On the Dutch Registry of catheter ablation of atrial fibrillation: Authors’ reply

We thank Jordaens et al.¹ for their reaction on our paper, and the issues for discussion they bring forward. They question whether we underestimate the number of complications following pulmonary vein isolation (PVI) due to incomplete follow-up. In our real-world registry study, we included all available data of patients with drug-refractory atrial fibrillation (AF) who underwent PVI in Netherlands Heart Registration (NHR) participating hospitals (14/16 ablation centres during study period) and were treated with conventional radiofrequency (C-RF), phased radiofrequency (Ph-RF) or cryoballoon (CB) ablation.² We believe that missing values in our manuscript are based on a systematic error and unlikely to differ among groups. Furthermore, the registry has a quality system in place and performs audits. To overcome the issue of underestimation, we reported the complication incidence of the data available. This resulted (in contrast to the calculation of Jordaens et al.) in a cumulative complication incidence of 3.9%, 2.5%, and 4.2% for C-RF, Ph-RF, and CB, respectively. Of note, phrenic nerve palsy (PNP) and vascular complications were added to the NHR registry from 2016 onward. Phrenic nerve palsy almost exclusively occurred after CB ablation.² The degree of reversibility of PNP is of great interest but beyond the scope of the current report.

Secondly, Jordaens et al. express their regret on the absence of efficacy data of the different ablation modalities. As the NHR registry reflects standard-of-care in the Netherlands, systematic follow-up on the absence of AF was incomplete. Therefore, we could not report the recurrence rate of any atrial tachyarrhythmia as an outcome parameter and used repeat left atrium (LA) ablation as a surrogate instead. Although repeat LA ablation within one year provides a fair comparison between the groups, we are cautious about concluding the effectiveness of the ablation modalities based on repeat ablation. Importantly, the rate of repeat ablation within 1 year depends on patient, physician and centre preferences, and the decision to re-ablate may be delayed by attempting antiarrhythmic drugs to reduce symptoms first. This introduced a selection bias for which we were not able to correct.² To date, there are no randomized controlled trials demonstrating relevant differences in atrial tachyarrhythmia recurrences; neither between C-RF and CB nor between C-RF and Ph-RF, irrespective of the follow-up method used.^3,4 Only one small randomized controlled trial has suggested the superiority of CB compared to Ph-RF.⁵ Although patients selected for CB had fewer redo ablations, based on the data provided in our study, we cannot conclude that CB is more effective compared to the other modalities.

Jordaens et al. further propose that patients should be informed about a higher risk for repeat LA ablation when undergoing C-RF ablation compared to CB. Although we sympathize with optimizing patient information and shared decision-making, we stress again that these data are registry data and that selection bias for an ablation strategy is expected to be present.

Finally, Jordaens et al. ask whether there is still a place for Ph-RF ablation. Also, this question cannot be answered, and it is beyond the scope of the current analysis, which describes the procedural complications of PVI in the Netherlands between 2012 and 2017. During that time, a considerable proportion of procedures were performed with Ph-RF, without a significant signal of more complications.⁶

To conclude, we described the complication rates of different ablation modalities based on standard-of-care collected clinical and outcome variables.² This may have caused underreporting, but we presume that—should that have been the case—this represents a systematic error, and are confident that the complications we report actually occurred. We hope that our data may inspire us and our colleagues to systematically follow-up our AF ablation patients concerning freedom of recurrences and procedural and late complications.

Conflict of interest: J.R.d.G. reports research grants through his institution from Abbott, Atricure, Boston Scientific, Bayer, Daichi Sankyo, Johnson&Johnson, and Medtronic. He reports speaker/consultancy fees from Atricure, Bayer, Daiichi Sankyo, Johnson&Johnson, Medtronic, Novartis, Servier, outside the submitted work.

References