During the past several years, surgical intervention

Transcription

1 Radial Approach: A New Concept in Surgical Treatment for Atrial Fibrillation I. Concept, Anatomic and Physiologic Bases and Development of a Procedure Takashi Nitta, MD, Richard Lee, MD, Richard B. Schuessler, PhD, John P. Boineau, MD, and James L. Cox, MD Divisions of Cardiothoracic Surgery and Cardiology, Washington University School of Medicine, St. Louis, Missouri Background. The maze procedure cures atrial fibrillation; however, it isolates the pulmonary vein area and results in discordant activation in certain adjacent left atrial segments, which affects left atrial function. To preserve a more physiologic atrial transport function, we developed a new concept of surgical treatment for atrial fibrillation the radial approach. The atrial incisions radiate from the sinus node toward the atrioventricular annular margins to allow a more physiologic atrial activation sequence and parallel the atrial coronary arteries to preserve blood supply to most atrial segments. Methods. We examined the atrial coronary arteries and the activation sequence during sinus rhythm in normal canine hearts to design the atrial incisions according to the concept of a radial approach. Results. The pattern of coronary artery distribution was centripetal, branching from the right coronary or left circumflex coronary artery at the right or left atrioventricular groove and spreading toward the sinus node. The endocardial mapping of the atria disclosed some important findings in designing the atrial incisions of the radial approach: the activation sequence at the left atrial septum and at the posterior left atrium between the pulmonary vein orifices. The atrial incisions were designed according to these findings. Conclusions. The radial approach may represent a more physiologic atrial transport function. (Ann Thorac Surg 1999;67:27 35) 1999 by The Society of Thoracic Surgeons Presented at the Thirty-fourth Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Jan 26 28, Address reprint requests to Dr Boineau, Division of Cardiothoracic Surgery, Washington University School of Medicine 660 S Euclid Ave, Box CSRB, St. Louis, MO During the past several years, surgical intervention has emerged as an important modality in the treatment of patients with atrial fibrillation (AF) [1 4]. The drawbacks of AF include uncomfortable symptoms, such as palpitation and an irregular heartbeat, risk of thromboembolism, and impaired hemodynamic function from loss of atrial contraction. Because pharmacologic treatment aimed at converting or preventing AF is less than optimal, medical treatment is ultimately directed at decreasing the ventricular response rate and anticoagulation. However, despite these treatments, the atria still fibrillate; thus, hemodynamic function remains impaired. Therefore, surgical intervention must overcome all these disadvantages and prevent any consequences of AF. The maze procedure is a surgical intervention to restore sinus rhythm and atrial contraction in patients with AF. The atrial incisions of the procedure block all the potential macroreentrant pathways and narrow the atrial tissue to block propagation of the microreentrant wavelets. As a result, the procedure restores sinus rhythm and some degree of atrial mechanical function. A high incidence of restoration of sinus rhythm and prevention of AF after operation has been reported in patients with and without structural heart disease [2 6]. Among 164 patients who underwent the maze procedure at our institution [4], AF or atrial flutter recurred in 12 (7%) postoperatively. All 12 patients were converted to sinus rhythm with medical therapy. Thus, all patients after the maze procedure are now free of AF. Atrial transport function was also studied in these patients. One hundred twentyfive patients were reevaluated approximately 6 months postoperatively specifically for atrial transport function. The evaluation revealed that 98% of patients had right atrial and 86% left atrial transport function detected by Doppler echocardiography, magnetic resonance imaging, or atrioventricular versus ventricular pacing at the same rate. Although the presence of atrial mechanical contraction has been demonstrated in most patients after the maze procedure, more important is whether the amount of atrial mechanical function is enough to provide sufficient atrial transport function and to eliminate the risk of systemic thromboembolism. Quantitative analysis of atrial function in patients after the maze procedure revealed that the left atrial transport function was significantly less than that in normal control subjects, whereas the right atrial function was comparable [6 9]. The reason for this finding is hypothesized as follows: The maze 1999 by The Society of Thoracic Surgeons /99/$20.00 Published by Elsevier Science Inc PII S (98)

2 28 NITTA ET AL Ann Thorac Surg RADIAL APPROACH FOR AF: PART ;67:27 35 procedure electrically and thus mechanically isolates the posterior left atrium between the pulmonary vein orifices, thereby eliminating approximately 29% of the left atrium [10] from contributing to the transport function. Because of the complex and detouring incisions, the activation and contraction of neighboring atrial segments across the incisions can be discordant. Also, the total atrial activation time is prolonged; thus, atrioventricular contraction coupling can be desynchronized. Moreover, some of the maze incisions interrupt the atrial coronary arteries; therefore myocardial circulation can be impaired by the procedure. The significance of atrial contraction to ventricular filling depends on the patient s ventricular diastolic function. In ventricles with normal compliance, relatively low atrial pressure is necessary to fill the ventricle and maintain normal cardiac output. This preceeding statement is true because the early rapid filling is accomplished by the pressure gradient between the atrium and ventricle, where the diastolic pressure is normal. Most of the ventricular filling is completed during the early diastolic period, and there is little need of atrial contraction (late filling), which is why patients with AF and normal ventricular systolic and diastolic function do not exhibit the symptoms of heart failure. In patients with diastolic dysfunction, however, the ventricular filling pressure is elevated because of the impaired compliance. Even if the left atrial pressure is raised and filling time is prolonged to compensate for the ventricular filling, diastolic ventricular filling is not accomplished during the early filling period. Atrial contraction during the late filling period contributes to ventricular filling and results in maintaining cardiac output within the normal range. Therefore, atrial contraction is an important determinant of hemodynamic status in patients with diastolic dysfunction. Abnormalities of diastolic ventricular filling are seen in patients with most forms of heart disease, including hypertension [11], coronary artery disease [12], valvular heart disease [13], cardiomyopathies [14], and a variety of systemic diseases. Diastolic dysfunction can be present and can be detected before the clinical manifestations of disease. More important, aging is another important factor that affects diastolic function [15]. In young adults with normal diastolic function, 85% to 95% of ventricular filling occurs in early diastole. By age 65, however, approximately 50% of flow may occur during late diastole because of the impaired diastolic function. Therefore the majority of the patients undergoing operation for AF can have some degree of ventricular diastolic dysfunction. We hypothesized that a procedure with atrial incisions radiating from the sinus node toward the atrioventricular annular margins, paralleling the activation sequence and paralleling the atrial coronary arteries, would provide a more physiologic atrial activation contraction sequence and optimize the atrial contribution to ventricular filling. The concepts of the maze procedure and the radial approach are compared in Figure 1. It is hypothesized that the activation wavefront from the sinoatrial node radiates centrifugally toward the atrioventricular annular Fig 1. Schema of the concepts of the maze procedure and the radial approach. The large outer circle denotes the atria, and its outer limit is bounded by the atrioventricular annular margins. The small circle indicates the sinoatrial node, and the shaded area indicates the isolated portion of the atrium. Arrows indicate the activation wavefront from the sinoatrial node, radiating toward the annular margins. The atrial coronary arteries, arising at the atrioventricular groove, are also schematically drawn. Note that the radial approach (right panel) preserves a more physiologic activation sequence and preserves blood supply to most atrial segments, whereas the atrial incisions of the maze procedure (left panel) desynchronize the activation sequence, and some of the incisions cross the atrial coronary arteries. margins in normal atria and that the atrial coronary arteries, originating at the atrioventricular groove, distribute centripetally toward the sinus node. In the left panel, the atrial incisions of the maze procedure desynchronize this physiologic activation sequence and prolong the total activation time of the atria. Some of the incisions cross the atrial coronary arteries. Also, there is a large electrically isolated region encompassing the pulmonary vein orifices. In the right panel, the atrial incisions of the radial approach parallel the direction of the physiologic activation and the direction of the blood supply. There is no electrically or mechanically isolated region. These incisions may not disturb the physiologic activation sequence and may preserve blood supply to most atrial segments. To design atrial incisions that follow the concept of the radial approach, atrial anatomy and electrophysiologic variables should be considered with particular regard to the coronary artery distribution and the activation sequence during normal sinus rhythm. The purpose of the present study was to examine the atrial coronary arteries and activation sequence in normal canine hearts and to design atrial incisions that conform to the concept of the radial approach. A subsequent report by our group describes the testing of the procedure in animals with respect to electrophysiologic and hemodynamic variables [16]. Material and Methods Atrial Coronary Arteries The anatomy of the atrial coronary arteries was studied in 26 adult mongrel dogs of either sex, weighing 20 to 30 kg. The heart was exposed through a midline thoracotomy. The heart was excised with the inferior vena

3 Ann Thorac Surg NITTA ET AL 1999;67:27 35 RADIAL APPROACH FOR AF: PART 1 29 cava, the superior vena cava, and all pulmonary veins attached. The heart was flushed with 500 ml of heparinized saline through a plastic cannula inserted into the coronary arteries, which flushed all blood from them. Care was taken not to inject air into the coronary arteries. The major ventricular branches of the coronary arteries were then ligated with silk stitches. Colored and diluted silicone rubber injection compound (Microfil; Canton Bio-medical Products, Inc, Boulder, CO) was injected into each coronary ostium through a 16-gauge plastic cannula. Orange compound was injected into the right coronary artery (RCA), and blue compound was injected into the left circumflex coronary artery (LCX) simultaneously. The compound solution was catalyzed with curing agent just before the injection. The ventricles were cut off with 1 cm of basal ventricular muscle attached to the atria. The atrial cavities were packed with cotton to maintain the shape of the atria during the following process. The heart was immersed in diluted ethyl alcohol for 5 to 7 days until dehydration and clearing were completed. Concentration of the ethyl alcohol was increased daily from 25% to 100%. Final clearing was carried out with methyl salicylate. Ventricular myocardium and the ascending aorta were carefully removed from the atria, so that only the atria were left. The specimen was observed and photographed with the atria immersed in methyl salicylate to maintain the clarity of visualization. Atrial Activation Sequence Because it was necessary to examine the activation sequence at the atrial septum and the posterior left atrium between the pulmonary vein orifices, which the maze procedure isolates, endocardial mapping of the atria was performed. The chest was opened through a median sternotomy, and the heart was suspended in a pericardial cradle. The dog was supported by normothermic cardiopulmonary bypass with vena caval and femoral artery cannulations. Two electrode molds carrying 104 and 108 unipolar electrodes each were inserted into the right and left atria through bilateral ventriculotomies across the atrioventricular annuli retrogradely. The mitral and tricuspid valve leaflets were excised to facilitate the insertion of electrode molds. The mapping system and method have been described previously [17]. The activation maps during sinus rhythm and atrial pacing were constructed. All animals received humane care in compliance with the Principles of Laboratory Animal Care formulated by the National Society of Medical Research and the Guide for the Care and Use of Laboratory Animals prepared by the National Academy of Science and published by the National Institutes of Health (NIH publication No , revised 1985). In addition, the study protocol was approved by the Washington University Animal Studies Committee. Results Atrial Coronary Arteries Figure 2 illustrates the distribution pattern of the atrial coronary arteries. In the present report, the anatomic Fig 2. Atrial coronary arteries in dogs. The upper panel represents the posterior epicardial aspect of the atria. The sinus node region is denoted as a dashed oval. The lower panel represents the right atrial aspect of the interatrial septum. Upper and posterior borders of the septum are denoted as a dashed line. (Ao aorta; CS coronary sinus ostium; FO fossa ovalis; IVC inferior vena cava; LAA left atrial appendage; LCX left circumflex coronary artery; MS membranous septum; PA pulmonary artery; PVs pulmonary veins; RAA right atrial appendage; RCA right coronary artery; SVC superior vena cava; TT tendon of Todaro; TV tricuspid valve; 1 anterior right atrial artery; 2 intermediate right atrial artery; 3 posterior right atrial artery; 4 anterior left atrial artery; 5 intermediate left atrial artery; 6 posterior left atrial artery; 7 atrioventricular node artery.) nomenclature of the atrial coronary arteries follows that of previously published descriptions [18, 19]. The RCA had two to three (mean, ) atrial arteries. The anterior right atrial artery was the first atrial artery of the RCA. This artery was small in most dogs and was absent in 2 (8%). However, in 3 dogs (12%), the anterior right atrial artery was highly developed, extended to the superior vena cava (SVC), and anastomosed with the intermediate right atrial artery (2 animals) or the anterior left atrial artery (1 animal). The intermediate right atrial artery, which was the second and usually the largest atrial artery of the RCA, arose at the midportion of the RCA. This artery was well developed and extended to the crista terminalis and sinus node in all animals except 1 in which the anterior left atrial artery and the anterior right atrial artery supplied the sinus node. The intermediate right atrial artery initially traversed the lower right atrium and then ran cranially along the crista terminalis toward the sinus node. This artery anastomosed with the

4 30 NITTA ET AL Ann Thorac Surg RADIAL APPROACH FOR AF: PART ;67:27 35 anterior left atrial artery and formed a vascular ring around the SVC in 78% of the animals. In 8 animals (31%), the posterior right atrial artery was present at the lower right atrium near the inferior vena cava. The LCX had two to three (mean, ) atrial arteries. The anterior left atrial artery was the first and largest atrial artery of the LCX. This artery supplied the superior left atrium, extended toward the SVC, and anastomosed with the intermediate right atrial artery of the RCA, forming a vascular ring around the SVC in 78% of the animals. There were one to two intermediate left atrial arteries that originated from the middle third of the LCX and supplied the base of the left atrial appendage and the posterior left atrium around the pulmonary veins. The posterior left atrial artery, which was the last epicardial atrial artery of the LCX and supplied the inferior left atrium, was present in 83% of the dogs. Coronary arteries of the interatrial septum were examined in 22 animals. Four atrial arteries were distributed in the interatrial septum, as illustrated in the lower panel of Figure 3: the anterior right atrial artery, the intermediate right atrial artery, the anterior left atrial artery, and the atrioventricular node artery. The anterior right atrial artery had branches to the anterior septum in 15 animals (68%). The branches of the anterior left atrial artery were distributed in the anteromedial part of the anterior limbus of the fossa ovalis in 88% of the animals. The branches of the intermediate right atrial artery were distributed in the crista terminalis and were further distributed posterolaterally in the posterior part of the anterior limbus in all animals except one in which the artery was small, as described earlier. The lower interatrial septum was perfused by the posterior left atrial artery and the atrioventricular node artery from the LCX. Atrial Activation Sequence During Sinus Rhythm Figure 3 shows the atrial endocardial activation sequence during normal sinus rhythm in a dog. The earliest activation was seen at the sinoatrial node region. The activation propagated along the crista terminalis and spread over the lateral right atrium. The activation at the posterolateral wall of the SVC collided with the other wavefront coming around the medial SVC from the opposite direction. The interatrial septum was activated by two wavefronts from the sinus node that propagated along the junction of the right atrium and the SVC: One wavefront propagated superiorly, then medially and inferiorly, and the other wavefront propagated in the opposite direction inferolaterally, then medially and superiorly. These two wavefronts reached the anterior limbus of the fossa ovalis 20 ms after sinus node activation, merged, and propagated anteriorly toward the atrioventricular annulus and node. Right atrial activation during sinus rhythm completed within 40 to 50 ms, and the median activation time in the right atrium was 24 ms. The latest activation of the right atrium was located laterally, adjacent to the atrioventricular annulus at 40 ms. The site of earliest left atrial activation in the posterior septal region (indicated by an asterisk in Fig 3) was depolarized from the adjacent site in the right atrial septum 20 to 30 ms after the onset of earliest activation in the right atrium. Wavefronts spread from this earliest left septal activation site in three directions. A wavefront traversed the anterosuperior left atrium (Bachmann s bundle) and activated the left atrial appendage and surrounding atrium. Other wavefronts propagated toward the posterior left atrium above, below, and between the pulmonary vein orifices, terminating in the left lateral wall of the left atrium at 60 ms. A third component of the wavefront propagated toward the atrioventricular annulus and inferiorly toward the lateral left atrium. These three wavefronts merged at the lateral left atrium, where the latest activation occurred. The left atrial activation was completed by 70 ms, and the median left atrial activation time was 45 ms after onset of activation in the right atrium. Comment Atrial Coronary Arteries and Activation Sequence in Dogs The observations in the canine atrial coronary arteries in the present study confirmed those of previously published reports [18, 19]. Briefly, the intermediate right atrial artery supplied the crista terminalis and the sinus node and formed a vascular ring around the SVC with the anterior left atrial artery in most dogs. The pattern of coronary artery distribution was centripetal, branching from the RCA or LCX at the right or left atrioventricular groove and spreading toward the sinus node. Furthermore, the present study demonstrated the anatomy of the canine coronary arteries supplying the interatrial septum, which has not been previously examined in detail. The important finding in designing the septal incision was that the anterior limbus was supplied by the anterior right or left atrial arteries anteriorly and by the intermediate right atrial artery posteriorly. After the maze procedure, the blood supply to the posterior septum may be interrupted by the transcaval longitudinal incision posteriorly and by the septal incision anteriorly. The atrial coronary arteries in humans have been previously examined either in cadaver hearts or angiographically in living humans [20, 21]. There are several differences in atrial coronary artery anatomy between humans and dogs. The intermediate right atrial artery is the sinus node artery in most canine hearts, whereas the sinus node is supplied by the anterior right atrial artery in approximately two thirds of humans and by the anterior left atrial artery in one third. The atrioventricular node artery arises from the distal LCX in dogs, whereas it arises from the distal RCA in humans. The intermediate right atrial artery branches to the crista terminalis and the posterior part of the anterior limbus of the fossa ovalis in dogs, but this artery is small in humans. The interatrial septum in humans is supplied by the anterior right or left atrial artery superiorly and by the atrioventricular node artery inferiorly [22]. The atrial activation during sinus rhythm has been

5 Ann Thorac Surg NITTA ET AL 1999;67:27 35 RADIAL APPROACH FOR AF: PART 1 31 Fig 3. Atrial endocardial activation maps during sinus rhythm in a normal dog. The boxed area on the electrocardiogram (ECG) is the data window analyzed to construct the activation maps. The atria were mapped endocardially with 212 unipolar electrodes mounted on sponge forms that were molded to fit the canine atria. The electrode molds were inserted into the atria through biventriculotomies across the atrioventricular valve annuli while the animal was supported by normothermic cardiopulmonary bypass. The wide QRS configuration in the electrocardiogram was the consequence of the ventriculotomies. The two middle maps represent the lateral (LAT) and septal (SEPT) surfaces of the right atrial (RA) endocardium. Three lower maps represent the left lateral, inferior (INF), and septal aspects of the left atrium (LA). The sinus node is indicated as an oval on the right atrium at the junction with the superior vena cava (SVC). The border of the interatrial septum is denoted as dashed lines. The activation sequence is indicated by arrows. The asterisk in the left atrial septum indicates the earliest activation site of the left atrial endocardium. (L.LAT left lateral; LLPVs left lower pulmonary veins; LUPV left upper pulmonary vein; MV mitral valve; RPVs right pulmonary veins; other abbreviations as in Fig 2.) previously demonstrated by epicardial [23] or right atrial septal maps [24]. The endocardial mapping of the atria in the present study disclosed the activation sequence at the left atrial septum as well as at the posterior left atrium between the pulmonary vein orifices, where epicardial mapping is technically difficult. There were some important findings in designing the atrial incisions of the radial approach. There are two paths between the sinus node and the interatrial septum; one rotating around the SVC clockwise and the other rotating counterclockwise. These two wavefronts merge and activate the septum at the posterior region of the anterior limbus. The left atrial

6 32 NITTA ET AL Ann Thorac Surg RADIAL APPROACH FOR AF: PART ;67:27 35 Fig 4. Location of the atrial incisions of the radial approach (right panels) and the maze procedure (left panels), as well as the postoperative activation sequence. Upper, middle, and lower panels represent the superior and posterior epicardium and septum of the atria, respectively. The small dark region of the right atrium at the junction with the superior vena cava (SVC) represents the sinus node. Dashed lines indicate the atrial incisions, and the arrows indicate the activation sequence after the procedures. The darkly shaded regions indicate the excised or isolated regions. The small lightly shaded circles indicate cryolesions. The asterisks in the middle and lower panels are the epicardial and endocardial aspects of the identical sites. (Abbreviations are as in Figs 2 and 3.) activation begins at the posterosuperior septum. Early activation of the left septum could be blocked by the septal incision of the maze procedure, delaying the onset of left atrial depolarization. Another important finding was that the posterior left atrium was activated by the wavefront propagating inferiorly from the superior left atrium through the atrial tissue between the right and left upper pulmonary veins. The conduction from the earliest left septal activation site to the posterior left atrium may be faster through this pathway than through the atrial tissue between the upper and lower right pulmonary veins. This statement is true because the right pulmonary veins form a common orifice with a narrow band of atrial tissue between each vein. Development of a Surgical Procedure On the basis of the coronary artery distribution and activation sequence of the atria described in the present report, we designed the atrial incisions following the concept of the radial approach. Figure 4 illustrates the atrial incisions and the subsequent activation sequence during sinus rhythm after the radial approach and after the maze procedure. The right atrial incisions are similar between the two procedures, except for the excision of the right atrial appendage. The right atrial appendage can participate in atrial reentry as an anatomic obstacle in conjunction with or without the SVC. It also has many bridging endocardial trabeculae that could form part of a potential reentrant pathway [25]. However, the right atrial appendage has been shown to secrete atrial natriuretic peptides [26 28], and reduced secretion of these peptides has been demonstrated in patients after the maze procedure [29, 30]. This reduced secretion of atrial natriuretic peptides could be one of the mechanisms for the postoperative complication of fluid retention frequently seen in patients after the maze procedure. To preserve secretion of atrial natriuretic peptides, the right atrial appendage is not excised in the radial approach but is incised to prevent reentry at the lateral right atrium. In addition, major bridging trabeculae in the right atrial appendage are divided to eliminate reentry using these structures. The right atrial appendage incision is extended down to the tricuspid valve annulus anteromedially and from the appendage tip in the opposite direction, toward the lower right atrium inferiorly, as in the maze procedure. The other two right atrial incisions are the same as in the maze procedure. The intercaval longitudinal incision at the posterior right atrium blocks the reentry that uses the vena cava as an anatomic obstacle. During development of the radial approach procedure, we found that reentry around the SVC could occur. Atrial flutter was inducible in two of four animals in which this incision was not extended to the SVC. The activation maps confirmed the reentry around the SVC as a mechanism for the atrial flutter. Therefore, we believe that this incision is necessary in surgical treatment for AF. This incision does not change the activation sequence at the lateral right atrium. It blocks one of the pathways from the sinus node to the interatrial septum; however, the other pathway can maintain the physiologic activation of the septum. Although the incision also interrupts the blood supply to the posterior septum from the intermediate right atrial artery, the other arteries, such as the anterior left atrial artery, may be enough to supply the septum. The transverse incision at the lower right atrium blocks the reentry around the tricuspid valve annulus. Although this incision alters the activation sequence of the right atrium by shifting the collision zone of the wavefronts from the septum to lateral wall of the lower right atrium, the difference in the activation sequence is insignificant. In addition, it would be technically difficult to place a transverse incision in the septum at the level of the coronary sinus ostium where the wavefronts collide in the normal canine right atrium. The transverse incision does not interrupt any major atrial arteries, as long

7 Ann Thorac Surg NITTA ET AL 1999;67:27 35 RADIAL APPROACH FOR AF: PART 1 33 Fig 5. Schematic representation of anatomic correlation between the atrial incisions and the atrial activation pattern during sinus rhythm in normal atria and in the atria after the maze procedure or the radial approach. Thick lines represent the surgical incisions and the solid area represents the part of atria that is surgically isolated or excised. Dashed lines represent Bachmann s bundle between the atrial appendages, the interatrial septum, and the crista terminalis. Arrows indicate the activation sequence. (AVN atrioventricular node; SAN sinoatrial node; other abbreviations are as in Fig 2.) as it is placed at the lower right atrium low enough to avoid injury to the intermediate right atrial artery in canine hearts. Therefore, these two incisions are essential in eliminating AF and do not result in significant changes in either atrial activation sequence or myocardial blood supply. The incisions in the left atrium and interatrial septum in the radial approach are entirely different from those of the maze procedure. One incision, beginning at the anterior limbus of the fossa ovalis, extends inferoposteriorly toward the lower posterior interatrial septum and to the right posteroinferior wall of the left atrium, passing near the right and left lower pulmonary vein orifices, and continues down to the mitral valve annulus at the commissure between the middle and posteromedial scallops. The other incision, beginning at the superior left atrium between the right and left upper pulmonary veins, connects with the left atrial appendage excision line and extends anteromedially further downward to the mitral valve annulus at the anterolateral commissure. These two incisions divide the left atrium into three segments: the upper, middle, and lower left atrium. These segments are connected to each other at the septum and at the superior left atrium. Because these incisions parallel the activation sequence during sinus rhythm, a synchronous left atrial activation sequence is maintained. The data in the present study suggest that the septal incision of the maze procedure interrupts the blood supply to the posterior septum from the anterior left atrial artery and that the incision locates the earliest activation site of the left septum. This incision may cause ischemia in certain areas of the septum and may impair the physiologic activation sequence of the septum. In the radial approach, the septal incision is placed from the posterior lower septum extending up to the anterior limbus, so that the blood supply to most areas will not be disturbed and the normal septal activation is preserved. The posterior left atrium, which is isolated in the maze procedure, is activated from the superior left atrium between the right and left upper pulmonary veins in the radial approach. This activation sequence is the same as the sequence in normal canine atria. To prevent reentry around the right upper pulmonary vein, the narrow isthmus of atrial tissue between the upper and lower

8 34 NITTA ET AL Ann Thorac Surg RADIAL APPROACH FOR AF: PART ;67:27 35 right pulmonary veins is cryoablated. Moreover, the atrial tissue between the left atrial incision and each pulmonary vein orifice is also cryoablated to prevent reentry around one or more orifices. At the superior left atrium between the right and left upper pulmonary veins, the area of the atrial myocardium can be broad and large enough to maintain reentry in this region, particularly in patients with a dilated left atrium. Therefore the upper left atrial incision, beginning at the left atrial appendage excision line, is extended over the left upper pulmonary vein and toward the right upper pulmonary vein orifice to reduce the effective atrial area and to prevent reentry at this region. The studies on atrial coronary artery anatomy and the activation sequence during sinus rhythm described in the present report suggest that the atrial incisions of the radial approach parallel the activation sequence and atrial coronary arteries in canine atria. Moreover, the basic patterns of the atrial coronary arteries and the activation sequence are similar between dogs and humans. Therefore the radial approach should also preserve a more physiologic activation sequence and blood supply to most atrial segments in patients. To characterize and compare the pattern of activation sequence after the two procedures, correlation between the atrial activation and the atrial incisions is illustrated in Figure 5, with reference to the preoperative activation sequence. The difference in the right atrial activation sequence is small after the radial approach or the maze procedure. However, the activation sequence in the left atrium is markedly different in the two procedures. The left atrial incisions of the radial approach barely alter the preoperative activation sequence because the incisions parallel the activation sequence. The difference between the procedures is evident in three regions: the posterior left atrium, the left lateral left atrium, and the interatrial septum. The most significant difference is that the posterior left atrium is activated physiologically in the radial approach, whereas this region is isolated and inexcitable in the maze procedure. The activation at the lateral left atrium is also different in the two procedures. In the maze procedure, this region is activated by the wavefront propagating from the superior left atrium and detouring around the left atrial appendage excision line. The path length from the sinus node to this region is longer than the preoperative path length, suggesting delayed activation in the region. In the radial approach, the activation sequence of this region is the same as the preoperative sequence. The activation at the posterior septum reaches a dead end in both procedures. However, the activation sequence is more physiologic in the radial approach than in the maze procedure, in which the activation turns around at the edge of the septal incision. These differences in the left atrial activation sequence should result in a significant difference in the left atrial transport function between the two procedures. This hypothesis is tested in a subsequent report [16]. Conclusions A new surgical method, the radial approach, was developed as an outgrowth of and an alternative to the maze procedure, to both eliminate AF and preserve normal electrophysiologic and mechanical function of the atria. The procedure has been evaluated in chronic canine studies with respect to electrophysiologic and hemodynamic variables, and the results are described in a subsequent report [16]. This study was supported by the National Institutes of Health (5 T32 HL007776, 5 RO1 HL32257, and 5 RO1 HL33722). References 1. Cox JL, Schuessler RB, D Agostino HJ, 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, Kater KM, Lappas DG. Five-year experience with the maze procedure for atrial fibrillation. Ann Thorac Surg 1993;56: Cox JL, Boineau JP, Schuessler RB, Jaquiss RD, 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, Schuessler RB, Lappas DG, Boineau JP. An year clinical experience with surgery for atrial fibrillation. Ann Surg 1996;224: Kosakai Y, Kawaguchi A, Isobe F, et al. Cox maze procedure for chronic atrial fibrillation associated with mitral valve disease. J Thorac Cardiovasc Surg 1994;108: Isobe F, Kawashima Y. The outcome and indications of the Cox maze III procedure for chronic atrial fibrillation with mitral valve disease. J Thorac Cardiovasc Surg 1998;116: Feinberg MS, Waggoner AD, Kater KM, Cox JL, Lindsay BD, Pérez JE. Restoration of atrial function after the maze procedure for patients with atrial fibrillation: assessment by Doppler echocardiography. Circulation 1994;90(Suppl II):II Itoh T, Okamoto H, Nimi T, et al. Left atrial function after Cox s maze operation concomitant with mitral valve operation. Ann Thorac Surg 1995;60: Yashima N, Nasu M, Kawazoe K, Hiramori K. Serial evaluation of atrial function by Doppler echocardiography after the maze procedure for chronic atrial fibrillation. Eur Heart J 1997;18: Tsui SS, Grace AA, Ludman PF, et al. Maze 3 for atrial fibrillation: two cuts too few? PACE Pacing Clin Electrophysiol 1994;17: Hartford M, Wikstrand J, Wallentin I, Ljungman S, Wilhelmsen L, Berglund G. Diastolic function of the heart in untreated primary hypertension. Hypertension 1984;6: Bonow RO, Bacharach SL, Green MV, et al. Impaired left ventricular diastolic filling in patients with coronary artery disease: assessment with radionuclide angiography. Circulation 1981;64: Appleton CP, Hatle LK, Popp RL. Relation of transmitral flow velocity patterns to left ventricular diastolic function: new insights from a combined hemodynamic and Doppler echocardiographic study. J Am Coll Cardiol 1988;12: Vanoverschelde JL, Raphael DA, Robert AR, Cosyns JR. Left ventricular filling in dilated cardiomyopathy: relation to functional class and hemodynamics. J Am Coll Cardiol 1990; 15: Klein AL, Tajik AJ. Doppler assessment of pulmonary venous flow in healthy subjects and in patients with heart disease. J Am Soc Echocardiogr 1991;4:

9 Ann Thorac Surg NITTA ET AL 1999;67:27 35 RADIAL APPROACH FOR AF: PART Nitta T, Lee R, Watanabe H, et al. Radial approach: a new concept in surgical treatment for atrial fibrillation II. Electrophysiologic effects and atrial contribution to ventricular filling. Ann Thorac Surg 1999;67: Cronin CS, Nitta T, Mitsuno M, et al. Characterization and surgical ablation of acute atrial flutter following the Mustard procedure: a canine model. Circulation 1993;88(Suppl II):II Meek WJ, Keenan M, Theisen HJ. Auricular blood supply in the dogs: 1. General auricular supply with special reference to the sino-auricular node. Am Heart J 1929;4: Moore RA. The coronary arteries of the dog. Am Heart J 1930;5: Hutchinson MC. A study of the atrial arteries in man. J Anat 1978;125: James TN, Burch GE. The atrial coronary arteries in man. Circulation 1958;17: Kugel MA. Anatomical studies on the coronary arteries and their branches. 1. Arteria anastomotica auricularis magna. Am Heart J 1927;3: Cox JL, Canavan TE, Schuessler RB, et al. The surgical treatment of atrial fibrillation. II. Intraoperative electrophysiologic mapping and description of the electrophysiologic basis of atrial flutter and atrial fibrillation. J Thorac Cardiovasc Surg 1991;101: Chang BC, Schuessler RB, Stone CM, et al. Computerized activation sequence mapping of the human atrial septum. Ann Thorac Surg 1990;49: Schuessler RB, Kawamoto T, Hand DE, et al. Simultaneous epicardial and endocardial activation sequence mapping in the isolated canine right atrium. Circulation 1993;88: Veress AT, Sonnenberg H. Right atrial appendectomy reduces the renal response to acute hypervolemia in the rat. Am J Physiol 1984;247:R Omari BO, Nelson RJ, Robertson JM. Effect of right atrial appendectomy on the release of atrial natriuretic hormone. J Thorac Cardiovasc Surg 1991;102: Stewart JM, Dean R, Brown M, et al. Bilateral atrial appendectomy abolishes increased plasma atrial natriuretic peptide release and blunts sodium and water excretion during volume loading in conscious dogs. Circ Res 1992;70: Nakamura M, Niinuma H, Chiba M, et al. Effect of the maze procedure for atrial fibrillation on atrial and brain natriuretic peptide. Am J Cardiol 1997;79: Kim KB, Lee CH, Kim CH, Cha YJ. Effect of the Cox maze procedure on the secretion of atrial natriuretic peptide. J Thorac Cardiovasc Surg 1998;115: DISCUSSION DR DAVID A. OTT (Houston, TX): I congratulate you and your associates for your conceptual innovation in designing this new variant of the maze operation based on atrial coronary anatomy. And you are also to be commended for evaluating it in an elegant and scientifically sound fashion. It appears that you have devised a physiologic operation that cures atrial fibrillation while preserving the atrial arterial blood supply and the atrial transport function. This also apparently avoids the electrical and mechanical isolation of the posterior portion of the left atrium, which is one of the shortcomings of the classical maze operation. Ideally we need an operation for atrial fibrillation that can be expeditiously performed concomitantly with mitral valve repair or other cardiac procedures during which definitive treatment of the associated atrial fibrillation would be a worthwhile ancillary procedure. My question is, Do you now have human experience with this operation, and do you believe that this new operation will prove to be easier and quicker to perform and thus more universally applicable than the classical maze procedure? DR NITTA: Thank you, Dr Ott, for the thoughtful comments and questions. It is always difficult to quantify how easy or difficult a procedure is; but I think it is safe to say that the radial approach is easier to perform because the incisions are more linear, and there is no isolation incision or T-shaped incision in the left atrial incision of the radial approach. The pulmonary vein isolation incision and the T-shaped incision of the left atrium make the maze procedure technically difficult. These technical difficulties of the maze procedure prompted some surgeons to modify the procedure. However, most of the modifications were only aimed at lowering the cross-clamp time or reducing the risk of bleeding from the suture lines. Unfortunately, none of the modifications were considered or evaluated in terms of atrial transport function. Regular heartbeats can be easily achieved by atrioventricular node ablation and pacemaker implantation. However, we need to remember the purpose of the atrial fibrillation operation, which includes not only the restoration of regular sinus rhythm but also the restoration of atrial transport function and the prevention of thromboembolism. We believe that the concept of the radial approach is extremely important to preserve a more physiologic atrial activation and optimize the atrial transport function. The duration of aortic cross-clamp time was similar between the procedures in this canine model. However, seven cryolesions are performed during the aortic cross-clamping in the radial approach, whereas three are performed in the maze procedure. Therefore, if you use multiple cryoprobes simultaneously, you can significantly decrease the cross-clamp time. I am currently using two cryoprobes with different tip sizes in patients and have found that approximately 30 to 40 minutes of additional aortic cross-clamp time is necessary to perform the radial approach concomitantly with mitral valve operation. We therefore believe that the radial approach is more universally applicable than the maze procedure.