Temporal Trends in Transcatheter Aortic Valve Replacement in France

FRANCE 2 to FRANCE TAVI

Discussion

The present analysis is the first report of the second national French TAVR monitoring program, with a systematic comparison with its predecessor for evaluating TAVR performance trends across France during a 6-year period (Central Illustration). The chief findings of our analysis are as follows: 1) significant changes in baseline characteristics of patients occurred with an important decrease in estimated surgical risk over time, thus reflecting lower rates of comorbidities within FRANCE TAVI despite the inclusion of older patients; 2) >80% of patients benefited from the “transfemoral first” policy adopted by centers, whereas transapical access declined significantly in favor of alternative procedural access; 3) approximately 50% and 70% of procedures were performed using conscious sedation and without transesophageal guidance, respectively, with sustained procedural success; 4) in-hospital and 30-day mortality rates were significantly lower in FRANCE TAVI (4.4% and 5.4%, respectively); 5) infrequent but life-threatening complications did not decline over time, with a significant increase in cardiac tamponade rates; and 6) PPI rates markedly increased, especially within ESV recipients.

• Download figure
• Open in new tab
• Download powerpoint

Central Illustration

Temporal Trends in TAVR in France

In this comparison of patients from the FRANCE 2 (French Aortic National CoreValve and Edwards 2) (2010 to 2012) and FRANCE TAVI (French Transcatheter Aortic Valve Implantation) (2013 to 2015) registries who underwent transcatheter aortic valve replacement (TAVR), patients in FRANCE TAVI had a lower risk profile and significantly lower mortality rates but significantly higher rates of tamponade and pacemaker implantation. AR = aortic regurgitation; ESV = Edwards SAPIEN valve (Edwards Lifesciences, Irvine, California); LES = logistic European System for Cardiac Operative Risk Evaluation (EuroSCORE); PPI = permanent pacemaker implantation; TEE = transesophageal echocardiography.

Unlike previous reports of large multicenter registries (13,14), we demonstrated highly significant changes in baseline characteristics and risk profile of TAVR recipients across France. The improved patient risk profile at baseline (evidenced by a steep decrease in the logistic EuroSCORE) was perhaps the most striking change. Several reasons could explain this trend. First, the favorable outcomes and sustained midterm hemodynamic performances demonstrated in high-risk patients with both ESV and MCV (16,17) could have encouraged the performance of TAVR in lower-risk patients. Second, operators may have been influenced by the current trends in published reports that highlight predictors of early and late death post-TAVR, as well as factors associated with procedural futility (14,16–19). Indeed, rates of coexisting conditions, such as chronic pulmonary disease, peripheral vascular disease, chronic kidney disease, or left ventricular dysfunction that have been independently linked to an increased mortality rate (12,14,16,18–20), were all lower in FRANCE TAVI. Despite a lack of specificity in the TAVR setting, a logistic EuroSCORE ≥40% is a well-known predictor of higher early and late death (14), and it was uncommon (6%) in FRANCE TAVI. The logistic EuroSCORE, albeit less accurate than the Society of Thoracic Surgeons (STS) scoring system among TAVR recipients, is the most established score in Europe for evaluating operative risk (21). This score is calculated on the basis of death from all cardiac surgical procedures; however, its development cohort mainly included patients undergoing coronary artery bypass grafting. It consists of an online EuroSCORE calculator that includes 17 baseline and procedural variables, which provide a 30-day predicted mortality rate with a cutoff of >20% representing high-risk patients. Similar to the STS score, it imperfectly captures some comorbidities, such as frailty, cirrhosis, porcelain aorta, or hostile chest. Interestingly, the STS score and the logistic EuroSCORE provide similar estimated mortality rates for low-risk patients, whereas the logistic EuroSCORE tends consistently to overestimate the operative risk among high-risk patients (21). Third, early nonrandomized reports demonstrating TAVR results comparable to those of surgical aortic valve replacement and superior to those observed in high-risk patients among lower–surgical risk groups also likely contributed to these patterns in procedural evolution (22,23). Finally, the aging of TAVR recipients could reflect the reluctance of some heart teams to perform operations in octogenarians, even patients with minimal comorbidities; thus the teams adopted a TAVR policy for these patients. Although differing commissioning structures may play a significant role in differences in the evolution of baseline characteristics observed over time from 1 national registry to another (13,14), the inclusion of patients treated in 2015 whose changes were even more pronounced likely accentuated discrepancies between the present report and previous reports from other national registries.

The evolution toward a more simplified TAVR procedure is another major finding of the present report. Although the contribution of approach per se has been debated, transfemoral access seems to have mortality and morbidity benefits that justify its use as a first-line strategy (24). The increasing availability of smaller delivery systems enabled this approach to be used increasingly over time, ultimately in >80% of FRANCE TAVI patients overall. Conversely, we observed a dramatic decline in transapical access, with rates much lower than the ∼25% to 30% reported in previous registries (9,10,14). However, such a trend was also reported by the STS/American College of Cardiology (ACC) transcatheter valve therapy (TVT) registry for procedures performed in 2014 within the United States (13). Future reports should focus on outcomes of newer alternative approaches (e.g., direct aortic, transcarotid).

Local anesthesia provided good clinical outcomes compared with general anesthesia in a propensity-matched analysis of the FRANCE 2 registry (25). Concerns also emerged regarding a higher rate of significant paravalvular leak associated with this technique, which does not allow periprocedural transesophageal guidance. However, despite the widespread use of local anesthesia in FRANCE TAVI, device success was greater, and moderate or severe paravalvular leaks were less frequent than in FRANCE 2. These findings are likely the result of growing experience of operators, refined annulus sizing using 3-dimensional imaging techniques, and the availability of devices with a dedicated sealing skirt. Interestingly, unlike with baseline characteristics, changes in procedural aspects were more pronounced among centers that did not participate in FRANCE 2. As previously demonstrated (26), these centers probably benefited from the global knowledge and shared experience of trained operators at the start of their TAVR program and therefore promptly achieved favorable technical and clinical outcomes.

The reported 4.4% and 5.4% in-hospital and 30-day mortality rates, respectively, in the present study were within the range of recent publications. Reporting outcomes of patients treated between 2011 and 2013, the German Aortic Valve Registry (GARY) (10) and STS/ACC TVT (9) registries demonstrated in-hospital mortality rates of 5.2% and 5.5%, respectively, whereas the 30-day mortality rate was 7.6% in the STS/ACC TVT registry. The gradual decrease in mortality rates demonstrated in the present study also was highlighted in the U.K. TAVR registry during the period from 2007 to 2012 (14). This trend was even more pronounced in 2015 in FRANCE TAVI because the 3.3% 30-day mortality rate achieved during this single year among experienced centers in an all-comers population compared favorably with the 3.9% mortality rate of the selected PARTNER 2 group of patients (7). Stroke rates were low (2.0%), stable over time, and comparable to those reported by the STS/ACC TVT registry, which provided central adjudication of these neurological events (13). The stability of infrequent but potentially life-threatening complications (annular rupture, aortic dissection) and the unexpected increase of cardiac tamponade should, however, be emphasized. Given the rapid expansion of TAVR indications, German centers without on-site cardiac surgery have been allowed to perform transfemoral procedures, provided they have documented cooperation with an external surgical center. An analysis of the German AQUA (Aortic Valve Replacement Quality Assurance) registry (27) demonstrated similar rates of complications likely to benefit from emergency cardiac operations, with comparable mortality rates related to these complications, among centers with and without on-site cardiac surgery. However, in the background setting of treating lower-risk patients, the findings of the present study suggest caution regarding the dispersion of TAVR in centers without on-site cardiac surgery, at least until a dedicated series confirms the results of the AQUA registry.

Although debate remains regarding the clinical impact of PPI post-TAVR (28), given that TAVR is set to expand to lower–surgical risk and potentially younger patients, the deleterious consequences of long-term pacing requires careful attention. Moreover, PPI prolongs hospitalization and may jeopardize the cost-effectiveness of TAVR (29). Therefore, the significant increase of PPI observed from FRANCE 2 to FRANCE TAVI is a notable and important finding. Because this trend was exclusively observed in ESV recipients, this is likely to reflect the availability of the latest iteration of ESV (the SAPIEN 3 valve) that was associated with higher rates of PPI during its early use (30,31). Implantation depth is a consistent predictor of PPI, and an implantation technique aiming at a 70% aortic valve position produced rates of PPI comparable to those observed with previous ESV iterations (31). Whether the implementation of this recommendation yields similar results in large multicenter settings requires careful evaluation in future studies.