Surgical treatment of atrial fibrillation (AF) remains the most effective therapy for the treatment of this morbid arrhythmia. Success rates vary depending on the approach, but Dr. Cox, creator of the Maze III procedure, has reported efficacies as high as 98%.1 The Maze III procedure, regarded as the gold standard, is based on the idea that AF is caused by multiple wandering wavelets of reentry, and that strategically placed lesions will disrupt those wavelets from perpetuating, thus restoring sinus rhythm. Perhaps fortuitously, this procedure also addresses other potential causes of AF, including focal areas of automaticity and micro-rotors of reentry by nature of its isolation of the pulmonary veins.2,3 However, despite excellent results with the Maze III procedure, it has not yet been widely adopted. Reasons for this include the complex nature of the procedure and its invasiveness (sternotomy and cardiopulmonary bypass). Recently, many authors have reported performing various surgical modifications of the Maze III in a variety of patient populations with encouraging results,4-7 which has lead to a significant increase in the number of procedures being performed. The primary modifications that have resulted in an increasing acceptance include: reducing the number of lesions performed and utilizing energy sources to create lesions and thus decrease bleeding risk, shortening procedure time and an applying lesions from the epicardium on the beating heart. Most of these cases have been performed using different types of radiofrequency energy, though other energy sources have been utilized as well. Radiofrequency energy is an attractive energy source since it has a long clinical history for cardiac ablation and has generally proven to be very safe. It is also advantageous due to its relatively inexpensive nature, and is adaptable via a number of different approaches. Herein is a description of one radiofrequency system used surgically for treatment of atrial fibrillation, as well as an overview of a center s clinical experience. The Cobra Radiofrequency Surgical System (Boston Scientific, San Jose, California) comes in three different forms: Cobra, Cobra Cooled, and Cobra Bipolar. The Cobra is a flexible 7-electrode monopolar RF catheter (Figure 1A). Each electrode can be activated individually, but when adjacent electrodes are concurrently activated, a linear lesion is created. The number of electrodes utilized depends on the desired length of the lesion; however, lengths from 1 cm to 9 cm can be achieved. The Cobra Cooled (Figure 1B) is a flexible irrigated temperature-controlled device with the same electrode configuration as the regular Cobra. The continuous, contained irrigation of the device allows for deeper lesion formation and thus is optimized for epicardial ablation. The Cobra Bipolar is a clamp device that provides focused bipolar ablation between the sides of the clamp. This device is ideal for rapid epicardial isolation of the pulmonary veins on the beating heart. Both the Cobra and Cobra Cooled are approved for use in the United States; however, the Cobra Bipolar is currently pending FDA approval. In addition, there have been clinical uses of all three devices outside the U.S. with early encouraging results. Screening. In order for a patient to undergo an ablation for atrial fibrillation, the patient must be having a cardiac procedure that requires the use of cardiopulmonary bypass, and be in a state of chronic AF for at least six months. Patients with AF for at least three months whom have also failed at least two attempts at cardioversion are also eligible. Obviously, there is consideration given of an individual patient in deciding whether or not they undergo an ablation. While we have no absolute contraindication, important factors to consider include: previous complications from AF, tolerance of anticoagulant and antiarrhythmic medications, the complexity and time requirement of the concomitant procedure, age of the patient, and other comorbid conditions. Intraoperative: Lesion Set. All patients undergo isolation of all four pulmonary veins. This is generally accomplished by creating one circumferential lesion in the posterior left atrium that contains all four pulmonary veins. On occasion, a bilateral isolation is performed so that the right and left veins are isolated separately. In these cases, a lesion is then performed to connect both sides to prevent the formation of an atypical left-sided atrial flutter. Finally, in both approaches a lesion is created to the mitral valve annulus. The left atrial appendage is then both anatomically and electrophysiologically isolated. This is accomplished either by simply amputating and repairing the defect or creating a lesion at its base and then closing the orifice. Right-sided lesions are not routinely performed unless the patient has a known history of atrial flutter, or if the planned procedure requires entry into the right atrium. In these cases, a standard isthmus lesion is created from the inferior vena cava to the tricuspid annulus. All lesions are created using the Cobra system as previously described. Ablations are performed at 80 °C for 60 seconds. Retrograde cardioplegia is empirically given during the creation of the mitral valve lesion to protect the circumflex artery. Postoperative. Patients are routinely started on IV amiodarone in the operating room and transitioned to an oral form as tolerated. The patient s protocol involves a dose of 200 mg/day of amiodarone for twelve weeks in addition to anticoagulation. At the end of the twelve-week postoperative period, the patients in sinus rhythm receive an echo to document left atrial contractility. If this is present, assuming there is no other indication, the anticoagulant and antiarrhythmic agents are discontinued. Results. From October 1999 to December 2002, we performed endocardial ablations on 99 patients. The patients were on average 64 ± 12 years old, 54% male, and had been in atrial fibrillation for an average of 8 ± 8 years. The patients had an average left atrial diameter of 7.1 cm ± 5.1 cm. Mitral valve or combined mitral valve procedures were performed in 84%, the others had aortic, congenital or coronary procedures. Patients were followed to an average of 319 days ± 256 days; with results varying over time (Figure 2), and with better results the longer the patients had been out of surgery. Interestingly, 73% of the population experienced AF early in the postoperative period, despite a 74% long-term success. There were no preoperative variables that predicted a successful procedure though patients with left atria larger than 8 cm tended to remain in atrial fibrillation. Surgical treatment of AF with endocardial RF during concomitant cardiac surgery is an exciting modality that has greatly simplified the options available to these patients. In our series we have found that the procedure is successful 74% of the time, which is very encouraging for many reasons. First, this data demonstrates that patients may not need a full maze for a successful procedure. Our results are similar to those of the largest compilation of patients that had a Maze III in addition to a mitral valve procedure with a 76% rate of success.8 In addition to the simplification of the procedure, this also further validates the claim that AF is an arrhythmia arising from the left atrium, specifically near the pulmonary veins. The data are also exciting in that excellent results can be obtained with the use of RF and thus the procedure is greatly simplified, theoretically reducing the risk for bleeding. This would likely translate into a greater number of surgeons that can offer this modality to their patients. However, our mortality was higher than we expected. Careful review of these deaths did not reveal any correlation either to the ablation procedure or their AF. In fact, five of patients expired in sinus rhythm. This experience has caused us to become more selective about our patients in terms of considering other comorbidities. One of the most important lessons is the extremely high postoperative AF rate, despite excellent long-term results. Initially, one can be very discouraged by early AF, but it is important to expect this and continue with care according to your protocol. This early AF rate explains our desire to keep patients anticoagulated for the first three months. These results provide the foundation as we proceed with this type of procedure, using minimally invasive techniques from the epicardium, as well as dealing with patients with AF as the sole indication for surgery. It is expected that mitral valve patients are probably the most difficult to treat for AF, given their extensive atrial pathology from the associated valve disease. Thus, one can hope that as we move out of this patient population we will see even better success with a more limited procedure. This transition will be further facilitated by the new technology that allows ablation from the epicardium on the beating heart, something that was previously impossible with traditional RF and cryo sources. Ultimately patients with AF may be treated with a totally endoscopic surgical procedure performed as an outpatient with a procedure time less than two hours.