CrossTalk opposing view: Rotors have not been demonstrated to be the drivers of atrial fibrillation

Despite extensive research, it is still not clear how atrial fibrillation (AF) perpetuates itself. Experimental studies have shown that various mechanisms may sustain this ‘eternal’ arrhythmia (Lewis et al. 1921; Moe & Abildskov, 1959; Moe, 1962; Allessie et al. 1977, 1985; Konings et al. 1994; Mandapati et al. 2000; Jalife et al. 2002; Waldo, 2003). Roughly, they can be divided into focal (repetitive ectopic discharges) and reentrant mechanisms (mother-wave, rotor, multiple wavelets). Recently, we introduced a third mechanism, independent of focal or reentrant activity (the double layer hypothesis; Allessie et al. 2010; de Groot et al. 2010; Eckstein et al. 2011). Using high-resolution optical mapping in isolated sheep hearts subjected to high concentrations of acetylcholine, Jalife and co-workers clearly demonstrated that pacing-induced AF was due to the presence of a very rapid rotor (15–20 Hz) in the left atrium (Mandapati et al. 2000; Jalife et al. 2002). However, although it is unquestionable that under experimental conditions a single rapid rotor can serve as a ‘driver’ of AF, the question still remains whether this is also the operative mechanism in patients with longlasting persistent AF. Based on a series of clinical studies, in which the electrical activity of the atria was mapped by two basket catheters, Narayan et al. (2012a,b,c, 2013) postulated that human AF is due to the presence of a stable rotor that serves as a ‘driver’ to sustain AF. Ablation of the centre of these rotors abruptly terminated or consistently slowed AF in the vast majority of cases, and substantially improved long-term freedom from AF compared to conventional ablation alone (Narayan et al. 2012b). High resolution mapping of AF has been performed in patients during open-chest surgery, by positioning a regular array of 128–256 electrodes (spatial resolution <2.5 mm) directly on the epicardial surface of the atria (Cox et al. 1991; Gerstenfeld et al. 1992; Konings et al. 1994; Harada et al. 1996; Holm et al. 1997; Wu et al. 2002; Kanagaratnam et al. 2004; Sahadevan et al. 2004; Allessie et al. 2010; de Groot et al. 2010; Yaksh et al. 2013; Lee et al. 2014). The mapping electrode was positioned in different areas, including the right atrial free wall, the atrial appendages, the posterior and lateral wall of the left atrium, and Bachmann's bundle. The patterns of activation recorded by this technique were invariably highly complex and varied from beat to beat. To enable quantitative analysis of the complex activation during AF, an algorithm was developed to separate the spatiotemporal process into its individual components (wave-mapping; Allessie et al. 2010). In patients with valvular heart disease and longlasting persistent AF, a high number of dissociated narrow fibrillation waves propagated in different directions over the subepicardium of both atria (Allessie et al. 2010). A remarkable observation was that a large proportion of fibrillation waves was of ‘focal’ origin (de Groot et al. 2010). These ‘focal’ fibrillation waves were not repetitive, occurred over the entire epicardial surface, often had a coupling interval longer than the dominant AF cycle length, and unipolar electrograms at the epicardial origin of these waves exhibited small but clear R-waves (de Groot et al. 2010). This strongly suggests that, rather than being the result of ectopic focal discharges, these ‘focal’ fibrillation waves originated from endo-epicardial breakthrough (Fig. 1). In a recent paper from Australia (Lee et al. 2014), more than one-third of the fibrillation waves in patients with persistent AF were also of ‘focal’ origin (disorganized activity 24.2%, large radial activations 11.3%), with no site demonstrating sustained focal activity. Rotational activity was either not detected at all (Allessie et al. 2010; de Groot et al. 2010), or observed only occasionally (Konings et al. 1994; Lee et al. 2014). If present, rotors were always transient and ceased to exist after only a couple of cycles. High-resolution mapping of human AF thus points to endo-epicardial dissociation as the main mechanism for longlasting AF. The presence of a dual layer of dissociated fibrillation waves will highly stabilize the fibrillation process. Each time a fibrillation wave dies out, it creates an opportunity for the multiple fibrillation waves in the other layer to conduct transmurally and to replace the extinguished wave by one or more focal-breakthrough waves. It is like the paradox of life itself, that one cannot survive without dying. A flexible rectangular array of 8 × 16 electrodes was positioned consecutively on different parts of the atria (10 s recordings). Breakthroughs occurred at multiple sites distributed over the entire epicardial surface of the left and right atrium. The diagram at the bottom illustrates how endo-epicardial breakthrough sites may serve as sources of ‘new’ fibrillation waves (modified from de Groot et al. 2010). Low-resolution mapping is done by inserting basket catheters (Boston Scientific, Inc., Massachusetts, USA) into the right and left atria. They consist of eight splines of eight electrodes with a diameter of 48 mm or 60 mm. Because it is not always easy to make good contact with the atrial wall and because 2–3 splines are floating in the blood pool at the tricuspid and mitral orifices, the effective number of recordings is often limited to less than 48. After filtering and computational processing of the signals, a special algorithm (Topera, Inc., San Diego, CA, USA) was applied that produces a video of the computed activation process. In 98 of 101 patients with sustained AF, this video-algorithm exhibited a large electrical rotor somewhere in the right or left atrium (Narayan et al. 2012b). Low-resolution maps computed from basket electrograms (Narayan et al. 2012a,b,c, 2013) thus differ completely from high-resolution maps obtained by direct-contact mapping of human AF (Allessie et al. 2010; de Groot et al. 2010; Lee et al. 2014). Whereas high-resolution mapping has been unable to detect stable rotors, the Topera algorithms invariably revealed the presence of a large and stable rotor in AF patients (Narayan et al. 2012a,b,c, 2013). Another emerging technique is to reconstruct the activation sequences of human AF by body surface mapping (CardioInsight, Inc., Cleveland, OH, USA; Cuculich et al. 2010; Haissaguerre et al. 2013). In 26 patients with AF, the most common patterns were multiple wavelets (92%) and focal sites (62%). Rotors were seen only rarely (15%; Cuculich et al. 2010) and were not stationary for more than two rotations (Haissaguerre et al. 2013). Ablation at the rotor locations abruptly converted AF into atrial tachycardia after 10 min of radiofrequency application (Haissaguerre et al. 2013). Of course this raises the question of whether high-resolution mapping is missing rotors because it is recording from only part of the atria, or whether the computed rotors are the result of a high degree of freedom of the algorithms to generate local activation times in case of a lack of sufficient data (Allessie & de Groot, 2014). The basket catheters were displayed as a rectangular grid of 8 × 8 electrodes. Yet the perimeter of the sphere formed by the eight splines of the basket is π times longer than its diameter. In order to give a more realistic view of the atrial endocardial surfaces, the maps thus have to be stretched in one direction (perpendicular to the splines) by a factor of 3.14. This is more than just an aesthetic matter, because it not only unmasks the poor spatial resolution between the splines (2–3 cm), but is also essential to avoid serious underestimation of the conduction velocity in one direction. A key feature of atrial fibrillation is its irregularity in time and space. Because the patterns of atrial activation change on a beat-to-beat basis, a series of consecutive AF maps is needed to cover the spatiotemporal variation in activation. The ever-changing size and direction of multiple fibrillation waves also readily explains why fibrillation electrograms are irregular and polymorphic. Yet, Narayan et al. (2012a,b,c, 2013) choose to represent human AF by just a single map. Although they claim that their computed fibrillation maps ‘…were consistent in multiple recordings over >10 min (equating to thousands of cycles)’ (Narayan et al. 2012b), in none of their studies do they actually show that the maps were reproducible over a prolonged period of time. Given the large size of the rotors (in the latest two examples they occupied the whole posterior or anterior wall of the left atrium; Narayan et al. 2013), even in the case of some spatial precession, one would expect predominantly regular and monomorphic electrograms in the area of stable repetitive circular activation. However all recordings from the rotor area presented in the publications of Narayan et al. are highly irregular and polymorphic (Narayan et al. 2012a,b,c, 2013). Only a few electrograms are given to support the presence of a rotor and in no case were local activation times assigned to them. Many electrograms are of poor quality and represent either far-field or injury potentials (Fig. 1 of Narayan et al. 2012b). In the most recent publication, only two or three bipolar electrograms at the rotor path were given (Narayan et al. 2013). Magnification of these electrograms reveals that the two electrograms recorded from opposite limbs of a rotor in the anterior wall of the left atrium (LA) (their Fig. 2; A54 and H65) were not in anti-phase, whereas the three electrograms along a rotor in the posterior LA (their Fig. 3; D45, F45 and E12) did not bear any fixed time relationships. Thus, in none of the maps of the publications of Narayan et al. (2012a,b,c, 2013) are the pathways of repetitive circular activation documented by a sufficient number of electrograms recorded around the rotor path. In the absence of convincing evidence that human AF is driven by a single rotor, it is puzzling how we should interpret the CONFIRM study, in which ‘rotor ablation’ almost doubled the long-term success rate of AF-ablation (Narayan et al. 2012b). In itself, this observation does not prove that human AF is driven by a single rapid source, and the reported clinical success needs to be confirmed by other centres, especially in patients with longstanding persistent AF. It also remains to be seen to what extent body surface mapping might improve the long-term efficacy of AF-ablation (Haissaguerre et al. 2013). At present, the mechanisms responsible for perpetuation of human AF remain unclear. Low-resolution mapping studies claiming that a stable rotor is the main ‘driving force’ behind AF need to be validated by high-resolution mapping. On the other hand, the hypothesis that the substrate of AF is primarily based on endo-epicardial dissociation requires additional proof by simultaneous endo-epicardial mapping in patients with AF. Understanding the mechanisms operative in the perpetuation of human AF is essential for the development of more effective strategies to modify the substrate of AF in order to restore and maintain sinus rhythm (Allessie & de Groot, 2013). Readers are invited to give their views on this and the accompanying CrossTalk articles in this issue by submitting a brief comment. Comments may be posted up to 6 weeks after publication of the article, at which point the discussion will close and authors will be invited to submit a ‘final word’. To submit a comment, go to http://jp.physoc.org/letters/submit/jphysiol;592/15/3167 Natasja M. S. de Groot was trained as cardiologist-electrophysiologist at the Leiden University Medical Center and Cardiovascular Research Institute in Maastricht. She received her PhD from Leiden University on the basis of a dissertation entitled ‘Mapping and ablation of atrial tachyarrhythmias; from signal to substrate’. Currently, she works as an associate professor at the Erasmus Medical Center and she is head of the Translational Electrophysiology Research Unit. Her research focuses on mapping studies of atrial fibrillation, the mechanism of post-operative atrial fibrillation, dysrhythmias in patients with congenital heart disease and the development of electro-anatomical mapping techniques. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. None declared.