The Moe model disproved: What next?
The Moe multiple reentrant wavelets model was based upon the observations of Gordon Moe in the 1950s, but the conclusions he drew from his experiments that inspired him to develop the first AF simulation have recently been disproven.
To that end, Albert Waldo (Case Western Reserve University, Cleveland, OH, USA) will speak to delegates this morning about putting multiple reentrant wavelets to rest – at least in the Moe model.
Computer models form an element in the battery of methodologies that help us to question and test the putative mechanisms at play in diseases such as AF that are still so poorly understood. The Moe model emerged in this way as one of the predominant theories of AF, and although various newer models have emerged since then, the legacy of his pioneering work endures. Dr Waldo described the experimental observations that led Moe to develop his simulation: “Moe used a canine heart, and he isolated the vagus nerve,” he said. “While stimulating the vagus nerve, he burst-paced the atria and made atrial fibrillation. Moe was unable to map the atrial fibrillation, but when he looked at the atria he said it looked like Brownian motion, the way the atria were moving; it’s often described as a ‘bag of worms’ when the atria are not contracting normally.
“He hypothesized that it was likely due to multiple wavelets reentering – so-called random reentry. These observation led him to do one of the very first computer models of an arrhythmia and he did so a few years later in 1962, which he said confirmed the multiple wavelet hypothesis of random reentry causing atrial fibrillation. So this was the only theory out there, although there had been much speculation before, but his model became the predominant hypothesis. So according to the multiple reentrant wavelet hypothesis, during atrial fibrillation, reentrant wavelets (which almost never manifest what we call head-tail interaction or circus movement) randomly propagate in both atria along routes primarily determined by local atrial refractory periods as well as conduction velocity and also tissue mass.”
Subsequently, the Allessie lab performed electrophysiological studies in vitro with a high dose of acetylcholine perfusing the heart in an effort to test Moe’s prediction of 20 to 40 wavelets being necessary to sustain fibrillation. Dr Waldo explained: “He mapped it with a new system of electrodes endothelially in an egg-shaped device that he put through the mitral and tricuspid valves, recording endocardially from about 172 sites in each atrium. He thought that his work supported the Moe hypothesis, but he said that he thought there were only four to six simultaneously circulating random reentrant wavelets that were necessary to sustain atrial fibrillation. The thing about Allessie’s study is not to say that it isn’t valid, but he infused a very big dose of acetylcholine – in that scenario all the atrial tissue gets the same dose, whereas when you stimulate the vagus nerve the acetylcholine is released locally at the nerve endings so it’s not uniform so it’s a very different process.”
Speaking a little about the similarities that draw alternate models of AF together, Dr Waldo described the process that prompted his own investigations into the Moe model: “Virtually all of these models have demonstrated a driver mechanism for atrial fibrillation: either a single focus or multiple foci, or a single reentrant circuit or unstable reentrant circuit,” he said. “These driver mechanisms are very rapid and all include the existence of a source – a single source – of very rapid rates which produce high frequency wavefronts that interact with the atrial tissue variably because of refractory periods and conduction properties, etc. This produces atrial fibrillation: fibrillatory conduction. So the driver mechanism is responsible for maintaining the atrial fibrillation in these models. And it’s not random reentry. It’s either ordered reentry or a single focus firing rapidly, and that’s the point in animal models.”
Recreating Moe’s original ‘bag of worms’ in the canine heart meant that Dr Waldo’s group was able to see if his observations were correct. “Our lab has a very sophisticated mapping system with 512 electrodes on the atria and we used that to repeat the study exactly how Moe did it,” Dr. Waldo enthused. “We isolated the vagus nerve, we burst-paced the atria for three to seven beats rapidly and we made atrial fibrillation. When we mapped it, we had very high resolution, and what we found was remarkable: we found virtually no reentry at all; we found that the atrial fibrillation was due to multiple foci firing rapidly. There were at least two going, and most of them were persistent foci. And we also had some transient foci that would come for a few beats, disappear and come back later, so these foci were very short-cycling.
“But what these foci did was almost never make fibrillatory conduction. What they did was create wavefronts that collided with each other, and because the cycle lengths were different (the rates were fast, but different), they kept colliding at different places. Also, sometimes they didn’t collide but they fused, so atrial fibrillation was really due to multiple variable sites of collision and fusion of atrial fibrillation but reentry was very rare. Occasionally we saw a little bit of fibrillatory conduction; the most we ever saw it go around was two cycles (three beats). So this is very very different, and it’s a whole new paradigm for atrial fibrillation.
“We also studied the termination of atrial fibrillation, which supported all this – there was no evidence of reentry. When the foci slowed the fibrillation slowed, and then the foci and the fibrillation stopped. So this is really very important, because many people still think that multiple reentrant wavelets is the basis of atrial fibrillation, and this study shows that the original Moe model didn’t even have it! So whether multiple reentrant wavelets really exist in human atrial fibrillation nobody really knows; but many have thought so. For instance, the original MAZE surgical procedure to cure atrial fibrillation was postulated on the basis that it was the MAZE procedure – all these randomly circulating wavelets – that if you made a whole bunch of incisions in the atria they would heal the fibrosis, and these multiple reentrant wavelets couldn’t conduct through the fibrosis so they would stop. When the MAZE procedure had some success they thought it was indirect confirmation of this theory. One of the things we learned is that the first lesion set that the MAZE surgery made was to isolate the pulmonary vein, and we know that the pulmonary vein is the trigger of atrial fibrillation, so it probably worked for reasons that were not understood at the time, and are still not really understood, because we know that isolating the pulmonary vein is not the whole story.”
No doubt, these experiments are emulating but one of many varied etiologies that culminate in symptoms of AF, and Dr Waldo acknowledged this: “This is a very special aetiology. Of course, there is vagally mediated atrial fibrillation in humans, and it may well be that what we are showing is exactly what happens in humans. That would be a very reasonable hypothesis. We also know that in the remodeling of the atria, there are all kinds of abnormalities, with calcium cycling etc. and the bottom line is that people have demonstrated that you get after-potentials (delayed after-depolarization) and that these can act as a focus firing rapidly as well. So there are people that have suggested that maybe delayed after-depolarization is one of the etiologies of atrial fibrillation. But what happens when you have atrial fibrillation for a long period of time with all the remodeling, a lot of things change, and what we do know is that the substrate is very important. Our study shows once again that if the drivers of the fibrillation are removed the atrial fibrillation stops. These drivers are foci, but drivers can also be reentrants. Most people think that it’s not random reentry, most people think that it’s actu ally circus movement: the impulse goes round and round, and the head chases the tail.
“We have a lot to learn about what sustains atrial fibrillation but this paradigm that we have described is a new one, that is important for several reasons, not only because it is new, but also it finally dismisses the multiple wavelet hypothesis – the notion that this is a cause of AF – at least in the Moe model. Whether it happens in other models, we have to see. But it’s certainly not a part of the Moe model, despite the original hypothesis and the idea that it had in fact been demonstrated. The Allessie study was not the Moe model at’s the avenue of research that’s probably the most useful.”