Atrial fibrillation (AF), the chaotic and disorganized activity of the upper chambers of the heart, is the most common cardiac rhythm disturbance. It affects approximately 1% of the US population, or almost three million people. It occurs more frequently in the aged, affecting about 5% of people over 70 and as many as 10% of people over the age of 80. AF is also considered to be the leading cause of stroke in the US, leading to as many as 30% of strokes in the elderly population.
Besides being a costly arrhythmia, AF is also extremely troubling to patients and their physicians. Most people experience symptoms ranging from a relatively minor feeling of discomfort or 'doomÔÇÖ to severe frank palpitations and even syncope. As patients go in and out of AF, they present frequently to emergency rooms for control of their heart rate and/or rhythm, consuming large amounts of time and energy for treatment.
Despite having to deal with AF for many decades, the medical profession has made remarkably little progress in treating it. Currently accepted therapies with drugs and electrical cardioversion are palliative, not curative. Many patients are told to 'live withÔÇÖ their arrhythmia as nothing can be done - they are managed chronically with anti-coagulation and rate control. Patients in whom ventricular rate may be particularly difficult to control sometimes have their cardiac conducting system purposefully destroyed (artioventricular (AV) nodal ablation) and a permanent pacemaker implanted.
In comparison with pharmacologic therapy, surgical therapy for AF has been practised for the last 15 years and has enjoyed as much as a 97% cure rate with long-term follow-up.1-4 As it is practised today, however, CoxÔÇÖs Maze-III procedure remains poorly accepted due to its complexity and morbidity. Over the last few years, there has been an explosion of interest and activity in developing less invasive techniques that would be more widely accepted by patients and physicians yet that maintain the high cure rate offered by the Maze-III.
Much of the lack of progress can surely be blamed on the paucity of basic science models in which this disease is studied. Even the classification system for this disease is confusing and antiquated. It has only recently been concluded that AF is probably not one disease, but more likely two or three, depending upon the stage at which the patient presents, how long they have had the arrhythmia, and what is the predominant rhythm disturbance. Accordingly, AF has now been divided into three classes - paroxysmal, persistent, and permanent.5-9
The main dividing point among these varieties of AF is the patientÔÇÖs predominant rhythm. Paroxysmal fibrillation is marked by the intermittence of AF. The patient suffers from self-limited episodes of AF that convert back to sinus spontaneously. These attacks may last anywhere from a few minutes to a few days. They may occur very infrequently or several times a day. With regard to natural history, paroxysmal AF is considered by most physicians to be the 'initialÔÇÖ type of AF, as patients classically present first with paroxysmal disease, and then describe a progression to the more sustained types. Of the approximately 160,000 patients newly diagnosed with AF each year, most of them will have paroxysmal disease.
Persistent fibrillation is rather similar to paroxysmal AF, except that it is marked by the tendency to remain in fibrillation unless an intervention such as cardioversion is performed, either with medications or with an electrical shock. Patients with persistent fibrillation may also have either infrequent or frequent attacks, but will always require cardioversion to restore sinus rhythm. Some patients have presented with a long history of multiple episodes of persistent AF, having had more than 20 DC cardioversions.
Prior to 1998, little was known about the atrial properties of these patients. Relatively few electrophysiological studies had been performed on paroxysmal or persistent fibrillators. What data were available indicated that the electrical properties of the atrium in these patients were only slightly abnormal, chiefly indicated by prolonged average activation times and dispersed recovery times.10-13 However, because these patients spent most of their time in sinus rhythm, it appeared that the search for an 'AF triggerÔÇÖwas most important.
In 1998, Haissaguerre and his colleagues published the first of a series of papers in which the triggers for paroxysmal AF could be found.14 They found that in this patient group, over 90% of the electrical impulses that initiated AF appeared in or around the pulmonary veins (PVs). Although others have reported other trigger sites, such as the superior vena cava,15-17 it still appears that the majority of AF episodes start in or around the PVs.
Over the succeeding five or six years, there has been an enormous interest in, and effort made toward, electrically separating the PVs from the remainder of the heart in order to isolate the AF trigger. This has been variably referred to as 'PV ablationÔÇÖ or 'PV isolationÔÇÖ, and typically involves creating a series of scars that block any impulse(s) originating in the PVs from propagating outward and onto the left atrium itself.
With the rapid development of radiofrequency techniques for the ablation of various cardiac arrhythmias, the electrophysiology community has tried to adapt these techniques to PV isolation. The first procedures focused on identifying a specific firing locus within a specific vein and ablating that site. Complications were many, and even short-term cures were few. As the procedures have evolved, ablations have become more 'anatomicallyÔÇÖ guided to encompass more of the left atrium and almost never travel within the cylinder of the vein itself. These more 'Maze-likeÔÇÖ procedures have had far fewer complications such as PV stenosis or perforation, and seem to have higher cure rates.18,19 However, the procedures remain quite lengthy and still do not approach the Maze-III with regard to cure rate.
Despite these problems, PV isolation remains a worthwhile goal. Further procedural and device refinement will eventually make this procedure fast, safe and reliable at curing a large number of patients with paroxysmal and persistent AF. However, there are some considerations that are quite important as we concentrate our efforts in this direction. The first is the direction of energy application. The endocardial application of a destructive energy source will always prove problematic - too shallow a lesion will not produce complete, permanent conduction block, and there will be no long-term cure. Too deep a lesion will place collateral structures at risk of injury and perforation, chiefly the esophagus and left atrium.
Therefore, it is quite likely that the best approach to cure patients with paroxysmal and persistent AF will be to perform a PV isolation procedure through an epicardial approach. This is much safer (vide supra) and likely can be performed much more quickly as direct visualization of the device location, and the lesion can be accomplished with conventional cameras and instruments. The open questions at this point in time center on which technology to use and how to obtain access to the heart in a minimally invasive manner that will be acceptable to patients and their referring physicians.
Permanent fibrillation is a rather different arrhythmia than paroxysmal or persistent. Patients with permanent fibrillation are continuously in AF and have, by definition, failed attempts to cardiovert them back to sinus rhythm. It is beginning to be understood that in these patients there have been fundamental changes in the electrical properties of the atrium. Most of the studies performed to date on either animal models of permanent AF, on tissue taken from patients with permanent AF, or in the electrophysiology laboratory indicate that the atrium demonstrates abnormally low conduction velocities, depressed ion channel currents, and decreased refractoriness to impulse conduction.20-22 All of these alterations translate into a decrease in the wavelength of electrical conduction such that AF can be sustained in these abnormal atria whereas it cannot in normal atria.
The lack of success seen during drug therapy, therefore, is not unexpected. If one were to design a drug that would restore the wavelength of conduction to normal and thereby prevent AF, one would either restore the conduction velocity or the refractory period. There is no drug which yet addresses the former, and our only successes have been with drugs that accomplish the latter. Unfortunately, the drugs that best prolong the atrial refractory period, such as amiodarone and dofetilide, are also the ones most plagued by annoying and dangerous side-effects. Therefore, even though drug therapy may prove effective for some patients, these agents are frequently withdrawn as time goes by.
If we cannot restore the electrical properties of the atrium such that AF can no longer be sustained, then perhaps we can modify the substrate in other ways. One way would be to reduce the available area for electrical conduction to something less than can sustain the arrhythmia. Over 40 years ago, Gordon Moe proposed that a minimum square area of atrium (6.0cm2) would be necessary to sustain AF.23-26 These experimental predictions were recently confirmed in a dog model where AF was found to sustain when the wavelength of conduction dropped below 8.7cm, corresponding to a circuit area of exactly 6.0cm2.27
It is therefore more easily understood why the Maze-III operation has been so successful at treating all patients with AF - paroxysmal, persistent, and permanent. For the first two, the PVs are isolated by circumferential incision, and for the latter group, the atrium is partitioned by incisions into small blind corridors, each of which is too small to permit re-entry of an excitation wave - less than 6.0cm2. However, because the operation preserves the connection between the sinus node and the atrioventricular node, the normal pacemaking mechanism is preserved. Also, because each blind corridor is connected to its neighbor, impulses originating in the sinus node can penetrate and excite those regions, resulting in atrial contraction and function.28
Although the principles of the Maze-III operation are sound, and can be applied to any patient with AF, the problems associated with a highly invasive and morbid procedure remain. Therefore, over the last five years, there has been an enormous effort expended upon reducing the complexity and invasiveness of the operation such that it is more palatable to patients and their referring physicians. Improvements have been appearing in two main directions - ablative energy sources and minimal access techniques.
The most progress has been seen regarding energy sources. Seeking to replace the scissor and scalpel, surgeons have been using mostly hyperthermic technologies to heat and destroy myocardial tissue (although the venerable cryothermy is still being used by many and is a safe and fast technique with which to destroy atrial tissue). New instruments using unipolar dry radiofrequency, unipolar irrigated radiofrequency, microwave, bipolar dry radiofrequency, irrigated bipolar radiofrequency, ultrasound, and laser emissions have all been developed and assayed to differing extents. Each has its strengths and weaknesses but all represent a significant improvement in speed and safety over cutting with a knife or scissors and sewing the tissue together again. The remainder of this article examines microwave energy as one of those energy sources.
Microwave energy has been used for several decades to ablate tissues of all kinds. Extensive experience has been gained in using it to destroy tumors of the uterus, breast, skeletal muscle, liver, and prostate. Some of the earliest interest in its use on the heart, however, can be attributed to work by Haines et al. and Wonnell et al., who tried to ablate cardiac tissue using microwave-tipped catheters that were passed into the heart via the venous circulation.29,30
These attempts were met with technical difficulties, particularly cardiac chamber perforations, and so progress in this area was slow initially. When the AFx Corporation (Fremont, CA) designed and released the Flex 2 'LynxÔÇÖ device, focus was shifted from a catheter-based, small diameter, forward-firing probe to a longer, rigid, topically applied device that emitted radiation over a 2cm length (see Figure 1). The linear lesion created by the Flex 2 could be made either inside the empty heart (endocardial) or outside the full, beating heart (epicardial). Its efficacy and safety were demonstrated during the MICRO-STAF and MICRO-PASS trials by Knaut and his co-workers,31 which have now enrolled more than 200 patients and have documented about an 80% cure rate of permanent fibrillation with no morbidity related to the ablation procedure.
The Flex 2 (and the newer Flex 4 and Flex 10 devices) emits microwave radiation at 2.54GHz. It penetrates the tissue to excite water molecules, creating dielectric heating by enhanced vibration and friction of the molecules. Studies on muscle tissue phantoms have shown that the depth of penetration is governed by the generator power and the time of application (see Figure 2). It has been shown that tissue radiated by microwave energy undergoes coagulation necrosis and after a healing period is replaced by non-conductive scar (see Figure 3).
In contrast to other energy sources such as radiofrequency, microwave does not heat tissue by conduction. Therefore it does not require direct tissue contact in order to create a lesion. Microwave radiation, like any other energy source, does follow the rule of squares wherein energy density falls off with the square of the distance from the antenna. This means that the energy delivered to the tissue will be less if the antenna is separated from its target by an insulator such as air or fat (see Figure 4). Although simulated and in vitro studies have shown that power applied at the manufacturerÔÇÖs recommended settings should produce a lesion between 7-9mm deep, even when there is blood flowing along the opposite surface of the tissue, there has not yet been a method developed by which the operator can determine that a lesion is truly transmural in realtime. This is a technical problem that plagues all unipolar ablation methods at this time (for a more detailed review of the physics of microwave radiation the reader is referred to Williams et al.32,33).
Methods of Application
The newer Flex 4 (see Figure 5) and Flex 10 (see Figure 6) probes have overcome some of the technical limitations of the Lynx/Flex 2. Chiefly, the Flex 4 and 10 have flexible antenna guides and shafts that permit the operator to place them against heart muscle from either within the heart (endocardial) or on its outer surface (epicardial). Although the early experiences were all gathered using an endocardial approach, as experience has grown with these devices, surgeons have been creating lesions using both methods quite frequently.
Whereas the endocardial approach offers the advantage of direct visualization and access to all intra-atrial structures, it does add to cardiopulmonary bypass time, cross-clamp ischemia time, and sometimes is plagued by difficulties in visualization of the intracardiac targets. The epicardial approach offers the advantage of beating heart access and therefore applicability to any patient without opening the heart. There is also the theoretical advantage to ablating inward toward the intracardiac blood pool, rather than toward juxtaposed structures, greatly increasing the safety margin of the ablation. Epicardial ablation does take longer for each lesion, however, and also does not appear to penetrate as well. Also, not all structures can be accessed directly from an epicardial approach, such as the right atrial isthmus and the left atrial region near the mitral valve apparatus.
There are some unique properties about the Flex 10, however, that allow the operator to position it around the PVs without opening the chest. Because the ablating element can be positioned over any of 10 different locations along its sheath, it is not necessary to move and reposition the device between energy applications. Because the antenna is shielded, energy is emitted in only one direction, protecting collateral structures from injury. Finally, because the device is 9mm in diameter, it has recently been used to perform ablation through the closed chest using thoracoscopic techniques.34
Endoscopic Ablation of AF
Since 2002, the Cardiothoracic Surgery at the University of Massachusetts has performed the completely endoscopic ablation of AF on 26 patients. Most of these patients (67%) presented with paroxysmal AF, with an average duration of 79 months. They were an average of 60 years old (range 39-82) and 67% were male. Half had undergone multiple cardioversions and 46% had failed a minimum of two attempts at controlling their arrhythmia with anti-arrhythmic drugs. Their left atria were an average of 49mm in diameter (range 27-100) and their average ejection fraction was 50% (range 10-65). None had had prior chest or heart surgery.
All procedures were performed with the Flex 10 device. Briefly, the patient was brought to the operating room and general anesthesia was induced. A double-lumen endotracheal tube was used to permit separate lung ventilation as required during the procedure. The patient was positioned supine with the arms out at 90┬║ to the side. The anterior and lateral chest, abdomen and groins were sterilely prepped and draped.
The right lung was then deflated and three access ports into the right hemithorax created. Using 5mm cameras and operating instruments, the right pericardium was visualized and opened longitudinally parallel and about 2cm anterior to the phrenic nerve. This permitted visualization of the aorta, the superior vena cava, the right atrium, the inferior vena cava, and the right PVs.
In order to pass the Flex 10 around the back of the heart and encircle the PVs, it was necessary to access and instrument the transverse and oblique sinuses. As the entrance to each is guarded by a fold of pericardium under the superior and inferior vena cava, these folds were taken down using gentle blunt dissection with a 5mm instrument. Each sinus was then instrumented with a thin rubber catheter as a guide. Once the guide catheters were placed, the instruments were removed from the right side and the right lung reinflated.
The left lung was then deflated and a mirror-image pattern of access ports created. Again, using 5mm instruments, the left side of the pericardium was opened posterior to the phrenic nerve and the left atrial appendage visualized. The two guide catheters were then retrieved under direct vision and their tips delivered outside the chest where they were sutured to each other. Traction on the right-sided ends of the catheters brought the tips back into the chest and now the PVs were completely encircled.
The Flex 10 device was then attached to the transverse sinus catheter and delivered around the PVs by placing traction on the oblique sinus catheter. Once position and orientation were confirmed by direct visualization from both the right and left sides, an ablation 'boxÔÇÖ was created by applying microwave energy at each of the ten separate locations for 90 seconds at 65 watts. Lesion creation was confirmed by inspection of the areas that were visible.
In order to prevent any possibility of failure due to left atrial flutter, a lesion is created along the back wall of the left atrium connecting the PV 'boxÔÇÖ to the left atrial appendage. This lesion was performed in lieu of the mitral annular lesion of the Maze-III operations, as it is not yet possible to perform a mitral lesion from the epicardium without likely injuring the coronary sinus and/or circumflex coronary artery.
Once the appendage lesion was created, the appendage was amputated with an endoscopic stapler. A small chest tube was placed in either pleural space and the lungs re-expanded. The instruments were withdrawn and the wounds closed and bandages applied. All patients were awoken in the operating room and transferred to the intensive care unit for observation. The chest tubes were removed later that day. The median length of stay was four days, although 47% stayed three days or less.
There was a 67% incidence of AF immediately after surgery and during the recuperative phase. Therefore all patients were maintained on their anti-arrhythmic drugs (usually amiodarone) and their anti-coagulation for at least three months. Follow-up was obtained at one, three, six, nine, and 12 months after surgery and the medications were stopped if there was no AF for at least a three-month consecutive period.
The success rate of achieving sinus rhythm continued to improve over time. By six and nine months follow-up, 100% of patients were in either a paced or sinus rhythm.
Microwave ablation is a familiar and mature technology. It has been used thousands of times to ablate cardiac tissue and is safe and easy to handle. Recent innovations in delivery vehicles have produced a device that is long enough to completely surround all of the pulmonary veins at once as well as thin enough to be introduced through a 10mm operating port. This has allowed the development of a new procedure using microwave ablation to treat AF in a safe and effective manner without opening the patientÔÇÖs chest, arresting the heart or making cardiac incisions. This type of procedure will surely enhance patient and referring physician acceptance of a surgical approach to the treatment of AF, as well as make surgeons more comfortable and more willing to offer such a therapy. Such treatments are aticipated to be used early for patients just presenting with AF, rather than subjecting them to drug trials and life-long anticoagulation.