calif) is being tested outside the U.S. and has a novel design with no metal at all. The valve consists of Dacron tubes that house a pericardial valve. The Dacron tubes are inflated with a saline–contrast mixture to seat the valve. If the operator is pleased with the seating of the valve, the saline mixture is exchanged for a rapidly setting polymer, which makes the tubular frame permanently solid. Another transfemoral valve being tested outside the U.S. and likely to enter U.S. trials soon is the Lotus™ Valve System (Sadra Medical, a unit of Boston Scientific Corporation; Los Gatos, Calif). Its unique design is a single nitinol wire frame with a pericardial valve. The frame is mechanically shortened, which causes it to expand into place. If the operator is not satisfied, the valve can be lengthened, shortened, or moved to a new location or, if satisfied, released. This valve also has an adaptive skirt to help mitigate PVL. The Portico™ Transcatheter Aortic Heart Valve (St. Jude Medical, Inc.; St. Paul, Minn) is a self-expanding, intra-annular valve that will be repositionable, recapturable, and redeployable; it is due to start U.S. trials soon. Three self-expanding transapical valves are in use outside of the U.S. and likely to see U.S. trials at some point in the future. They are the JenaValve™ (JenaValve Technology GmbH; Munich, Germany), the Ventor Embracer™ Heart Valve (Medtronic), and the Acurate™ Aortic Valve Replacement System (Symetis SA; Ecublens, Switzerland). Transcatheter aortic valve replacement is also being tested in new indications, such as valve-in-valve for degenerated bioprostheses.

In summary, it should be remembered that TAVR is only a decade old and that future advances are likely to mitigate many of the current problems. Transcatheter aortic valve replacement is well accepted in extreme-risk and high-risk patients with symptomatic severe aortic stenosis. The procedure is likely to move into intermediate-risk patients, but at present this should be done only in randomized trials.