Impact of Timing of Atrial Fibrillation Ablation on AF Recurrence and Clinical Outcomes in Patients With HFpEF and HFrEF

We appreciate the thoughtful comments by Drs. Kataoka and Imamura regarding our study “Impact of Timing of Atrial Fibrillation Ablation on AF Recurrence and Clinical Outcomes in Patients With HFpEF and HFrEF [1].” Their observations highlight important considerations in interpreting AF recurrence and understanding real-world practice patterns across HF phenotypes. We appreciate the opportunity to respond to their comments as noted below. The authors note that our recurrence definition—redo ablation, direct current cardioversion (DCCV), or anti-arrhythmic drug (AAD) initiation—may be influenced by clinician preference. While no surrogate perfectly captures true AF recurrence in large databases, these markers are widely accepted indicators of clinically meaningful recurrence that prompt escalation of rhythm-control therapy. Redo ablation, DCCV, and class IC/III AAD use after AF ablation almost always represent AF recurrence in practice and are commonly used in real-world AF outcomes research [2-4]. We acknowledge that this will not capture all recurrence, as this would require long-term loop recording to ensure comprehensive and objective measurement. However, such intervention is not standard of care and only done in some randomized trials with a smaller number of patients, which brings its own caveats and critiques. Our intention was to identify events that reflect true clinical burden as best as possible given the limitations of the data set. Importantly, this definition was applied uniformly across HFpEF and HFrEF cohorts, which was an attempt to minimize systematic bias. We agree that rhythm control is often pursued more aggressively in HFrEF compared to selected HFpEF patients. Nonetheless, the overall pattern in our study was consistent: delayed ablation was associated with a higher burden of recurrence-related interventions in both HFpEF and HFrEF groups. This aligns with the established concept that atrial remodeling progresses over time irrespective of HF subtype, resulting in a more complex substrate encountered with ablation. In our analysis, all components of the composite measure were lower with early ablation in HFrEF, whereas in HFpEF only AAD initiation differed—potentially reflecting variation in clinical management. We also recognize that clinician-driven biases could influence our secondary analysis comparing HFpEF with HFrEF. To mitigate this, we applied propensity score matching to balance key confounders. Notably, differences between HF subtypes appeared only in the early ablation cohort and disappeared in the later ablation cohort despite identical matching. This pattern suggests that ablation timing—more than HF phenotype—is a principal driver of recurrence. Our analysis incorporated the diagnosis-to-ablation interval, an important factor often overlooked in prior studies. We agree with the authors that many patients probably present more than 1 year after AF onset. However, we do not believe ablation should be withheld in these patients, either in clinical practice or based on the findings from our study. Our data suggest that early ablation was associated with fewer rhythm-control interventions. However, later timing of ablation still appeared safe, with similar rates of stroke, HF hospitalization, and mortality when compared with early ablation. Furthermore, prior randomized trials support that benefits in these patients are still achievable even with relatively later ablation. In CASTLE-AF, 28% of patients in the ablation arm had AF duration > 1 year, yet still experienced improved outcomes with ablation [5]. Similarly, in CAMERA-MRI, patients with HFrEF demonstrated significant improvement in LVEF despite a mean AF duration of 23 months [6]. These findings underscore that AF ablation can remain effective even when performed later in the disease course. Our study, if anything supports the collective data that the earlier the intervention the better the outcomes appear to be in AF patients. We also agree with the authors that these data suggest improvement in pharmacologic and procedural burden with respect to earlier ablation timing. However, we cannot infer causation from these data, either for or against HF or mortality benefit, as this would need to be further evaluated in randomized controlled studies. In summary, our study suggests that early intervention for AF ablation in patients with HF can optimize rhythm-control outcomes. Such ablation appears to remain a reasonable and safe option even when intervention timing is not ideally early. We appreciate the opportunity to clarify these points and thank the authors for their thoughtful engagement with our work. The authors declare no conflicts of interest.