Successful Performance of Cox-Maze Procedure on Beating Heart Using Bipolar Radiofrequency Ablation: A Feasibility Study in Animals

Transcription

1 Successful Performance of Cox-Maze Procedure on Beating Heart Using Bipolar Radiofrequency Ablation: A Feasibility Study in Animals Sydney L. Gaynor, MD, Yosuke Ishii, MD, Michael D. Diodato, MD, Sunil M. Prasad, MD, Kara M. Barnett, BS, Nicholas R. Damiano, Gregory D. Byrd, BS, Samuel A. Wickline, MD, Richard B. Schuessler, PhD, and Ralph J. Damiano, Jr, MD Division of Cardiothoracic Surgery, Department of Surgery, and Division of Cardiology, Department of Internal Medicine, Washington University School of Medicine, St. Louis, Missouri Background. The Cox-Maze procedure is the gold standard for the surgical treatment of atrial fibrillation with proven long-term efficacy. However, its application has been limited by its complexity and significant morbidity. The purpose of this study was to test the feasibility and safety of performing the Cox-Maze procedure using bipolar radiofrequency ablation on the beating heart without cardiopulmonary bypass. Methods. After median sternotomy, 6 Hanford minipigs underwent a modified Cox-Maze procedure using bipolar radiofrequency energy. The animals survived for 30 days. Atrial function, coronary artery, pulmonary vein anatomy, and valve function were assessed by magnetic resonance imaging. At reoperation, pacing documented electrical isolation of the pulmonary veins. Induction of atrial fibrillation was attempted by burst pacing with cholinergic stimulation. Histologic assessment was performed after sacrifice. Results. There were no perioperative mortalities or neurologic events. At 30 days, atrial fibrillation was unable to be induced, and pulmonary vein isolation was confirmed by pacing. Magnetic resonance imaging assessment revealed no coronary artery or pulmonary vein stenoses. Although atrial ejection fraction decreased slightly from to (p 0.18), atrial contractility was preserved in every animal. Histologic assessment showed all lesions to be transmural, and there were no significant stenoses of the coronary vessels or injuries to the valves. Conclusions. Virtually all of the lesions of the Cox- Maze procedure can be performed without cardiopulmonary bypass using bipolar radiofrequency energy. There were no late stenoses of the pulmonary veins. Clinical trials of this new technology on the beating heart are warranted. (Ann Thorac Surg 2004;78:1671 7) 2004 by The Society of Thoracic Surgeons The Cox-Maze III procedure is the most successful surgical treatment for atrial fibrillation (AF), with a long-term cure rate in excess of 90% [1 5]. However, this operation has not been widely adopted because of its invasiveness and complexity. There has been much recent interest in simplifying this procedure by using various energy sources to create linear lines of ablation to replace the multiple surgical incisions of the traditional cut-and-sew procedure [6]. Most of these new energy sources use unipolar catheters. Because these devices do not possess any means of determining the transmurality of the lesions, it often has been necessary to use prolonged ablation times and multiple lesions. These prolonged application times have resulted in injuries to collateral structures [7, 8] while still not guaranteeing transmural lesions [9]. Our laboratory has chosen to investigate bipolar radiofrequency (RF) energy [10 13]. With a bipolar device, Accepted for publication April 20, Address reprint requests to Dr Damiano, Division of Cardiothoracic Surgery, Washington University School of Medicine, 660 S Euclid Ave, Box 8234, St. Louis, MO 63110; damianor@msnotes.wustl.edu. alternating current is delivered between two closely approximated electrodes embedded into the jaws of a clamp [11]. This focused energy delivery minimizes lesion width and reduces the possibility of collateral damage [14]. An advantage of this technology is the ability to have a real-time indicator of lesion transmurality by measuring tissue conductance between the two electrodes [12, 13]. In previous work, our laboratory has shown that bipolar RF energy creates discrete, transmural atrial lesions within 10 seconds. Approximately 1 mm from the electrode edge, tissue temperature falls below that required for irreversible cellular injury. If bipolar RF is to be used to replace the surgical incisions of the Cox-Maze procedure, it would have to cross tricuspid and mitral valve tissue and the coronary sinus, and be used in close proximity to the right and circumflex coronary arteries. The purpose of this chronic animal study was to examine the feasibility of using bipolar RF to perform the Cox-Maze procedure on the beating heart Dr R. Damiano discloses that he has a financial relationship with AtriCure, Inc by The Society of Thoracic Surgeons /04/$30.00 Published by Elsevier Inc doi: /j.athoracsur

2 1672 GAYNOR ET AL Ann Thorac Surg BEATING HEART MAZE PROCEDURE WITH BIPOLAR RF 2004;78: Fig 1. Cox-Maze lesion set. (IVC inferior vena cava; Lt. left; RF radiofrequency; Rt. right; SVC superior vena cava.) and to examine the effect of this energy on heart valvular tissue, coronary arteries and veins, and pulmonary veins (PVs) at 1 month. Material and Methods Experimental Protocol Six Hanford mini-pigs weighing 40 to 50 kg were used in this study. All animals received humane care in compliance with the Guide for the Care and Use of Laboratory Animals, published by the National Institutes of Health (National Institutes of Health publication 85-23, revised 1985). Preoperatively, each animal was anesthetized and intubated and had a magnetic resonance imaging (MRI) study to assess (1) global cardiac function, (2) the mitral and tricuspid valves, (3) PV anatomy and flow, and (4) coronary artery anatomy and flow. After the initial MRI, the animals underwent a modified Cox-Maze procedure on the beating heart. The animals were premedicated, intubated, and anesthetized with 2% to 4% isoflurane and monitored continuously during the procedure. The heart was exposed through a median sternotomy. The left and right PVs and the inferior vena cava were isolated with umbilical tapes. The heart was paced from the right atrial appendage, the right atrial body, and the left and right PVs to establish pacing thresholds. The animals were given intravenous heparin (350 U/kg) to maintain an activated clotting time greater than 250 seconds. Intravenous bretylium was administered to prevent arrhythmias with manipulation of the heart. A bipolar RF ablation device (AtriCure, Inc, Cincinnati, OH) was used to create the lesion set. The device consists of a handpiece with embedded bipolar electrodes (5 cm long, 1 mm wide) and a generator. The RF energy was delivered at 75 V and 750 ma. Ablations were terminated when the tissue conductance decreased and achieved a steady state for 2 seconds [12, 13]. A laptop computer with Labview version 5.1 (National Instruments, Austin, TX) was used to monitor and record temperature, time, current, voltage, impedance, conductance, and energy in real time. The temperature was recorded by a thermistor 1 mm from the edge of the electrode in the jaw of the device. Surgical Procedure A modified Cox-Maze procedure was performed, which included virtually all of the right and left atrial lesions of the traditional procedure (Fig 1). All lesions were performed with the bipolar RF device. Because the procedure was performed without cardiopulmonary bypass, it was necessary to make the following modifications of a classic Cox-Maze III procedure: (1) the right and left PVs were isolated separately rather than as one large island, (2) a connecting lesion was performed between the right and left PVs, and (3) the transseptal incision, which functions principally to aid in exposure of the posterior left atrium, was omitted. The heart was manipulated with the aid of a suction retractor (AXIUS Xpose 3; Guidant Corp, Santa Clara, CA). The device was passed around the right and then left PVs, and a cuff of surrounding atrial tissue was clamped and ablated on each side. Electrical isolation was confirmed by pacing the PVs distal to the lesions at a stimulus strength of 20 ma. The bipolar RF device was introduced through a pursestring suture in the anterior aspect of the left atrial appendage with the tip directed into the left superior PV to create the connecting lesion from the left atrial appendage to the left PV. The bipolar RF device was manipulated so that its inner jaw traversed the mitral valve orifice into the left ventricle and the outer jaw crossed the atrioventricular groove with the tips resting on the endocardial and epicardial surfaces of the left ventricle. The distal circumflex artery was incorporated purposefully into this ablation. The left atrial appendage was then circumferentially oversewn with a running suture along the entirety of the ablation line. The heart was repositioned to expose the intraatrial groove, and a pursestring suture was placed in the left atrial wall at the site of the circumferential ablation of the

3 Ann Thorac Surg GAYNOR ET AL 2004;78: BEATING HEART MAZE PROCEDURE WITH BIPOLAR RF 1673 right PVs. The device was introduced into the left atrium and manipulated such that its inner jaw was introduced into the left inferior PV while its outer jaw was manipulated below the inferior vena cava and across the posterior wall of the left atrium with its tip on the epicardial surface of the left inferior PV. The interposing tissue was then ablated, completing the vein-connecting lesion between the left and right PVs. The bipolar RF device was introduced through a pursestring suture at the tip of the right atrial appendage and positioned to create the right atrial free wall lesion (Fig 2A). The device was then placed across the tricuspid valve crossing the atrioventricular groove, and all included tissue was ablated. The right atrial appendage was circumferentially ablated. A pursestring suture was placed midway between the confluence of the superior vena cava and inferior vena cava. The bipolar RF device was introduced through this pursestring suture, and superior vena cava and inferior vena cava lesions were created (Fig 2B). The snare was relaxed, and the device was carefully repositioned into the right ventricle. This lesion incorporated the tricuspid valve, atrioventricular groove, and right coronary artery. The bipolar RF device was removed and hemostasis secured. This completed the modified Cox-Maze lesion set (Fig 1). The pericardium was closed and the sternum reapproximated. The entire surgical procedure was completed in less than 1 hour. Postoperative Data Collection Postoperatively the animals were closely monitored for 48 hours, and survived for 30 days. Each animal received 62.5 mg of aspirin a day, commencing on the first postoperative day. No antiarrhythmic drugs were used during this period. At 1 month, each animal underwent an MRI scan to reassess the cardiac measurements, followed immediately by a redo median sternotomy. Pacing at 20 ma (5-millisecond pulse duration) was used to document chronic electrical isolation of the left and right PVs. Neostigmine (2 3 mg/50 kg) was administered intravenously. Burst pacing was performed from the right atrial appendage and left and right atrial bodies at a cycle length of 30 to 100 milliseconds continuously for 30 seconds in an attempt to induce AF [15, 16]. Fig 2. (Top) Right atrial free wall lesion. (Bottom) Superior vena cava lesion. Histology and Microscopic Analysis The animals were sacrificed, and the hearts were removed en bloc and examined grossly for any evidence of intraatrial thrombus formation, PV stenosis, or thrombosis. The hearts were immediately placed in 1% 2,3,5- triphenyl-tetrazolium chloride solution and incubated at room temperature for 45 minutes. Each bipolar RF lesion was sectioned at four levels, 5.0 mm apart, perpendicular to the long axis of the ablation line. In the regions where the lesions crossed the right and circumflex coronary arteries, sections were taken axially 5.0 mm apart inclusive of the lesion. The sections were then fixed in formalin, molded in paraffin, sectioned, and stained with Gomori s trichrome, hematoxylin and eosin, and Masson s trichrome stains. The sections were microscopically examined to assess transmurality of the lesions and the effects of bipolar RF energy on the coronary arteries, coronary sinus, and mitral and tricuspid valves. Magnetic Resonance Imaging Protocol The animals were anesthetized and placed supine in a clinical 1.5-T MRI scanner (NT Intera CV; Philips Medical Systems, Best, The Netherlands) and fitted with a fiveelement dedicated cardiac surface coil for image reception. The imaging protocol consisted of multiple anatomic and functional cine images and velocity-encoded images. All images were synchronized to the cardiac

4 1674 GAYNOR ET AL Ann Thorac Surg BEATING HEART MAZE PROCEDURE WITH BIPOLAR RF 2004;78: rhythm using a four-lead vectorcardiogram system. The cine images, triggered to each R-wave peak, were based on a steady-state gradient echographic technique (BFFE) with 15 phases per cardiac cycle spaced 20 to 25 milliseconds apart. The cine MRI views consisted of (1) horizontal long axis, (2) vertical long axis, (3) short axis (20 planes, 9 mm thick), and (4) left ventricular outflow track. The reconstructed pixel size was approximately 1.5 mm 1.2 mm. Transmitral and transtricuspid blood flow were assessed using velocity-encoded MRI [17]. Magnetic Resonance Imaging Data Analysis Analysis of the acquired images was performed offline using a commercial software package (Easy Vision, Philips Medical System, Best, The Netherlands). The volume of the left atrium was assessed by planimetry. The area of the left atrial blood pool was identified in each short-axis slice. The image was hand-traced using a cursor, and the volumes were calculated and summed over all slices to give the total volume of the left atrium at end systole and end diastole. The left atrial volume was determined for these two phases. Atrial ejection fraction was then defined as [(end-diastolic volume end-systolic volume) divided by end-diastolic volume] [13]. Analysis of quantitative flow across the tricuspid and mitral valves was performed using velocity-encoded images. Magnetic resonance imaging was used to assess anatomic and physiologic flow patterns in the pulmonary venous system. Coronary Artery Anatomy and Patency The technique of Botnar and colleagues [18], which uses a navigator-gated free-breathing T 2 -weighted prepared turbo field echo acquisition, was used to assess coronary artery anatomy and patency. Results Operative Results All animals survived the initial operative procedure. There was no evidence of cardiac arrhythmias, and all animals remained in sinus rhythm during the operation and at the time of sacrifice. There were no signs of gross neurologic dysfunction in any of the animals after the procedure. Fig 3. Representative magnetic resonance image demonstrating right pulmonary vein patency (arrow) at 1 month after ablation with bipolar radiofrequency device. Radiofrequency Lesions The beating heart lesion set consisted of 12 bipolar RF ablations (Fig 1). Two lesions traversed the tricuspid valve, and one crossed the mitral valve. The coronary sinus and the right and circumflex coronary arteries were deliberately crossed with the bipolar RF device. Acutely, the PVs were electrically isolated with a single ablation in all instances. This was documented by failure of the heart to capture while pacing from the PVs at 20 ma. All lesions were created with a single application of the bipolar RF device. The average ablation time for the left PVs was seconds and for the right PVs was seconds. The average ablation time for all lesions was seconds. The average tissue temperature, as measured 1 mm from the electrodes, was C. The average energy requirement per lesion was J. Cardiac Magnetic Resonance Imaging Data Comparative analysis was made between preoperative and postoperative MRI data in 5 animals. In 1 animal, no preoperative MRI data were available. There was no evidence of PV stenosis in any animal (Fig 3). Quantitative flow study revealed no significant evidence of mitral or tricuspid regurgitation in either the preoperative or postoperative MRI. The regurgitant fraction was less than 5% in every animal. Atrial ejection fraction was assessed by MRI. Preoperatively, the average atrial ejection fraction was Postoperatively, atrial ejection fraction fell to (p 0.18; Table 1). Atrial contractility was Table 1. Ejection Fraction at Baseline and at 1 Month Pig No. Preop EF Postop EF Average a a p value EF ejection fraction; Postop postoperative; Preop preoperative.

5 Ann Thorac Surg GAYNOR ET AL 2004;78: BEATING HEART MAZE PROCEDURE WITH BIPOLAR RF 1675 Fig 5. Photomicrograph of ablated circumflex coronary artery 1 month after ablation with bipolar radiofrequency device. (Masson s trichrome stain, 20 magnification.) Fig 4. Magnetic resonance image of right coronary artery (arrow) at 1 month after ablation with bipolar radiofrequency device. preserved in every instance, as confirmed by the presence of the peak velocity of atrial filling, the a wave. Magnetic resonance imaging evaluation did not detect any coronary artery stenosis or thrombosis in the areas ablated. The right coronary and circumflex coronary arteries remained patent in each animal (Fig 4). Chronic Pacing At 30 days, pacing demonstrated that the right atrial appendage and the right and left PVs were isolated from the body of the atria in every animal. During normal sinus rhythm or atrial pacing from the body of the atria, the right atrial appendage was electrically silent. Attempts to induce AF from both left and right atria with burst pacing coupled with intravenous neostigmine were unsuccessful in all cases. Neostigmine was used because it increases cholinergic tone and shortens the effective refractory period, thus facilitating reentry [15, 16]. In previous work from our laboratory, this has been shown to induce AF in 100% of normal animals [15]. Histology On gross inspection of the heart, there was no evidence of intraatrial thrombus formation or stricture of the PVs. At necropsy, the hearts were placed in 1% 2,3,5- triphenyl-tetrazolium chloride solution and examined. The atrial lesions appeared pale, discrete, linear, and approximately 1 to 2 mm in width. Four representative sections were taken from each of the 12 ablative lesions (Fig 1). A total of 288 samples were examined from the 6 animals. Transverse and linear histologic sections of ablative tissue demonstrated continuous full-thickness replacement of myocardial fibers with septae of maturing fibrous connective tissue (collagen). Microscopic examination of the atrial samples stained with hematoxylin and eosin, Gomori s trichrome, and Masson s trichrome revealed that the atrial lesions were all transmural (288 of 288 samples). There was no evidence of stenosis or thrombosis of the circumflex and right coronary arteries or coronary sinus. The vessels remained patent as demonstrated by histology (Figs 5, 6). Inflammatory cell infiltrates, associated with ablated tissue, were minimal to mild within developing connective tissue septae. Histologic examination of the ablated tricuspid and mitral valve tissue revealed no apparent injury or perforation of the leaflets, and the Fig 6. Photomicrograph of ablated coronary sinus 1 month after ablation by bipolar radiofrequency device. (Gomori s trichrome, 20 magnification.)

6 1676 GAYNOR ET AL Ann Thorac Surg BEATING HEART MAZE PROCEDURE WITH BIPOLAR RF 2004;78: Fig 7. Radiofrequency lesion crossing the posterior mitral valve leaflet. subvalvular apparatus remained intact (Fig 7). There was no valvular thrombus. Comment Our laboratory and others have been examining less invasive approaches to the surgical treatment of AF [11, 12]. The goal of these approaches is to preserve the efficacy of the Cox-Maze procedure and, at the same time, decrease the morbidity of this procedure. In our opinion, the most invasive aspect of the Cox-Maze procedure is the need for a prolonged period of cardiopulmonary bypass and cardiac arrest. Thus, our efforts have been to develop a beating-heart approach to perform this operation that would not require cardiopulmonary bypass. To facilitate the procedure, we elected to use bipolar RF ablation to replace the surgical incisions. The use of ablation lines in lieu of surgical incisions has several advantages. First, clamping can be performed without cardiopulmonary bypass, and does not require an arrested heart as does the traditional cut-and-sew approach. It is faster, and thus decreases the time to perform the lesions. The lack of long suture lines may reduce the risk of postoperative bleeding. Finally, the use of a clamp as opposed to a complicated set of incisions simplifies the procedure and, hopefully, makes it accessible to all surgeons. In this study, our goal was to examine the feasibility of a beating-heart Cox-Maze procedure. In the traditional procedure, both left and right atrial appendages were resected. However, there has been increasing evidence that preservation of the atrial appendages preserves the production of atrial natriuretic peptide, thus reducing the incidence of postoperative fluid retention [19, 20]. The left and right atrial appendages were ablated in this experiment to document electrical isolation and to generate all possible lesions historically associated with the Cox- Maze III procedure. The septal lesion was deliberately eliminated in this study because traditionally this lesion was performed only for exposure. It is possible that reentry could develop around the fossa ovalis; however, this has not ever been seen during clinical mapping of AF at our institution. Our results have demonstrated that bipolar RF energy can be used on the beating heart to create transmural lesions that replace the standard lesions of the Cox-Maze III procedure. These lesions resulted in chronic transmural conduction block documented by pacing, and they were transmural on histologic examination. Atrial fibrillation could not be induced in any animal with cholinergic stimulation. To evaluate the safety of this approach, the animals underwent both preoperative and postoperative MRI evaluation to assess cardiac anatomy and atrial function. Cardiac MRI was chosen instead of echocardiography because of its superior edge detection and the ability to delineate structures as small as 1 to 2 mm [21, 22]. There was no evidence of PV stenosis with bipolar RF ablation. This is in contrast to the late PV stenosis seen with catheter-based, unipolar RF ablation [23]. Although atrial ejection fraction was slightly decreased after this procedure, all of the animals had atrial contractility, as documented by the presence of an a wave. The chronic effects of bipolar RF ablation on the mitral valve, tricuspid valve, coronary sinus, and the right and circumflex coronary arteries were evaluated. Surprisingly, both MRI angiography and chronic histology documented no evidence of coronary artery or coronary sinus stenosis or thrombosis after the procedure. The valvular function remained intact as documented by MRI in vivo and histology after sacrifice. In summary, this study demonstrated the feasibility of a beating-heart operation that included all of the standard lesions of the Cox-Maze procedure. Bipolar RF ablation was effective in creating transmural lesions, both by electrophysiologic and histologic evaluation. This energy source appears to be safe and did not result in injury to the PVs, coronary arteries, or heart valves. Initial clinical trials of this technology on the beating heart are warranted, and may lead to a more minimally invasive Cox-Maze procedure. Study Limitations This study has significant limitations, and the results must be cautiously applied to the clinical situation. First,

7 Ann Thorac Surg GAYNOR ET AL 2004;78: BEATING HEART MAZE PROCEDURE WITH BIPOLAR RF there are significant differences between human anatomy and that of the porcine model. There is minimal fat around the PVs, and there is significantly less tissue in the porcine model. In this study, complete isolation of the PVs required only one early application. In our initial clinical experience, complete isolation of the PVs often required two or more applications. Moreover, this study was performed on normal porcine atria, and not on the diseased and often thickened human atria. However, our initial clinical experience on cardiopulmonary bypass has similarly documented the efficacy of this technology [24]. This experimental procedure involved using bipolar RF ablation across both the mitral and tricuspid valve and the circumflex and right coronary arteries. Although there was no evidence of thrombosis, stenosis, or perforation found by MRI or histology, the follow-up was at only 1 month and does not rule out the possibility of late injury, particularly to the coronary arteries [25]. It is our belief that this energy should not be used across coronary arteries. This procedure can easily be modified in the clinical situation to avoid these structures by reflecting the atrioventricular groove fat pad, or crossing the groove distal to the last circumflex marginal coronary artery. Finally, a significant shortcoming to the clinical use of this procedure is the need to place the clamp inside the right and left atria. Although this likely can be performed with little difficulty on the right side of the heart, there would be a significant chance for both bleeding and air embolus on the left side. Thus, this could limit the clinical utility of this procedure and may require the development of other parallel technology. Supported in part by National Institutes of Health grants R44 HL67535, 5R01 HL References 1. Cox JL, Sundt TM III. The surgical management of atrial fibrillation. Ann Rev Med 1997;48: 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: Prasad SM, Maniar HS, Camillo CJ, et al. The Cox-Maze procedure for atrial fibrillation. Long-term efficacy in patients undergoing lone versus concomitant procedures. J Thorac Cardiovasc Surg 2003;126: Cox JL, Canavan TE, Schuessler RB. The surgical treatment of atrial fibrillation: intraoperative electrophysiologic mapping and description of the electrophysiologic basis of atrial flutter and atrial fibrillation. J Thorac Cardiovasc Surg 1991; 101: Cox JL, Schuessler RB, D Agostino HJ Jr, Stone CM, Chang BC, Corr PB. The surgical treatment of atrial fibrillation: development of a definitive surgical procedure. J Thorac Cardiovasc Surg 1991;101: Viola N, Williams MR, Oz MC, Ad N. The technology in use for the surgical ablation of atrial fibrillation. Semin Thorac Cardiovasc Surg 2002;14: Gillinov AM, Petterson G, Rice TW. Esophageal injury during radiofrequency ablation for atrial fibrillation. J Thorac Cardiovasc Surg 2001;122: Doll N, Borger MA, Fabricius A, et al. Esophageal perforation during left atrial radiofrequency ablation: is the risk too high? J Thorac Cardiovasc Surg 2003;125: Thomas SP, Guy DJR, Boyd AC, et al. Comparison of epicardial and endocardial linear ablation using handheld probes. Ann Thorac Surg 2003;75: Damiano RJ Jr. Alternative energy sources for atrial ablation: judging the new technology. Ann Thorac Surg 2003;75: Prasad SM, Maniar HS, Moustakidis P, Schuessler RB, Damiano RJ Jr. Epicardial ablation on the beating heart: progress towards an off-pump maze procedure. Heart Surg Forum 2001;5: Prasad SM, Maniar HS, Schuessler RB, Damiano RJ Jr. Chronic transmural atrial ablation by using bipolar radiofrequency energy on the beating heart. J Thorac Cardiovasc Surg 2002;124: Prasad SM, Maniar HS, Diodato MD, Schuessler RB, Damiano RJ Jr. Physiological consequences of bipolar radiofrequency energy on the atria and pulmonary veins: a chronic animal study. Ann Thorac Surg 2003;76: Fayad G, Modine T, Le Tourneau T, et al. Circumflex artery stenosis induced by intraoperative radiofrequency ablation. Ann Thorac Surg 2003;76: Schuessler RB, Grayson TM, Bromberg BI, Cox JL, Boineau JP. Cholinergically mediated tachyarrhythmias induced by a single extrastimulus in the isolated canine right atrium. Circ Res 1992;71: Geddes LA, Hinds M, Babbs CF, et al. Maintenance of atrial fibrillation in anesthetized and unanesthetized sheep using cholinergic drive. Pacing Clin Electrophysiol 1996;19: Naylor GL, Firimin DN, Longmore DB. Blood flow imaging by cine magnetic resonance. J Comput Assist Tomogr 1986; 10: Botnar RM, Stubewr M, Danias PG, Kissinger KV, Manning WJ. Improved coronary artery definition with T 2 -weighted, free-breathing, three-dimensional coronary MRA. Circulation 1999;99: Isobe F, Kumano H, Ishikawa T, et al. A new procedure for chronic atrial fibrillation: bilateral appendage-preserving maze procedure. Ann Thorac Surg 2001;72: Kondo N, Takahashi K, Minakawa M, Daitoku K. Left atrial maze procedure: a useful addition to other corrective operations. Ann Thorac Surg 2003;75: Plein S, Bloomer TN, Ridgway JP, Jones TR, Bainbridge GJ, Sivananthan MU. Steady-state free precession magnetic resonance imaging of the heart: comparison with segmented K-space gradient-echo imaging. J Magn Reson Imaging 2001;14: Grothues F, Smith GC, Moon JCC, et al. Comparison of interstudy reproducibility of cardiovascular magnetic resonance with two-dimensional echocardiography in normal subjects and in patients with heart failure or left ventricular hypertrophy. Am J Cardiol 2002;90: Arentz T, Jander N, von Rosenthal J, et al. Incidence of pulmonary vein stenosis 2 years after radiofrequency catheter ablation of refractory atrial fibrillation. Eur Heart J 2003;24: Damiano RJ Jr, McCarthy PM, Gillinov AM, et al. Prospective multicenter clinical trial of surgical pulmonary vein isolation for the treatment of atrial fibrillation [Abstract]. J Am Coll Cardiol 2003;41(Suppl A):498A. 25. Menasse E, Medici D, Ghiselli S, Ornaghi D, Gallotti R. Left main coronary arterial lesion after microwave epicardial ablation. Ann Thorac Surg 2003;76:276 7.