We read with great interest the meta-analysis by Pearman et al.1 on the comparison between epicardial ablation for atrial fibrillation and the hybrid approach. The authors should be commended on the way they handled the data. The imbalance between the two meta-analytic cohorts – in terms of the type of atrial fibrillation and left atrial diameter – may have worked to the detriment of the hybrid arm, but the sensitivity analysis shows quite convincingly, at least from a statistical point of view, that this is probably not the case.
However, one cannot fail to note that statistics can only go so far as the quality of available data allows. An all-encompassing meta-analysis that includes not only randomised data but also small series of patients from observational and/or retrospective studies may dilute the true signals of difference between the two strategies studied. Therefore, it would be interesting to report results only from studies with a prospective design.
Furthermore, considering that clear definitions of the parameters studied are the basis of any sound analysis, one may wonder what exactly is meant as a ‘hybrid’ technique. Is it merely a combination of epicardial and endocardial ablation? Do epi- and endocardial ablation happen at the same time or in a staged manner and, when performed simultaneously, is the confirmation of electrical isolation reliable? Aggregating different approaches – all termed ‘hybrid’ – in the same analysis may cloud the picture.
Finally, another point of concern would be the method of follow up. Considering that the minimally-invasive epicardial technique antecedes the hybrid approach, it is conceivable that in a number of the earlier studies (thus more frequently in studies of epicardial technique) follow-up was merely clinical/electrocardiographic, while in later studies (thus more frequently in studies of hybrid technique) higher-yield techniques of patient monitoring may have been used. This would result in a lower detection rate of arrhythmia recurrence in less intensively followed patients, i.e. artificially higher ‘success’ rates. Are there any data regarding these aspects of follow-up for the two patient groups?
Georgios Giannopoulos, Yale School of Medicine, New Haven, CT, USA
Spyridon Deftereos, Yale School of Medicine, New Haven, CT, USA