Author + information
- Jean-Michel Paradis, MD and
- Josep Rodés-Cabau, MD∗ ()
- ↵∗Reprint requests and correspondence:
Dr. Josep Rodés-Cabau, Quebec Heart & Lung Institute, Laval University, 2725 Chemin Ste-Foy, G1V 4G5 Quebec City, Quebec, Canada.
- aortic stenosis
- GARY Registry
- transcatheter aortic valve implantation
- transcatheter aortic valve replacement
Although randomized trials have been pivotal in establishing transcatheter aortic valve replacement (TAVR) as a treatment strategy for patients with severe aortic stenosis and high surgical risk (1–3), several registries have also featured prominently in consolidating this technology. Registry-based data have inherent limitations, including data accuracy and completeness, potential selection bias due to nonrandomized allocation of interventions and devices, frequent lack of data monitoring and endpoint adjudication, and the presence of unmeasured confounders. Nevertheless, registries allow the comparison of large “real-world” patient populations, permit post-market approval device surveillance, minimize bias in reporting individual center and operator outcomes, and could eventually lead to an international network of registries for global device surveillance.
The TAVR field comprises several important industry-driven registries, although this type of registry generates data with an inherent selection bias. Alternatively, registries with mandatory data collection on a national scale utilizing an all-comers approach would ultimately provide the most reliable data on both patient characteristics and real-world results. Such was the case with both the FRANCE 2 (French Aortic National CoreValve and Edwards) and TVT (Transcatheter Valve Therapies) registries (4,5), which helped pioneer an all-comers approach fulfilling the requirements of the French Ministry of Health and the Centers for Medicare & Medicaid Services, respectively.
The German Aortic Valve Registry (GARY), founded in July 2010, was assembled as a complete survey for all invasive therapies in patients with significant aortic valve diseases (6). Surgical aortic valve replacement (SAVR) and new treatment options such as TAVR are being monitored. Importantly, data acquisition is based on 3 sources: 1) the mandatory German database for external performance measurement; 2) a specifically-developed registry dataset; and 3) a follow-up data sheet, which was filled in by telephone interview (6). Apart from the 30-day clinical status, all other follow-up evaluations (1, 3, and 5 years) are accomplished independently by the Institute for Quality and Patient Safety.
In this issue of the Journal, Walther et al. (7) report the acute results and periprocedural complications of 15,964 patients included in GARY who underwent TAVR in 88 German centers between 2011 and 2013 (7). Notably, GARY not only provides solid evidence on TAVR results from an all-comers and real-world perspective, it also represents the largest published TAVR registry to date. Similar to prior registries, the mean age of patients was >80 years; however, the median Society of Thoracic Surgeons (STS) risk score was 5%, much lower than the risk profile reported in most previous TAVR studies including randomized trials (1–3,8) as well as the all-comers FRANCE 2 and TVT registries (4,5). Furthermore, patient risk progressively decreased over time, highlighting the shift in Germany toward treating lower surgical-risk patients with TAVR. This specific feature of GARY exemplifies the likely worldwide shift of TAVR in the ensuing years. However, further confirmation of TAVR’s noninferiority compared with SAVR in lower-risk populations is mandatory, and there are randomized trials in the recruitment stage that will certainly help answer this unresolved question. Meanwhile, the recently-reported results from a prospective randomized trial showed lack of mortality and stroke differences at 1-year follow-up between TAVR and SAVR in 280 low-risk patients (mean STS: 3%) (9). These promising results, in addition to 5-year data from the PARTNER (Placement of Aortic Transcatheter Valves) trial showing absence of structural failures of the transcatheter and surgical valves (10), sets the scene for the likely expansion of TAVR toward treating both lower-risk and younger patients.
The work of Walther et al. (7) focused on in-hospital outcomes post-TAVR. Unfortunately, the authors failed to report the (more standard) 30-day data, making it somewhat difficult to compare with most prior TAVR studies. Most patients were treated transfemorally (71%), with the transapical approach used in 27% of patients. About one-half of the patients received a balloon-expandable valve and one-half received a self-expanding valve system. The in-hospital mortality rate of 5.2% was close to the pre-operative risk determined by the STS score (median: 5%) or the German Aortic Valve score (median: 5.6%) but was much lower than the logistic EuroSCORE median (18.3%). GARY adds to the widely demonstrated outcome that the logistic EuroSCORE significantly overestimates the procedural risk of TAVR candidates. The acute mortality reported in GARY compares favorably with prior all-comer registries including higher-risk patients like FRANCE-2 (mean STS: 14%; 30-day mortality: 9.7%) and TVT (median STS: 7.1%; 30-day mortality: 7.0%) (5,6). Up to 20 baseline and periprocedural variables (mostly procedural complications) were associated with in-hospital mortality. Hence, these data essentially confirm what has already been reported in numerous prior TAVR studies (8). Interestingly, and unlike other registries (4,5), the transapical approach was not an independent predictor of mortality in GARY.
The in-hospital stroke rate (1.5%) was lower than those reported in randomized trials including neurological assessment both before and following TAVR (between 4.9% and 6.7%) (1–3). This might be explained by GARY’s lower-risk population, the learning curve effect, or technology advances reducing valve embolization rates, need for a second valve, and so on. Additionally, it would have been interesting to know the rate of embolic protection device use (if any) during TAVR. This lower stroke rate aligns with other contemporary reports of periprocedural cerebrovascular event rates <3% (5,9).
Moderate or severe aortic regurgitation (AR) has been identified as one of TAVR’s most important drawbacks (1–3,8). Importantly, moderate or severe AR was reported in <6% of GARY cases, almost one-half the rate reported in prior studies. Although this finding should be interpreted with caution due to the high variability in AR evaluation in the absence of a central core laboratory, this lower rate of AR reflects the current trend in the TAVR field, which is likely related to better transcatheter valve sizing and positioning, as well as the use of newer-generation devices with enhanced antiparavalvular leak properties (8), although the rate of use of such devices was not provided.
To further evaluate procedural safety, Walther et al. (7) created 2 novel composite endpoints for reporting periprocedural complications: 1) severe vital complications (SVC) including death on the day of intervention, conversion to sternotomy, acute percutaneous coronary intervention, low cardiac output requiring treatment, aortic dissection, or annular rupture; and 2) technical complications (TCO) including the repositioning or retrieval of the valve prosthesis, valve-in-valve procedure, valve embolization, or paravalvular leak closure. SVC and TCO occurred in 5.0% and 4.7% of patients, respectively, and both categories of complications decreased significantly over time. Again, this favorable tendency could be explained by technical progress (retrievable valves, antiparavalvular leak properties), constant learning, improved screening, and treating progressively lower-risk patients. Of note, the incidence of complications requiring urgent sternotomy was 1.3%, with >50% of patients surviving such complications. This underscores the importance of performing these procedures in centers with surgical facilities.
Although these novel combined endpoints (SVC and TCO) help provide a rapid overview of procedural safety, they have several drawbacks. First, they are vastly different from the consensus endpoint definitions established by the Valve Academic Research Consortium (11). Although original, this new paradigm of outcomes assessment may render comprehensive evaluation of GARY rather complex and make the comparison of interstudy results extremely challenging. Second, identifying the predictors of life-threatening complications is key for implementing appropriate preventive measures. The authors identified the following independent predictors of SVC: female sex, New York Heart Association functional class IV heart failure, ejection fraction ≤30%, pre-operative inotropic medication, peripheral vascular disease, higher degree of valve calcification, and grade IV aortic stenosis. For TCO, the independent predictors were male sex, ejection fraction ≤30%, neurological dysfunction, cardiac decompensation, intravenous inotrope therapy, and year of the procedure. However, the complications included in each combined endpoint are very diverse, raising questions as to the validity and clinical application of these data. For example, it is well known that the predictors of mechanical complications such as coronary obstruction or annulus rupture are inherently different from those of low cardiac output syndrome. Not surprisingly, the authors failed to find any predictor for conversion to sternotomy. Again, because this complication can be related to a myriad of causes (ventricular perforation, coronary obstruction, annulus rupture, and so on), each one of them should probably be analyzed separately to ascertain clinically-meaningful predictors. In our opinion, GARY represents a unique opportunity to analyze the factors associated with each individual life-threatening or technical complication of TAVR. This would generate more clinically-meaningful data that could be applied when evaluating TAVR candidates as well as for pre-operative planning.
Compared with other industrialized nations, TAVR has expanded most rapidly in Germany. With the data from GARY, it is reassuring that TAVR’s growth appears safe. The heart team approach, involving the collaborative work of cardiologists, cardiac surgeons, anesthesiologists, and imaging specialists, remains fundamental for optimizing clinical outcomes. GARY highlights the ability to monitor the evolution of risk profile, procedural success, and acute complications in a vast number of patients. This may allow us to detect, among others, changes in patients' characteristics and risk profile as well as major differences between valve types, approaches, or other specifics relating to individual treatment centers that could permit corrective actions on the basis of unanticipated negative outcomes and treatment gaps. Nonetheless, in the burgeoning TAVR arena, standardizing clinical research is essential. We strongly encourage international societies to utilize the accepted Valve Academic Research Consortium definitions for improving comparability and interpretability of the ever-expanding TAVR data bank.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. Rodés-Cabau has received research grants from Edwards Lifesciences and St. Jude Medical. Dr. Paradis has reported that he had no relationships relevant to the contents of this paper to disclose.
- American College of Cardiology Foundation
- Walther T.,
- Hamm C.W.,
- Schuler G.,
- et al.
- Thyregod H.G.H.,
- Steinbrüchel D.A.,
- Ihlemann N.,
- et al.
- Mack M.J.,
- Leon M.B.,
- Smith C.R.,
- et al.
- Kappetein A.P.,
- Head S.J.,
- Généreux P.,
- et al.