With summer and conference season approaching, as we await updates from major trials PARTNER 3 at 5 years and Evolut-Low Risk at 6 years—both increasingly favorable to surgical options—the DEDICATE study emerges as the latest “hit” in support of percutaneous solutions. This German multicenter experience distinctly diverges in design from previous, heavily sponsored trials. Essentially, it seems to address design and interpretation issues that previously raised impartiality concerns in these clinical trials. While published in The New England Journal of Medicine (NEJM), a journal that, in my personal view, tends to obscure data in numerous supplementary appendices, this work brings a refreshing level of transparency, characteristic of German rigor. This publication seems crafted to present a concise manuscript where “what you see is what you get.” Delving into the results at the source requires a careful review of the supplementary material, which, despite its transparency, almost acts as a “deterrent by complexity” for those seeking a straightforward message. 

Before analyzing the study, I feel compelled to issue two critiques relevant to future works addressing this controversy. In an era where TAVI implants have significantly outpaced surgical prostheses, continuing to perform non-inferiority analyses solely to achieve easier statistical significance seems inappropriate. TAVI has now established itself as a leading therapeutic alternative and, therefore, should be evaluated on equal footing with surgery. Secondly, while discussions of surgical risk remain necessary, as low-risk is the last stronghold for surgery, it is striking that the leading therapeutic alternative lacks widely accepted risk estimation systems in clinical practice. This likely stems from a need that, evidently, it does not have. 

Moving into the study details, this was a randomized, multicenter trial (38 German centers), analyzed by intention to treat with a somewhat redundant non-inferiority design, including 1414 patients, of whom 701 underwent TAVI and 713 underwent surgical aortic valve replacement. This trial thus represents the largest published to date in this setting, following the Evolut-Low Risk with 1403 patients, PARTNER 3 with 950, UK-TAVI with 412, and NOTION with only 276 patients. Both groups had low surgical risk, with a mean STS score of 1.8% across the series. Eligible patients were those in whom either therapeutic option was feasible and who presented with severe symptomatic aortic stenosis. Regarding age, study protocol modifications allowed for the inclusion of patients under 75 years old (approximately 38% in both groups) and those younger than 70 years (11-13%), reflecting typical German clinical practice. Although the mean surgical risk was low, 3.9% of patients in the TAVI group and 4.9% in the surgical group had intermediate risk (STS score >4%). Notable exclusion criteria included congenital bi- or unicuspid valve disease, significant concomitant coronary or valvular disease, left ventricular dysfunction with an ejection fraction <20%, or left ventricular hypertrophy with outflow tract obstruction. Importantly, associated procedures were far fewer than in previous trials: 11 coronary revascularizations, 6 AF surgical ablations, 6 ascending aortic replacements, 1 mitral valve surgery, and 2 tricuspid valve surgeries. However, there were still 26 procedures that did not involve pure aortic valve replacement, whereas none of the TAVI procedures included concomitant interventions. In fact, percutaneous revascularization within a month prior to randomization was an exclusion criterion. 

The procedures performed also differed significantly from previous clinical trials. In this non-sponsored study (notwithstanding the extensive conflicts of interest among the authors), operators had complete freedom in choosing the devices and access route. This freedom was made possible by funding from the German Center for Cardiovascular Research and the German Heart Foundation, independent from the biomedical industry. Regarding the surgical approach, the valve sizes typically associated with the German population predominated, with the majority being Edwards® pericardial prostheses in various versions, followed by a concerning 15% St. Jude Trifecta, given concerns over long-term structural degeneration. Sutureless prostheses accounted for 15.8% of implants, divided between the Edwards Intuity® and Corcym Perceval® models. Minimally invasive approaches comprised only 38% of the series. For the percutaneous option, 97.3% used the transfemoral approach, with transapical and transaxial routes being rare. The implanted prostheses were two-thirds balloon-expandable (almost exclusively Edwards Sapien® 3/3 Ultra), with the remaining third being self-expandable (split between Boston Acurate® Neo/Neo 2 and the Medtronic CoreValve/Evolut® variants). 

Although the NEJM article does not detail short-term outcomes, these results are of particular interest and may hold the key to understanding long-term findings. Notably, after randomization, 100 patients assigned to the surgical arm were not operated on, with 70 crossing over to TAVI treatment. This crossover occurred in only 18 TAVI-assigned patients, of whom 12 subsequently underwent surgery. As a result, the entire published intention-to-treat analysis is somewhat questionable, with a more reliable approach found in the per-treatment analysis, accessible in supplementary materials. As previously mentioned, patients were eligible for both surgical and percutaneous options. Surgical outcomes, particularly in the German healthcare context, were remarkably ordinary and lacked the excellence associated with such a high-standard setting. For instance, mean extracorporeal circulation time was 88 minutes; mean aortic cross-clamp time was 61 minutes; mean ICU stay was 2 days, and mean hospital stay was 9 days, with only 40.4% of patients discharged home (43.9% were discharged to rehabilitation centers). Major bleeding events requiring transfusion occurred in 13.8% of cases, with transfusions exceeding three units of red blood cells in 13% of patients. The surgical group had a stroke rate of 3.1% (compared to 1.9% in the TAVI group, where cerebral embolic protection devices were used in over 5% of cases), mild or greater paravalvular leakage at 4.9% (compared to 20.9% in the TAVI group), and a pacemaker implantation rate of 3.4% (compared to only 8.7% in the TAVI group). This resulted in a 30-day mortality rate of 1.5% versus 0.7% for the TAVI group. 

The authors conducted a one-year follow-up, although they promise to extend this to five years per the study protocol. While the one-month mortality curve did not show significant divergence, the upward trajectory over the subsequent year raises concerns, with one-year mortality reaching 5.6% for TAVI and 10.1% for surgery. These results are similar in the intention-to-treat analysis published. Throughout this period, the surgical group experienced a range of adverse outcomes leading to the death of 42 patients (compared to 18 in the TAVI group) from causes such as sepsis (8 cases), cardiogenic shock or heart failure (8 cases), COVID-19, endocarditis, intracranial hemorrhage, cancer, respiratory failure, and arrhythmias (2 cases each). Despite seemingly comparable preoperative characteristics, the adverse progression in some surgical patients significantly impacted one-year survival. In the TAVI group, outcomes were also worse compared to clinical trials, where one-year mortality in Evolut-Low Risk was 2.4% for TAVI and 3% for surgery, while PARTNER 3 reported 1% for TAVI and 2.5% for surgery. However, for the surgical group, the results were comparatively even less favorable. 

The authors conclude that among patients with severe aortic stenosis at low surgical risk, due to the study’s design, TAVI was not inferior to SAVR regarding all-cause death or stroke at one year. 

COMMENTARY: 

The subsequent promotion and dissemination of DEDICATE trial findings came promptly after its recent publication. Social media and interventional cardiology forums have resonated with this “new success story,” as it has been labeled. Some of the more aggressive advocates, based on the inclusion of younger patients below the age of 75, have questioned the timeline for lowering the TAVI age threshold in this patient population. Others, observing one-year trends, argue that these results provide sufficient grounds for robust conclusions to further TAVI indications in all contexts. It is likely we are witnessing the emergence of a new hallmark study, as previous sponsored clinical trial trends have not been as favorable as expected. 

In presenting the raw results, we aimed to shed light on the factors behind the more than 5% difference between the two options at one-year follow-up. Given the limited losses to follow-up (13 in the TAVI group and 18 in the surgery group), this discrepancy likely lies in details that, as mentioned before, NEJM publication conceals within a plethora of data and appendices. One of the primary factors could be patient crossover; 100 patients in the surgical group, a seventh of the cohort, were not treated as initially planned. Alongside the questionable validity of the published intention-to-treat analysis, the per-treatment results may reflect an imbalance in pre-procedure characteristics between groups. Indeed, preoperative morbidity is presented only in the intention-to-treat analysis, not in per-treatment appendices, which detail only 30-day and mid-term outcomes. Another notable aspect is the worse postoperative course for surgical patients, with an average hospital stay of 9 days and less than half discharged home. Whether this reflects a peculiarity of the German healthcare system is uncertain, but these results are more consistent with those reported for TAVI, with 75% of cases discharged home and a mean hospital stay of 5 days. Minor penalizations in the surgical group likely contributed to the unfavorable short-term outcomes, such as an exceedingly high bleeding event rate, the disproportionate impact of associated procedures between both therapeutic options, and unaccounted morbidities in preoperative characteristics or risk scores that contributed to mortality events like COVID-19, cancer, sepsis, or intracranial hemorrhage (14 cases not found in the TAVI group, where mortality was mainly cardiovascular, including one case of aortic dissection). 

In conclusion, it is essential to highlight the apparent impartiality of this study. Despite receiving funding from public organizations, the abundance of conflicts of interest among the authors is notable. Given the reimbursement system in Germany’s healthcare, there remains a reasonable concern regarding potential biases favoring a more profitable therapeutic option. In other words, instead of being sponsored by a single entity, this study appears to have garnered the support of most transcatheter industry representatives. 

There are still questions to be resolved regarding the long-term follow-up results. Trends do not bode well for the surgical option, and it is unclear how various factors will influence future outcomes, including prosthetic degeneration (notably in sutureless and St. Jude Trifecta® prostheses in the surgical group), progression of mild or greater paravalvular regurgitation from 4.9% to 8.8% in the surgical group and from 20.9% to 24.3% in the TAVI group after one year, moderate or greater mitral insufficiency in one-year follow-up at 5.4% for TAVI and 4.7% for surgery, or moderate or greater tricuspid insufficiency at 5.5% in TAVI and 9.8% in surgery. For now, we can only wait and hope that these results will be taken cautiously… but the replay of this summer’s “hit” seems assured until we know the outcome by heart. 

REFERENCE: 

Blankenberg S, Seiffert M, Vonthein R, Baumgartner H, Bleiziffer S, Borger MA, et al.; DEDICATE-DZHK6 Trial Investigators. Transcatheter or Surgical Treatment of Aortic-Valve Stenosis. N Engl J Med. 2024 May 2;390(17):1572-1583. doi: 10.1056/NEJMoa2400685.