https://orcid.org/0000-0001-8237-7095
Quality metrics of TAVI programmes. Left panel: volume as quality indicator. Middle panel: additional quality indicators. Right panel: alternative outcomes.
This editorial refers to ‘Transfemoral aortic valve implantation: procedural hospital volume and mortality in Germany’, by K. Bestehorn et al., https://doi.org/10.1093/eurheartj/ehac698.
The association between hospital volume and clinical outcomes has been well established in the field of surgical aortic valve replacement (SAVR).1 The so-called ‘volume–outcome effect’ has also been shown in transcatheter aortic valve implantation (TAVI) procedures,2–4 and this has prompted many national cardiology and cardiac surgery societies to establish a minimum annual number of TAVI procedures (ranging from 20 to 100 per centre/year) as a measure of quality of hospital TAVI programmes, with potential implications regarding programme reimbursement and maintenance. To date, most data on TAVI volume–outcome effects have been obtained in the era of older generation transcatheter heart valve (THV) systems and included a significant proportion of high to prohibitive surgical risk candidates. However, the TAVI field has experienced major developments in recent years, with continuous iterations of THVs along with the simplification of the procedures and post-procedural management, and the inclusion of lower risk patients. This has translated into significant improvements in clinical outcomes over time, including a major decrease in peri-procedural complications and 30-day mortality.5 Thus, the volume–outcome relationship as well as the optimal number of annual procedures remains to be determined in the current TAVI era.
In this issue of the European Heart Journal, Bestehorn et al.6 evaluated the association between transfemoral TAVI hospital volume and in-hospital mortality in 2018 (81 centres) and 2019 (82 centres) in Germany. Also, the observed/expected (O/E) mortality ratio was used as a marker of quality of the TAVI programme (with >95th percentile O/E of all hospitals considered as not acceptable quality). Expected mortality was calculated based on two different and pre-defined scores, the German AKL/KATH score (O/E ratio) and the Euroscore II (O/E2 ratio). The study involved 40 900 patients, with a moderate risk profile (mean EuroSCORE II of 6.9% in 2018, and of 6.5% in 2019), and an average in-hospital mortality rate of 2.6% and 2.4% in 2018 and 2019, respectively. Hospital TAVI volume presented a significant but mild correlation with in-hospital mortality. When analysing the correlation between hospital volume and observed/expected mortality, this correlation was non-significant for both O/E and O/E2. Furthermore, when dividing hospitals into quartiles according to annual TAVI volume, no differences were found regarding in-hospital mortality, risk-adjusted mortality rate, O/E, and O/E2 between the first quartile (low-volume centres) and the fourth quartile (high-volume centres). In the same direction, there were no significant differences in major peri-procedural complications between low- and high-volume centres.
The results from Bestehorn et al. would challenge the use of hospital volume as a relevant marker of programme quality, and would invite reconsideration of the minimal annual volume requirements for programme competency maintenance. However, several aspects of the work of Besthorn et al. merit further consideration (Graphical Abstract).
In Germany, the mean yearly number of TAVI procedures (2018–2019) per hospital was >230, with the majority of centres performing >100 procedures/year. Also, among the hospitals in the lowest quartile, the mean number of procedures/year was still >80. These data contrast with most previous studies evaluating the volume–outcome relationship in TAVI, where the centres in the lowest tercile/quartile performed <50 cases/year, and high-volume centers were those performing >75–100 procedures/year.2–4 This would probably preclude generalization of the results of the German experience reported in Besthorn’s work, and highlights the importance of developing TAVI quality metrics according to the reality and experience of each nation/territory.
While the results were adjusted by patients’ surgical and TAVI risk, the fact that available data and risk adjustment were based on hospital aggregated (not patient-level) data was an important limitation. Also, a multiplicity of factors not included in the risk evaluation may have had an influence on TAVI outcomes. First, patient selection bias by anatomical features and procedural complexity. Low-volume centres may tend to select patients with fewer anatomical risk features such as extremely calcified valves, extreme aortic angle, risk of coronary obstruction or annular rupture, need for additional techniques such as basilica or coronary protection, suboptimal transfemoral access, etc. Second, on top of hospital volume, the number of operators (operator yearly volume) along with operators’ experience (a low-volume centre may have operators with a large prior experience) may have an impact on peri-procedural results.7 Third, experience with alternative accesses (probably more frequent in high-volume centres) may also influence the heart team decision in patients with unfavourable femoral access. The fact that non-transfemoral TAVI cases were excluded from the study of Bestehorn et al. may have provided a biased view of the overall quality of the TAVI programmes. Fourth, the number of THV types should probably be a factor to consider when evaluating volume–outcome effects. Each valve system requires specific skills and a learning curve process, which may be different and remains to be determined for most THV systems.8 While using a tailored approach with multiple THV systems may be associated with improved outcomes in high-volume centres, a multiplicity of valve systems may have a deleterious effect on intermediate- to low-volume centres. Finally, some studies have shown a positive impact of a high-volume cardiac surgery activity on TAVI outcomes.9 While TAVI and SAVR volumes are usually concordant, this may not be the case in some centres and, in some cases (particularly low-volume centres), it may be worth adjusting for this variable in order to obtain a better idea of the TAVI programme quality. In other words, in low-volume TAVI centres, the results in those with no or low cardiac surgery activity may differ from those with a higher surgical volume.
In-hospital or 30-day mortality rates have been the most common outcomes used for evaluating TAVI volume–outcome effects, but recent studies have shown that including other performance metrics such as 30-day rehospitalization or the number of days out of hospital may be more appropriate to determine the quality of a TAVI programme.10,11 Also, including complications with potential permanent sequelae (e.g. in-hospital stroke or permanent pacemaker) may also be important to refine quality programme evaluations.12 The significant reduction in TAVI mortality rates across all patients’ risk profiles makes it more difficult to detect performance differences amongst TAVI programmes, and extremely large cohorts may be needed to detect differences, if any, between centres in the future. Thus, an earlier hospital discharge without increasing the risk of death, permanent sequelae complications, or rehospitalization within the weeks following the procedure may best determine the performance of a TAVI centre. Indeed, the post-procedural hospital stay in German centres in 2018–2019 was of 9 days as a mean, which may not reflect current clinical practice in many other countries (e.g. mean hospital stay of 2 days in the TVT registry).13
In conclusion, the work of Bestehorn et al. showed that the TAVI volume/mortality relationship seems to be attenuated in the current era, particularly among centres with TAVI numbers close to or beyond three digits. The confirmation of these data in other countries with very different TAVI penetration levels and healthcare systems may reduce waiting times and facilitate the access of the population to this technology by having more TAVI programmes available. However, technology maturity and lowering patients’ risk has translated into a major decrease in peri-procedural mortality rates, and more sophisticated performance metrics should probably be used for evaluating the quality of TAVI programmes in the future.
Acknowledgements
J.R.-C. holds the Research Chair ‘Fondation Famille Jacques Larivière’ for the Development of Structural Heart Disease Interventions (Laval University, Quebec City, Canada). J.N. was supported by a grant from the Foundation Alfonso Martin Escudero (Madrid, Spain).
Funding
None.
References
Author notes
The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.
Conflict of interest: J.R.-C. has received institutional research grants from Edwards Lifesciences, Medtronic, and Boston Scientific, and is a consultant for or has received speaker fees from Edwards Lifesciences and Medtronic. J.N. reports no conflict of interest to disclose.
