Health technology assessment on catheter ablation of atrial fibrillation in Belgium

This editorial refers to ‘The right recommendations for the wrong reasons?’ by M. Duytschaever et al., on page 615

Introduction

In a separate editorial that was published ahead of print on 23 January 2013, Belgian cardiologists comment on a health technology assessment (HTA) on catheter ablation of atrial fibrillation (CA-AF) that we prepared at the Belgian Health Care Knowledge Centre (KCE).¹

We would like to thank the editor of Europace to provide us the opportunity to rebut our colleagues' contentions.

What's the evidence in favour of catheter ablation?

In contrast to the claims made by Duytschaever et al., our HTA report does not limit itself to an evaluation of administrative data. We performed a systematic review of the literature on the clinical effectiveness, the safety, and the cost effectiveness of CA-AF. We critically appraised randomized controlled trials (RCTs) that compared catheter ablation with antiarrhythmic drug (AAD) treatment, as well as recent observational studies.^2,3 The variation in the clinical effectiveness of CA-AF reported by different authors is huge. One-year recurrence rates of AF after ablation off-AAD ranges from 11 to 44% in RCTs and from 9 to 80% in observational studies. The total number of patients studied in RCTs is very small when compared with the hundreds of thousands of patients who have worldwide been treated with catheter ablation so far.

What did our study on Belgian data reveal?

From a nationwide health insurers' database, we retrieved claims data of all patients (n = 830) who underwent a first CA-AF in Belgium from November 2007 to December 2008. Although official administrative data have the disadvantage of not containing clinical information, unlike registries and surveys they are not prone to selection bias.^2,3

Based on data on reimbursed procedures and drugs, we assessed AF recurrence. We regarded any redo ablation as a failure of the index procedure. We also considered it failed if a patient underwent an electric cardioversion or used an AAD beyond the blanking period of 3 months. This is in agreement with the definitions proposed in the 2012 Expert Consensus Statement on Catheter and Surgical Ablation of Atrial Fibrillation.⁴ The same document also defines partial success as a 75% or greater reduction in AF occurrence in the presence or absence of AAD. It is however not clear if ‘partial success’ is a meaningful outcome and the long-term outcome of a partially successful ablation is unknown. Of note, it has been shown that over a third of patients go for CA-AF because of their desire for a drug-free lifestyle.⁵

In our view, it is self-evident that the assessment of the effectiveness of CA-AF considers the index ablation only. Although repeat procedures may lead to better AF control, they also multiply complications and costs involved.

The combination of the occurrence of a repeat ablation, the need for electric conversion and the use of an AAD beyond a 3-month blanking period, resulted in our study in an estimate of failure of the index ablation of 59.8% and 65.9% after 12 and 24 months respectively. Although we found these numbers disappointing, they still may represent an underestimation of the actual recurrence rate. We may have missed patients in whom the procedure failed but the arrhythmia became accepted and was no longer treated with AADs, cardioversion, or a repeat catheter ablation.

We discussed these results with Belgian expert electrophysiologists. In their view, our estimate could represent an overestimation of AF recurrence since a continued use of AADs as such may not reflect failure of the ablation. It may indeed have resulted from the fact that both physicians and patients were reluctant to stop the AAD after CA, even if it was successful. Furthermore, it was contended that CA might have been partially effective, allowing a patient with previously refractive AF to become drug responsive after CA. Therefore, these experts suggested that the use of AADs would be a better measure of ablation failure if it was only counted in case an AAD was reinitiated after a drug-free time interval, arbitrarily put at 1 month beyond the blanking period. We thought that this might not apply to the use of amiodarone, an AAD with dreaded side effects and a very long half-life. Consequently, we took these points of view into consideration and constructed two additional models to assess the effectiveness of the ablation. Adding a drug-free interval of at least 1 month beyond the blanking period for any AAD, but not for amiodarone, resulted in an estimate of failure of the index ablation of 45.5 and 55.9% after 12 and 24 months, respectively. Adding a drug-free interval of at least 1 month for any AAD resulted in an estimate of failure of the index ablation of 37.3 and 49.9% after 12 and 24 months, respectively. In contrast to the claims of the Belgian cardiologists, in the message we conveyed to the Belgian mass media, we only presented the latter most optimistic estimate of the effectiveness of CA-AF.

Based on one single reference,⁶ the editorialists feel that the cost of CA-AF is acceptable. From a systematic review of the economic evaluations of CA-AF we found the most determining input variables in full economic evaluations are the impact of CA-AF on utility and/or stroke, and the duration of these effects. However, it appears that there are no good utility data, especially not in the long-term, and that evidence on stroke impact is lacking, making the outcomes of the models rather hypothetical. Based on current knowledge, it is hard, if not impossible, to assess whether CA represents efficient use of limited resources. There is no doubt that CA-AF is associated with high initial costs and certain complications. Its cost-effectiveness depends on the belief one places on long-term outcomes. We preferred not to model ourselves the cost-effectiveness of the procedure since our calculations would be prone to the same uncertainty as previous models by other authors. This is in line with the conclusion of SBU, the Swedish HTA agency, which in 2010 concluded that the available scientific evidence is insufficient for drawing conclusions about the cost-effectiveness of CA-AF because its long-term effects are uncertain.⁷

Conclusion

Taken into consideration the high cost of a CA-AF (€9600) and the potentially severe complications, we confirm that the real-world effectiveness of CA in Belgium, even though located within the wide bounds of the off-AAD effectiveness of a single CA-AF as reported in other observational studies, is rather disappointing.

Conflict of interest: none declared.

References

Author notes