The Cox-maze procedure is the most effective surgical. Intraoperative Radiofrequency Maze Ablation for Atrial Fibrillation: The Berlin Modification

Transcription

1 Intraoperative Radiofrequency Maze Ablation for Atrial Fibrillation: The Berlin Modification Miralem Pasic, MD, PhD, FETCS, Peter Bergs, MD, Peter Müller, MD, Michael Hofmann, MD, Onnen Grauhan, MD, Hermann Kuppe, MD, and Roland Hetzer, MD, PhD, FETCS Deutsches Herzzentrum Berlin, Berlin, Germany Background. The Cox-maze procedure combined with an operation for organic heart disease is highly successful in the elimination of chronic atrial fibrillation. However, it prolongs significantly the aortic cross-clamp and operating time. In this study, a simplified left atrial maze procedure, which is a short procedure performed using a surgical radiofrequency ablation probe, is added to elective open heart procedures in patients with atrial fibrillation. Methods. Forty-eight adults with atrial fibrillation (duration, 6 months to 36 years) underwent elective open heart operations (isolated valve procedures or coronary artery bypass grafting, n 27 patients; combined procedures, n 21 patients) combined with intraoperative radiofrequency ablation of the left atrium. The postoperative follow-up period ranged from 1 to 11 months (mean, 4 months). Possible predictors for persistent postoperative atrial fibrillation were determined among 40 variables by univariate and multivariate analyses. Results. Intraoperative radiofrequency ablation prolonged the aortic cross-clamp time for 6 to 14 minutes (mean, 11 minutes). Freedom from atrial fibrillation was 100% intraoperatively, 25% at 1 week after operation (12 of 48 patients), 59% at 1 month postoperatively (16 of 27 patients), 64% at 3 months postoperatively (16 of 25 patients), and 92% at 6 months postoperatively (12 of 13 patients). The only predictor of postoperative atrial fibrillation was the presence of coronary artery disease (odds ratio, 7.5; 80% confidence interval, ). Conclusions. Intraoperative radiofrequency ablation of the left atrium combined with an operation for organic heart disease effectively eliminates atrial fibrillation without significant prolongation of the aortic cross-clamp and operative time. The presence of coronary artery disease decreases the success rate during the first 6 postoperative months. (Ann Thorac Surg 2001;72: ) 2001 by The Society of Thoracic Surgeons Presented at the Thirty-seventh Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Jan 29 31, Address reprint requests to Dr Pasic, Deutsches Herzzentrum Berlin, Augustenburger Platz 1, D Berlin, Germany; pasic@dhzb.de. The Cox-maze procedure is the most effective surgical treatment for patients with chronic atrial fibrillation [1 4]. The method can be combined with an operation for organic heart disease or can be performed as an isolated surgical procedure for patients with lone atrial fibrillation refractory to medical therapy [1 4]. Our experience with the Cox-maze operation correlates well with the results from most centers with a long-term success of 80% to 90% in restoring the sinus rhythm [5 7]. However, from a surgical point of view, it is a demanding procedure that prolongs significantly the aortic cross-clamp and operating time. Therefore, this procedure is not widely accepted. The next logical step is to find a modification of the original Cox-maze procedure that will have a similar success rate without the described drawbacks. Intraoperative radiofrequency ablation is a novel surgical principle for the treatment of atrial fibrillation in combination with a standard open heart operation [8 15]. It is based on the original concept of the maze procedure developed and introduced by James Cox [1]. Application of radiofrequency current replaces the incisions and sutures of the standard maze technique. There is a spectrum of modifications regarding the types of surgical probe used for ablation, modes of application (endocardial or epicardial), and the directions of the maze lines [8 15]. Since 1999, we in Berlin have used an easy method of intraoperative radiofrequency ablation performed with a surgical radiofrequency ablation probe. The method can be carried out in a short time and was added to elective open heart procedures in patients with organic heart disease and atrial fibrillation. In this study we report our preliminary experience with this simplified maze procedure. Patients and Methods Patients Between September 1999 and July 2000, 48 consecutive adults with atrial fibrillation underwent elective open heart operations (isolated valve procedures or coronary artery bypass grafting, n 27 patients; combined procedures, n 21 patients) (Tables 1 and 2) combined with intraoperative radiofrequency ablation of the left atrium. There were 28 females and 20 males (age range, 33 to 78 years; average age 64 years). The preoperative atrial fibrillation lasted from 6 months to 36 years with a mean 2001 by The Society of Thoracic Surgeons /01/$20.00 Published by Elsevier Science Inc PII S (01)

2 Ann Thorac Surg PASIC ET AL 2001;72: RADIOFREQUENCY MAZE ABLATION 1485 Table 1. Demographic Data Patient number 48 Age (years) Sex (male/female) 20/28 Height (cm) Weight (kg) Body surface area (m 2 ) of 7 years. Four patients had previous pacemaker implantations. The mean left ventricular ejection fraction assessed by ventriculography and echocardiography was 55%, (range, 27% to 73%) (Table 3). The modified maze procedure was offered to all patients with atrial fibrillation who were scheduled for elective procedures, to be performed concomitantly with open heart operations by two of the authors (MP and RH). The exclusion criteria for combined modified maze procedure and open heart operation were left ventricular function less than 25%, extensive and diffuse coronary artery disease, acute valve endocarditis, calcification of the left atrium, extensive calcification of the mitral valve annulus requiring annular decalcification, and patients unwilling to undergo the procedure. Operative Technique The patient is placed in the standard supine position. The two large indifferent patch electrodes are placed on the patient s back. The indifferent electrodes and the surgical probe are connected to the radiofrequency generator with the connecting cables. The system is positioned according to the space available in the operating room. The operation is carried out by median sternotomy using separate caval cannulation, total cardiopulmonary bypass, mild systemic hypothermia of 32 C, and a current cardioplegic technique. Both venae cavae and the posterior wall of the left atrium are mobilized. If used, a transesophageal probe should be removed from the esophagus at this time. The radiofrequency maze operation is performed prior to the operation for organic heart disease. The radiofrequency maze operation (Berlin modification) is based on the principle of the left-atrial part of the Cox-maze procedure [1 4] combined with the method of isolation of the pulmonary veins popularized by Melo and colleagues [9, 10]. (Figs 1 and 2). The Berlin modification consists of the standard approach to the left atrium from the right side through the interatrial sulcus (Waterston groove) and four standard left atrial maze lines; two maze lines isolate the ostia of the pulmonary veins and the other two are connecting lines. Instead of the cut and sew technique, the maze lines are performed with a flexible surgical radiofrequency probe and the coagulation system (Coagulation ThermaLine System, Boston Scientific, San Jose, CA). The system consists of a cardiac ablation controller, the electrosurgical probe, connecting cables, and a generator of radiofrequency energy. The probe (ThermaLine probe, Boston Scientific, San Jose, CA) (Fig 3) has seven electrodes on the distal part that coagulate the tissue when in contact with the electrodes. During current delivery, the monopolar radiofrequency current flows through the tissue toward the indifferent electrodes. At the location of direct contact between the probe and the atrial endocardium, radiofrequency energy coagulates atrial tissue and creates a deep linear lesion in the atrial wall. The generator settings are 70 C for temperature and 2 minutes for duration of ablation of each line, (ie, the ablation for each line lasts 2 minutes at the local temperature of 70 C). The temperature is mea- Table 2. Types of Surgical Procedures Combined With the Simplified Maze Procedure and Intraoperative Data Type of Surgery No. of Patients Cross-Clamp Time (min) Perfusion Time (min) All procedures Mitral valve replacement Mitral valve repair Mitral valve replacement and coronary artery bypass grafting Mitral valve repair and coronary artery bypass grafting Mitral valve replacement and tricuspid valve repair Mitral valve repair and tricuspid valve repair Aortic valve replacement Aortic valve replacement and mitral valve replacement Aortic valve replacement and mitral valve repair Aortic valve replacement and coronary artery bypass grafting Aortic valve replacement and mitral valve replacement coronary artery bypass grafting Coronary artery bypass grafting Ascending aortic replacement and aortic valve repair Mitral valve repair and tricuspid valve repair and atrial septal defect closure Aortic valve replacement and mitral valve replacement and atrial septal defect closure Tricuspid valve repair and atrial reduction

3 1486 PASIC ET AL Ann Thorac Surg RADIOFREQUENCY MAZE ABLATION 2001;72: Table 3. Preoperative Data Left ventricular ejection fraction (%) echo Left ventricular ejection fraction (%) ventriculography Left ventricular end-diastolic diameter (mm) 55 8 Left ventricular end-diastolic pressure (mm Hg) 14 6 Fractional shortening (%) 30 8 Right ventricular ejection fraction (%) 56 9 Right ventricular end-diastolic diameter (mm) 30 4 Left atrial diameter (mm) 47 8 Left atrial diameter index (mm/m 2 ) a a Left atrial diameter/body surface area. Fig. 1. The left atrial part of the standard Cox-maze III procedure encompasses excision of the left atrial appendage, multiple incisions with isolation of the pulmonary veins and consecutive continuous sutures of the left atrial incisions, cryoablation of the dissected coronary sinus and mitral annulus. (The illustration is slightly modified and reprinted with permission from the American College of Cardiology [Journal of the American College of Cardiology 1998;32: ]) Fig. 2. In the modified method (the Berlin modification), the incisions and sutures of the standard maze technique are replaced by radiofrequency ablation lines (dashed lines). The line directions are slightly changed. The right and left pulmonary veins are isolated separately using two ablation lines instead of one encircling line as used in the standard maze procedure. The left appendage is not excised. sured by thermocouples integrated into the probe. The pre-set temperature usually is achieved during the first 10 to 15 seconds of ablation. During ablation the operative field should be kept as dry as possible; this can usually be achieved by using two suckers. Also the lungs are freed from blood by forced manual ventilation, and then the endotracheal tube is disconnected from the ventilator. Alternatively the pulmonary veins can be separately encircled with tapes and temporarily snagged. The contact between the electrode and the endocardium should be good, otherwise the impedance rises and current delivery is shut off. The impedance can also rise if there is blood around the probe. The heat may cause formation of small coagula, and in such cases the probe should be cleaned with wet gauze. A standard right-sided left atriotomy is made surgically parallel to the interatrial groove and then the radiofrequency maze ablation is performed. All radiofrequency maze lines are made endocardially from within the cavum of the left atrium. The first ablation maze line connects both ends of the surgical atriotomy, completing isolation of the right pulmonary veins (Fig 4A). The second ablation maze line isolates the left pulmonary veins (Fig 4B). This is an encircling line around the orifice of the pulmonary veins. If the veins are enlarged, the encircling line is completed by two applications of the probe. Both lines encircling the pulmonary veins should be placed approximately 10 mm from the venous orifices to prevent possible late stenosis of the pulmonary veins. The third ablation maze line (Fig 4C) connects the encircling line of the left pulmonary veins and the middle part of the posterior mitral valve annulus. The line should be placed inferiorly in order to prevent heat damage to the circumflex artery. The last incision (Fig 4D) connects the left-sided and the right-sided encircling lines of the pulmonary veins. It runs along the roof of the atrium in order to prevent possible damage to the esophagus or vagal nerves. The left atrial appendage is left intact, and is neither excised nor oversewn. If chronic thrombotic material is found within the appendage, it should be oversewn in order to prevent possible later embolization. At the end of the procedure, if there are some coagula on the maze lines, they are removed with a wet gauze and then the left atrium is flushed with saline. After finishing the ablation maze procedure, the planned operation for organic heart disease is completed. Postoperative Antiarrhythmic Medication The patients were not given any special prophylactic medication postoperatively for the prevention of atrial fibrillation. If a patient remained in sinus rhythm, the patient did not receive any antiarrhythmic medication.

4 Ann Thorac Surg PASIC ET AL 2001;72: RADIOFREQUENCY MAZE ABLATION 1487 Follow-up All patients were followed up by their cardiologists or family physicians for adjustment of medication and control of anticoagulation and rhythm. After the operation, the patients were prospectively followed up at 1, 3, and 6 months. The postoperative rhythm was evaluated with 12 lead electrocardiogram. However, 24-hour Holter monitoring was not performed on all patients. The postoperative follow-up ranged from 1 to 11 months, with a mean follow-up of 4 months. Statistics and Data Presentation All data are reported as mean standard deviation. After univariate analysis, variables with p less than 0.01 were analyzed by multivariate logistic regression to identify predictors for postoperative atrial fibrillation. A value of p less than 0.05 was considered to be significant. Fig. 3. A flexible surgical radiofrequency probe (Thermaline, Boston Scientific, San Jose, CA) can be bent manually to adjust it to the anatomy of the left atrium. The probe has seven coagulation electrodes at its distal part; six of them are 12.5 mm in length each and the tip electrode is only 8 mm long. There is a 2-mm distance between each electrode. Any combination of the electrodes can be selected for ablation. Early postoperative patients with junctional rhythm were treated with continuous intravenous administration of orciprenaline sulfate (Alupent; Boston Scientific, San Jose, CA). Patients with atrial fibrillation during hospitalization were treated with digoxin in combination with verapamil or digoxin in combination with sotalol or with sotalol or amiodarone alone. After discharge from the hospital, the same regimen of antiarrhythmics was recommended for patients with episodes of postoperative supraventricular tachyarrhythmias, and we also recommended tapering the medication if the rhythm became stable. If patients did not convert into sinus rhythm under medical therapy, a direct current shock was recommended. Before cardioversion, transesophageal echocardiography was necessary to exclude atrial thrombi. Postoperative Anticoagulant Medication Postoperatively, anticoagulation medication consists of heparin and then peroral phenprocoumon (Marcumar; Boston Scientific, San Jose, CA). If a patient did not need long-term anticoagulation for other reasons (eg, implantation of a mechanical valve), we recommended anticoagulation with phenprocoumon for at least 6 months. Thereafter, this medication could be discontinued and replaced with aspirin in patients with stable sinus rhythm by 24-hour Holter monitoring and normal or nearly normal left atrial contractions determined by echocardiographic examination. Results There were no perioperative deaths and no reoperations for bleeding. Only 5 patients (10%) needed postoperative ventilation longer than 24 hours. The aortic cross-clamp time needed for intraoperative radiofrequency ablation ranged from 6 to 14 minutes with a mean of 11 minutes. The early postoperative recovery was uneventful in all patients, but two patients died (4% hospital mortality) on postoperative days 16 and 26 after an initial uneventful course. The causes of death were stroke in 1 patient after mechanical mitral valve replacement with postoperative sinus rhythm, adequate anticoagulation, and no evidence Fig. 4. Dashed lines show the position of the radiofrequency maze lines of the Berlin modification in comparison to the standard surgical maze lines. (A) The first line completes the isolation of the ostia of the right pulmonary veins, (B) and the second maze line isolated the ostia of the left pulmonary veins. The pulmonary veins isolation lines are performed about 10 mm away from the ostia of the pulmonary veins. (C) The third maze line connects the encircling line around the ostia of the left pulmonary veins (the isolation line of the left pulmonary veins) and the middle of the posterior part of the mitral valve annulus. (D) The fourth maze line is a connecting line between the two lines that isolate the ostia of the pulmonary veins. The line lies on the roof of the left atrium.

5 1488 PASIC ET AL Ann Thorac Surg RADIOFREQUENCY MAZE ABLATION 2001;72: of atrial thrombosis using transesophageal echocardiography, and in the other patient late pericardial tamponade after coronary bypass revascularization was suspected. Two patients developed fluid retention with leftsided pleural effusion, which was treated with diuretics. All 48 patients left the operating room with sinus rhythm or atrial pacing, but only 12 patients had stable sinus rhythm during the first postoperative week. The other patients demonstrated electrocardiographic characteristics of sinus node dysfunction, such as severe sinus bradycardia, sinus pause or sinus arrest, sinoatrial exit block, atrial tachyarrhythmias, alternating periods of atrial bradyarrhythmias and tachyarrhythmias, and atrial fibrillation. Temporary wires were used for temporary pacing if necessary or to overdrive the atrium. Cardiac rhythm was continuously monitored after the operation until stable rhythm returned. The patients stayed at our institution for 3 to 7 days until the rhythm stabilized. Thereafter, patients were transferred back to a referral cardiology department. Five patients (11%) received a pacemaker postoperatively because of bradycardia and persistent sinus node dysfunction. The occurrence of counterclockwise atrial flutter was not recorded in patients after their discharge from hospital. At 3 and 6 months postoperatively, all patients without atrial fibrillation had a sinus rhythm confirmed by clearly visible P waves in surface electrocardiogram, and 3 patients with pacemaker implantation still had sinus node dysfunction with pacemaker rhythm. The freedom from atrial fibrillation was 100% intraoperatively, 25% at 1 week postoperatively (12 of 48 patients), 59% at 1 month postoperatively (16 of 27 patients), 64% at 3 months postoperatively (16 of 25 patients), and 92% at 6 months postoperatively (12 of 13 patients) (Fig 5). The function of the left atrium assessed by transesophageal echocardiography demonstrated contractility and atrial transport of different grades of both atria in all patients according to the evidence of Doppler-recorded A waves. Forty preoperative and intraoperative variables were analyzed to determine predictors for postoperative atrial fibrillation. After univariate analysis, only four variables had p less than 0.01 (presence of mitral valve stenosis, coronary artery disease, duration of cardiopulmonary bypass time, duration of operation time [Table 4]) and were used for multivariate logistic regression. This identified the presence of coronary artery disease as the only variable that determined postoperative recurrence of atrial fibrillation ( p ; odds ratio 7.5; 80% confidence interval, ). Comment We describe an easy method of intraoperative radiofrequency ablation, a simplified left-atrial maze procedure Fig. 5. Postoperative freedom from atrial fibrillation (% of patients). (OP. operation.) Table 4. P Values of Univariate Analyses for Determination of Predictors for Postoperative Atrial Fibrillation Age (years) Gender (female, male) Height (cm) Weight (kg) Body surface area (m 2 ) Mitral valve disease (yes, no) 1.0 Mitral regurgitation (yes, no) Mitral stenosis (yes, no) Previous mitral valve surgery (repair, replacement, no) Tricuspid valve disease (yes, no) 1.0 Tricuspid incompetence (yes, no) 1.0 Aortic valve disease (yes, no) 1.0 Aortic regurgitation (yes, no) Aortic stenosis (yes, no) 1.0 Coronary artery disease (yes, no) Previous myocardial infarction (yes, no) Previous coronary artery bypass grafting 1.0 (yes, no) Ascending aortic aneurysm (yes, no) 1.0 Atrial septal defect and persistent foramen ovale 1.0 (yes, no) Pulmonary hypertension (yes, no) Mitral valve repair (yes, no) Mitral valve replacement (yes, no) 1.0 Tricuspid valve repair (yes, no) Aortic valve repair (yes, no) 1.0 Aortic valve replacement (yes, no) Coronary artery bypass grafting (yes, no) Ascending aortic replacement (yes, no) 1.0 Closure of atrial septal defect or persistent 1.0 foramen ovale (yes, no) Redo-operation (yes, no) 1.0 Aortic cross-clamp time (min) Cardiopulmonary bypass time (min) Operative time (min) Duration of atrial fibrillation (years) Echocardiography-assessed left ventricular ejection fraction (%) Left ventricular end-diastolic diameter (mm) Echocardiography-assessed right ventricular ejection fraction (%) Right ventricular end-diastolic diameter (mm) Left atrial diameter (mm) Left atrial diameter index (mm/m 2 ) Ventriculography-assessed left ventricular ejection fraction (%) Left ventricular end-diastolic pressure (mm Hg)

6 Ann Thorac Surg PASIC ET AL 2001;72: RADIOFREQUENCY MAZE ABLATION 1489 performed with a flexible surgical radiofrequency probe. Our preliminary experience demonstrated results comparable with those of the standard Cox-maze III procedure with a very high percentage of patients with stable sinus rhythm at 6 months postoperatively. The presence of coronary artery disease decreases the success rate during the first 6 months postoperatively. The procedure is simple and can be carried out in a short time. It eliminates the need for multiple incisions in the left atrium and does not prolong the aortic cross-clamp time significantly as does the standard technique. The method used is one of several modifications of the intraoperative radiofrequency ablation technique that have been developed recently [8 15]. We observed a relative high recurrence rate of atrial fibrillation early postoperatively. We believe that the possible reason is initial inadequate ablation of the atrial tissue producing incomplete maze lines. It can be postulated that about 3 to 6 months are needed for complete healing of the lines. Then, after complete scarring of the maze lines, they become a total barrier for an electrical atrial impulse and eliminate atrial fibrillation. Sinus Node Dysfunction After Radiofrequency Maze Ablation Mechanisms involved in the disturbance of cardiac rhythm after the Cox-maze surgical procedure occur as either primary events of the maze procedure itself or as general postoperative complications of open heart operation [5]. The surgically performed bi-atrial Cox-maze III procedure includes isolation of the pulmonary veins and multiple incisions in both atria, corresponding to partial autotransplantation and partial denervation of the heart. During the follow-up, gradual improvement of sinus node function and atrial contractions was observed with significant functional normalization 1 year after the Coxmaze procedure, corresponding to functional reinnervation and recovery of the autonomic nervous system [5, 6]. One of the advantages of this simplified approach is that sinus node function cannot be jeopardized by the placement of the maze lines and it is expected that the incidence of postoperative pacemaker implantation should be lower than after the Cox-maze surgical procedure. However, other possible factors that can cause sinus node dysfunction are not eliminated, such as intraoperative traction on the heart with retractors and manipulation, suture injury, placement of caval cannulas, inadequate atrial preservation during cardiopulmonary bypass, hemorrhage, ischemia, necrosis, disturbance of autonomic neural control, humoral and electrolyte disbalance, and the effect of cardiovascular drugs, such as sympatholytic agents or antiarrhythmics [5]. Furthermore, the patients can already have sinus node dysfunction before the operation because of atrial fibrosis that follows chronic atrial fibrillation. Latent concomitant dysfunction can be masked by atrial fibrillation, being recognized only postoperatively after sinus rhythm is established. This could not have been excluded only by preoperative electrocardiographic findings or anamnestic data [5]. Therefore, it should not be expected that all postoperative patients will have normal sinus node function after radiofrequency maze ablation. Policy Not to Exclude or to Excise Left Atrial Appendage The main difference between both the original Cox-maze procedure and the isolation of the pulmonary veins and our modification is that the left atrial appendage is left intact and is not excised or oversewn. We do not excise the appendage for several reasons. First, it simplifies the surgical procedure significantly, eliminating the need for cutting and oversewing of tissue in this region, which is sometimes very fragile and has the circumflex artery in the vicinity. Furthermore, in patients with sinus rhythm, the left atrial appendage is important for mechanical function of the left atrium. Also, we do not routinely exclude the left atrial appendage in other patients with atrial fibrillation who do not undergo the maze procedure. If chronic thrombotic material is found within the appendage, it should be oversewn in order to prevent possible later embolization. However, in the reported group of patients we saw 2 patients with left atrial thrombi, but the atrial appendage was always free from coagula, and therefore the appendage was not oversewn. Possible Complications of the Method Although we did not observe any complications regarding the procedure per se, there are some possible theoretic complications, such as perforation of the left atrial wall and bleeding, or acute or chronic damage of the neighboring structures and organs (eg, perforation of the esophagus, damage of the n. vagus or coronary arteries and stenosis of the ostia of the pulmonary veins). However, these possible complications can be prevented by correct placement of the maze lines. The connecting maze line between the two encircling lines of the left and right pulmonary vein should be performed on the roof of the left atrium and not in the middle between the veins in order to stay away from the neighboring organs. The encircling lines of the ostia of the pulmonary veins should be approximately 10 mm away from the ostia to prevent possible late stenosis of the pulmonary veins. To prevent possible damage to the circumflex artery, the connecting maze line between the encircling line around the ostia of the left pulmonary veins (the isolation line of the left pulmonary veins) and the mitral valve annulus should be directed to the middle of the posterior part of the mitral valve annulus. Furthermore, the echocardiographic probe should be removed from the esophagus during the procedure to exclude possible interference with radiofrequency energy and possible esophageal injury. Limitations of the Study We reported our preliminary experience with this simplified method and therefore, our study has several limitations. These limitations comprise a small number of patients without a control group, rhythm evaluation using only 12 lead electrocardiogram without 24-hour electrocardiogram in all patients, and short follow-ups. This study can only be considered an observational study

7 1490 PASIC ET AL Ann Thorac Surg RADIOFREQUENCY MAZE ABLATION 2001;72: because of its limitations, and a prospective randomized study should be the next step. However, we were able to demonstrate that the method is easy to perform, and yields a very good 6-month rhythm control result. In conclusion, the modified maze procedure using intraoperative radiofrequency maze ablation of the left atrium combined with an operation for organic heart disease effectively eliminates atrial fibrillation without significant prolongation of the aortic cross-clamp and operative time. Because of the excellent results of this study, we have changed our standard surgical policy by adding this modified, short-lasting left atrial maze procedure to elective open heart procedures in all patients with atrial fibrillation. However, the long-term results and comparisons to other ablation systems (ie, microwave, cryo, and laser) are needed. We thank Anne Gale for editorial assistance, Helge Haselbach and Annette Gaussmann for the illustrations, and Julia Stein for statistical analyses. References 1. Cox JL, Schuessler RB, D Agostino HJ Jr, et al. The surgical treatment of atrial fibrillation. III. Development of a definitive surgical procedure. J Thorac Cardiovasc Surg 1991;101: Cox JL, Boineau JP, Schuessler RB, Jaquiss RDB, Lappas DG. Modification of the maze procedure for atrial flutter and atrial fibrillation. I. Rationale and surgical results. J Thorac Cardiovasc Surg 1995;110: Cox JL. The surgical treatment of atrial fibrillation. IV. Surgical technique. J Thorac Cardiovasc Surg 1991;101: Cox JL, Jaquiss RDB, Schuessler RB, Boineau JP. Modification of the maze procedure for atrial flutter and atrial fibrillation. II. Surgical technique of the maze III procedure. J Thorac Cardiovasc Surg 1995;110: Pasic M, Musci M, Siniawski H, Edelmann B, Tedoriya T, Hetzer R. Transient sinus node dysfunction after the Coxmaze III procedure in patients with organic heart disease and chronic fixed atrial fibrillation. J Am Coll Cardiol 1998; 32: Pasic M, Musci M, Siniawski H, Grauhan O, et al. The Cox-maze procedure: parallel normalization of sinus node dysfunction, improvement of atrial function, and recovery of the cardiac autonomic nervous system. J Thorac Cardiovasc Surg 1999;118: Pasic M, Musci M, Edelmann B, Siniawski H, Bergs P, Hetzer R. Identification of p waves after the Cox-maze procedure: significance of right precordial leads V3R through V6R. Ann Thorac Surg 1999;67: Chen M-C, Guo B-F, Chang J-P, Yeh K-H, Fu M. Radiofrequency and cryoablation of atrial fibrillation in patients undergoing valvular operations. Ann Thorac Surg 1998;65: Melo J, Adragao PR, Neves J, et al. Electrosurgical treatment of atrial fibrillation with a new intraoperaive radiofrequency ablation catheter. Thorac Cardiovasc Surg 1999;47(Suppl): Melo J, Adragao PR, Neves J, et al. Surgery for atrial fibrillation using radiofrequency catheter ablation: assessment of results at one year. Eur J Cardiothorac Surg 1999;15: Benussi S, Pappone C, Nascimbene S, et al. A simple way to treat chronic atrial fibrillation during mitral valve surgery: the epicardial radiofrequency approach. Eur J Cardiothorac Surg 2000;17: Melo J, Adragao P, Neves J, et al. Endocardial and epicardial radiofrequency ablation in the treatment of atrial fibrillation with a new intraoperative device. Eur J Cardiothorac Surg 2000;18: Patwardhan AM, Dave HH, Tamhane AA, et al. Intraoperative radiofrequency microbipolar coagulation to replace incisions of maze III procedure for correcting atrial fibrillation in patients with rheumatic valvular disease. Eur J Cadiothorac Surg 1997;12: Mottkamp H, Hindricks G, Hammel D, et al. Intraoperative radiofrequency ablation of chronic atrial fibrillation: a left atrial curative approach by elimination of anatomic anchor reentrant circuits. J Cardiovasc Electrophysiol 1999;10: Sie HT, Beukema WP, Ramdat Misier AR, Elvan A, Ennema JJ, Wellens HJ. The radiofrequency modified maze procedure. A less invasive surgical approach to atrial fibrillation during open-heart surgery. Eur J Cardiothorac Surg 2001;19: DISCUSSION DR MICHAEL ARGENZIANO (New York, NY): I would like to begin by thanking the program committee for inviting me to discuss this paper and by congratulating Dr Pasic and his colleagues for an excellent presentation and for sending me a copy of their manuscript in advance. In this paper, Dr Pasic and colleagues report their experience with a modified left atrial radiofrequency ablation procedure performed concomitant with coronary bypass or valvular surgery, or both. Despite the lack of randomization and absence of control groups, the authors are to be congratulated on this impressive clinical series, which demonstrates the feasibility, safety and efficacy of surgical endocardial radiofrequency ablation of atrial fibrillation. I have several comments and questions for the authors. Other groups have reported inferior success rates for radiofrequency isolation in patients with left atria larger than 6 or 7 cm, and you state in your paper that you excluded patients with giant left atrium, but provide no definition of this entity. Our group at Columbia, which recently reported its results in a similar number of patients, demonstrated comparable success rates at 6 months, as shown in this slide, but found no relationship between left atrial size and the rate of successful ablation. What is your definition of giant left atrium, and do you think that these patients are untreatable with this method? Second, your method of radiofrequency ablation is different from that used by our group and others, as you have noted, in that you isolate the pulmonary veins and then add only a lesion from the left pulmonary vein isolation line to the mitral annulus. As we have reported, we make a single encircling lesion around the pulmonary veins and a lesion to the mitral annulus, but also isolate the left atrial appendage and add a connecting line from the pulmonary vein isolation line to the appendage isolation line. We add these appendage lesions in order to exclude arrhythmogenic foci within the left atrial appendage and to prevent arrhythmias based on reentrant conduction around the base of the appendage. Did you skip the appendage isolation and connecting lines because you consider them unnecessary or merely to save time? Also, with respect to your technique of

8 Ann Thorac Surg PASIC ET AL 2001;72: RADIOFREQUENCY MAZE ABLATION 1491 separate isolation of left and right pulmonary veins, which actually takes longer than isolation of all four veins with a single encircling lesion, do you think that the small intervening strip of posterior left atrium that is thus spared makes any real contribution to left atrial contractile function? This slide depicts a specimen taken from one of our early patients who underwent combined mitral repair and radiofrequency ablation of the left atrium. This patient had a normal preoperative and intraoperative transesophageal echocardiogram, revealing no thrombus in the left atrium or appendage. We can see here that the left atrial appendage is in fact filled with thrombus. Because of concern about the potential for bleeding, we now routinely exclude the appendage mechanically by ligating the orifice from within the atrium with a purse string suture rather than by resecting it as we did here. Regarding your patient who died of a stroke postoperatively, what was the rhythm before death, and do you think it is possible that the source of the embolus could have been from the left atrial appendage? Finally, you stated that you did not perform right atrial ablation because surface mapping studies have indicated that atrial fibrillation is usually associated with left-sided arrhythmogenic foci. Although this may be the case, it has also been shown that patients with atrial fibrillation are predisposed to atrial flutter as well, which in fact is often initiated by right atrial foci. In our series, 8 of 42 patients had right-sided as well as left-sided lesions created. None of these patients had postoperative atrial flutter, while 7 of 34 patients receiving only left-sided lesions had at least one episode of atrial flutter. Two of these patients, in fact, on medium term follow-up are in persistent atrial flutter. It is therefore our current practice to add right-sided lesions in patients with a history of atrial flutter or those who are undergoing concomitant right-sided procedures. Did any of your patients, all of whom received only left-sided lesions, have atrial flutter postoperatively? I would like to close by again congratulating the authors on a well-conducted and elegantly presented study that I believe is an important contribution to the expanding body of literature in this exciting new field of surgical interest, and by thanking the Society for the privilege of the floor. DR PASIC: Thank you very much, Dr Argenziano, for your comments and your questions. They are important issues regarding surgery for atrial fibrillation and the maze procedure. I would like to emphasize that there is no clear definition of giant left atrium. This definition is mostly established according to the echocardiographically measured diameter of the left atrium. However, I believe that an echocardiograhically measured diameter, although mostly used, is a confusing parameter and probably should not be used for the definition. The volume of the left atrium is the better parameter for the definition of giant left atrium. An enlarged left atrium has three diameters and usually we receive only one measured by echocardiography. There are three distances that should be measured: (1) between the right and left pulmonary veins, (2) between the mitral valve and the roof of the atrium, and (3) the height of the left atrium. Therefore, I suggest that all three distances should be measured in order to assess the volume of the atrium. A volume of 250 ml presents a giant left atrium. At our institution, giant left atrium was considered a contraindication for the surgically performed Cox-maze III procedure. However, the radiofrequency maze procedure is an easy method that can be carried out in a short time and, at present, we do not consider giant left atrium as a contraindication for radiofrequency maze ablation. We are aware that it is possible to obtain good results in these patients regarding electrical function of the left atrium and conversion into sinus rhythm, but the postoperative mechanical function of the left atrium is not good because of extensive histological changes in the wall of the dilated atrium. Your second question was about our policy not to exclude or excise the left atrial appendage on a routine basis. We do not routinely exclude the left atrial appendage in patients with atrial fibrillation who do not undergo the maze procedure. Therefore, we also do not do so in patients undergoing radiofrequency maze ablation. There are several reasons not to excise the appendage. First, it simplifies the surgical procedure significantly, eliminating the need for cutting and oversewing of tissue in this region, which is sometimes very fragile and has the circumflex artery in the vicinity. Furthermore, in patients with sinus rhythm the contribution of left atrial appendage is important for mechanical function of the left atrium. We would exclude the atrial appendage if chronic thrombotic material is found within the appendage. In this case it should be oversewn in order to prevent possible later embolization. However, in the reported group of patients we have seen two patients with left atrial thrombi, but the atrial appendage was always free from coagula and therefore, the appendage was not oversewn. Your further comment was about our technique of separate isolation of the pulmonary veins using two separate isolation lines, one for the left and one for the right veins, instead of one encircling line as in the classic maze procedure. There are two reasons why we do this: (1) the simplicity of the technical procedure, because we use a long and flexible radiofrequency probe that enables us to perform one isolation line in 2 minutes by a single application of the probe, and (2) we believe that the posterior wall should not be excluded from the mechanical activity of the left atrium. Using our technique, the posterior atrial wall is not electrically isolated and it may contribute to mechanical atrial function but, if the standard maze technique is applied, this part of the atrial wall is excluded from the mechanical activity of the left atrium. One patient died of stroke after mechanical valve replacement. At the time, the patient had sinus rhythm and was under adequate anticoagulation. Furthermore, there was no evidence of atrial thrombosis using transesophageal echocardiography. We can not rule out that the source of the embolus could have been from the left atrial appendage or from the maze lines, but also it was not possible to exclude other sources of embolization, such as from the ascending aorta or aortic arch or carotid artery. Our policy to perform only the left atrial maze procedure is based on our goal to perform an easy and short procedure. The bi-atrial procedure would certainly increase the success rate of elimination of atrial fibrillation, but it would also prolong the operating time in patients who do not need a right atrial procedure. Except in some patients in the early postoperative course, the occurrence of atrial flutter was not recorded in patients after discharge from hospital. Thank you.