Resetting the clock: Short-term narratives and long-term realities in transcatheter valve trials

Temporal evolution of TAVI evidence. The trajectory shown reflects consistent empiric patterns documented across multiple randomized trials and large registries.

Central Message.

Early trials showed short-term advantages of TAVI in low-risk patients. By 5 years, mortality and major outcomes converge. Beyond 5 years evidence is sparse and fragmented. Long-term data are needed.

Veritas filia temporis (“Truth, the daughter of time”)

Randomized controlled trials (RCTs) are the gold standard for evaluating cardiovascular therapies, but how their results are analyzed and communicated often shapes clinical practice as much as the data itself—nowhere more evident than in transcatheter aortic valve implantation (TAVI), where rapid innovation and short-term narratives frequently overshadow long-term evidence.

The trajectory of TAVI can be seen in landmark studies such as the Placement of Aortic Transcatheter Valves (PARTNER) trial series and other trials like Nordic Aortic Valve Intervention Trial (NOTION) and Transcatheter Aortic-Valve Replacement with a Self-Expanding Valve in Low Risk Patients (EVOLUT low risk trial).1, 2, 3 These examples illustrate a recurring cycle: early perceived advantages based on softer end points, diminishing or even reversing benefits at mid- and long-term follow-up, and the continual introduction of new devices and trials before existing long-term data are fully incorporated into clinical practice. No institutional review board approval or consent was required because no specific patient data were used.

Early Gains in Low-Risk Trials: Driven by Soft End Points

The noninferiority sponsor-initiated PARTNER-3 RCT studied 1000 low-risk patients with severe aortic stenosis, comparing transfemoral TAVI using the Sapien 3 valve (Edwards Lifesciences) to surgical aortic valve replacement (SAVR).⁴ At the 1-year mark, the composite end point, which included death, stroke, or cardiovascular rehospitalization, was 8.5% for transcatheter AVR and 15.1% for SAVR (P = .001). The claim of noninferiority was primarily due to differences in rehospitalization, a factor influenced by institutional practices and surveillance bias rather than by hard end points like mortality or disabling stroke, which were similar between groups.⁴

Similarly, the NOTION trial showed early non-inferiority based on composite end points² driving rapid TAVI adoption and guideline changes well before long-term data matured. These early findings set the stage for understanding how midterm results begin to diverge and how longer-term follow-up increasingly reshapes the comparative narrative.

Five-Year Results: Convergence and Emerging Surgical Advantage

At 5 years, the story changed in the PARTNER-3 trial. The primary composite end point was 22.8% for TAVI and 27.2% for SAVR (P = .07), which was no longer statistically significant. All-cause mortality was 10.0% for TAVI and 8.2% for SAVR (odds ratio, 1.24; 95% CI, 0.79-1.97), and disabling stroke occurred in 5.8% of TAVI patients compared with 6.4% of SAVR patients (hazard ratio, 0.87; 95% CI, 0.51-1.48).⁵ Thus, no survival or neurological benefit was seen for TAVI over surgery. Additionally, valve-related complications started to become apparent. Valve thrombosis occurred in 5.8% of patients undergoing TAVI patients versus 0.6% of patients undergoing SAVR, and mild paravalvular leak persisted in 20% of those receiving TAVI.

Paravalvular leak (PVL) remains a persistent structural vulnerability. Though moderate/severe PVL has declined with newer-generation devices, even mild PVL has been linked to increased late mortality, heart failure rehospitalization, and adverse remodeling.^6,7 Although its influence may be muted in elderly, high-risk patients, decades of mild regurgitation in younger, low-risk patients carry cumulative consequences. Leaving a younger patient with residual PVL is therefore not a neutral decision; it represents the acceptance of a known, measurable, and time-dependent risk.

A recent meta-analysis presented at the 39th European Association for Cardio-Thoracic Surgery Annual Meeting compared 5-year all-cause mortality between randomized trials and propensity score-matched cohorts of SAVR and TAVI. Although outcomes for SAVR remained broadly concordant between RCT and real-world propensity score-matched cohorts, TAVI exhibited a marked divergence, with significantly higher 5-year mortality in propensity score-matched cohorts compared with RCT populations (hazard ratio, 0.37; 95% CI, 0.18-0.79; P = .01).⁸

This divergence likely reflects the protective effect of trial selection, procedural standardization, and close follow-up, which attenuates over time in real-world practice. Outside controlled research environments, wider anatomic variability, uneven operator experience, and heterogenous follow-up protocols can all influence outcomes, contributing to the upward drift observed in real-world TAVI mortality curves. As early procedural and rehospitalization advantages fade, differences in durability and survival increasingly favor surgery, particularly in younger patients with longer life expectancies.^5,9

This should not be interpreted as confounding within randomized trials, but rather as the loss of protective trial structures when translated into broader clinical practice. To appreciate the clinical implications of this convergence, it is essential to examine the emerging 10-year-data.

Ten-Year Data: Durability and Survival Favor Surgery

Long-term data from earlier studies provide additional perspective. In the Observational Study of Effectiveness of TAVI with New Generation Devices for Severe Aortic Stenosis Treatment (OBSERVANT) study, at 10 years, survival was 18.2% in the TAVI group compared with 37% in the SAVR group (P = .01).¹⁰ Surgical valves also showed better durability and lower rates of reintervention than first-generation TAVI valves.¹⁰

Similarly, the NOTION trial 10-year follow-up found that whereas structural valve deterioration rates were similar, long-term survival was higher for surgical patients, and they experienced fewer instances of bioprosthetic valve failure and reintervention.¹¹ Although NOTION provides valuable 10-year insights, it reflects early-generation devices and an all-comers population whose mean age was 79 years; therefore, its results may not directly apply to contemporary, younger, low-risk patients.

Although structural valve deterioration in surgical bioprostheses accelerates after year 7 through 10, currently available 5- to 10-year data consistently favor SAVR regarding durability, survival, and reintervention. Emerging anticalcification technologies may improve future valve durability, but comparable long-term evidence incorporating these newer platforms is still limited.

The Resetting Cycle: New Devices, New Trials, Old Lessons Forgotten

When long-term results reveal these trends, the typical response is not critical integration, but the introduction of a new valve iteration and the initiation of more short-term trials, often in the same patient populations. Each cycle emphasizes 1- to 2-year composite end points, claims noninferiority or superiority, and restarts the evidentiary clock. By the time 5- or 10-year data are available, often showing diminished or reversed early advantages, clinical focus has shifted to the next device iteration.

This cycle is not limited to 1 but is seen across all TAVI platforms. Each new generation of self-expanding and balloon-expandable devices is launched with updated features and followed by short-term trials. Although innovation is important, the lack of continuity between device generations results in an evidence gap. Trials like NOTION and EVOLUT have repeated this process: early short-term benefits, midterm convergence,^5,11 and the launch of new studies before mature long-term outcomes are reached.

Regulatory and Transparency Gaps

Regulatory agencies are supposed to oversee the accumulation and disclosure of long-term evidence. In practice, key 5-year data are often buried in appendices of the Food and Drug Administration (FDA) documents, whereas public communications focus on early positive results. For example, the PARTNER-3 FDA Clinical Study Report documented 5-year all-cause mortality of 10.1% for TAVI versus 8.2% for SAVR yet sponsor communications did not highlight this outcome.⁴ Although this numerical difference might appear modest at first glance, it actually represents an almost 20% relative increase in mortality in the TAVI cohort. In absolute terms, this equates to 10 deaths per 100 patients treated with TAVI versus 8 deaths per 100 patients undergoing surgery, an excess of 2 deaths per 100 patients. In contemporary clinical practice, such an excess is far from trivial. To put this into perspective, early mortality after isolated SAVR in low-risk patients at experienced centers is consistently below 1%¹² (often in the 0.4%-0.7% range); in other words, we have long passed the era when the loss of 2 additional lives per 100 could be dismissed as statistical variability. This is a clinically significant signal that warrants clear and prominent communication.

The Pacemaker Problem and Other Long-Term Costs

Permanent pacemaker implantation remains a significant long-term concern. In PARTNER-3, the 30-day pacemaker implantation rate was 6.5% for TAVI and 4.0% for SAVR. In other low-risk TAVI studies, particularly those using self-expanding valves, pacemaker rates reach as high as 33%.¹³ In older or high-risk patients, this has often been considered clinically acceptable. However, in young, low-risk patients with longer life expectancy, the consequences are potentially significant. Chronic right ventricular pacing can induce ventricular dyssynchrony, progressive left ventricular dysfunction, tricuspid regurgitation, and heart failure, whereas persistent left branch bundle block has been associated with adverse remodeling and increased mortality in observational studies.¹⁴ These risks unfold slowly, often beyond the 2-year follow-up windows that dominate early trial narratives. Yet they represent a fundamental difference between surgical and transcatheter therapy. In clinical practice, pacemaker incidence may be underreported: Many centers prolong observation or delay device implantation. These strategies may obscure the true burden of conduction system injury in TAVI.

Additional late complications, including leaflet thrombosis, subclinical hypoattenuated leaflet thickening, and structural valve degeneration are often reported as secondary or mechanistic findings rather than as primary trial end points. Leaflet dysfunction in heavily calcified or bicuspid annulus may arise from asymmetric expansion, underexpansion, or reduced frame circularity. These deployment characteristics have been associated with hypoattenuated leaflet thickening, restricted leaflet motion, and possibly faster structural deterioration over time. These issues should be central in assessing the long-term safety of TAVI devices.

The Silent Years: Sparse Reporting Beyond 5 Years

Although pivotal TAVI trials are characterized by dense early reporting and clear 5-year publications, evidence between years 6 and 10 remains strikingly thin. Beyond 5 years, randomized data are virtually absent; follow-up becomes fragmented across device iterations, postapproval studies, and registry snapshots. Even where follow-up is mandated such as in FDA postapproval surveillance, publicly available data are often limited to interim tables buried deep within regulatory databases rather than peer-reviewed literature. Structural valve deterioration, bioprosthetic valve failure, and late reintervention rates—arguably the most critical questions for younger patients—are typically planned end points but remain unreported or unpublished years after trial completion. This vacuum is not simply a neutral gap; it shapes perception. Accessing these long-term data is itself revealing. Even for experienced clinicians and researchers, locating 6- to 10-year outcomes requires navigating a fragmented, nonintuitive landscape. There are no consolidated datasets or straightforward trial updates. Instead, one must conduct iterative search engine queries, sift through regulatory portals, and examine individual FDA postapproval study pages often to uncover only partial or tabular interim data. This complexity stands in stark contrast to the polished, prominently disseminated 1- and 2-year trial results. When crucial long-term outcomes are effectively hidden behind layers of regulatory bureaucracy and unindexed dashboards, the barrier is not merely logistical; it shapes which evidence actually reaches the profession and the public.

Post Hoc Analyses and Propensity Matching

Post hoc analyses and propensity-matched registry studies, such as the bicuspid PARTNER-3 Registry¹⁵ add further layers of complexity to interpretation. These analyses often involve highly selected patient groups, frequently younger individuals with favorable anatomy, and typically compare outcomes to tricuspid TAVI cohorts rather than the original randomized surgical arm, thereby undermining the balance achieved through randomization. Although useful for generating hypotheses, such approaches are inherently limited by their retrospective design, potential for unmeasured confounders, and model-dependent matching assumptions. Propensity score matching can only adjust for known variables; it cannot account for subtle anatomical or clinical selection factors that may influence outcomes. Moreover, these studies are often underpowered for detecting long-term differences and are susceptible to selection and publication bias. Nevertheless, their findings are frequently given disproportionate weight in shaping perceptions and practice, especially in areas where randomized evidence is lacking such as low-risk bicuspid disease leading to a narrative that may exceed the strength of the underlying data.

Ethical Imperatives

Scientific editorials should avoid accusatory language, yet when structural methodological practices consistently shape narratives to emphasize short-term benefits while obscuring long-term risks, these patterns must be addressed. The concern is not individual blame but collective responsibility.

• •Long-term follow-up should be mandatory and transparently reported.
• •New device trials must acknowledge and reconcile previous outcomes instead of ignoring them.
• •Durability, survival, and major complications should be primary end points, not secondary considerations.
• •Regulatory oversight should ensure that public communications reflect the entirety of available data, not just selective highlights.

Conclusions

The history of TAVI is about more than technological progress; it is also about how evidence is constructed, presented, and remembered. Although early composite advantages at 1 to 2 years are genuine, they are not the whole story. The true influence of a therapy emerges at 5 and 10 years, especially for younger, lower-risk patients.

TAVI's trajectory follows a familiar arc: Early results dominate headlines and shape guidelines, whereas critical long-term outcomes emerge quietly years later, often buried in regulatory appendices. This disconnect has real consequences for clinical decisions and patient counseling. Innovation must be matched by accountability; how truth is managed over time determines whether evidence serves patients or marketing cycles.

Conflict of Interest Statement

The authors reported no conflicts of interest.

The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.