Several groups have reported the use of linear lesions

Transcription

1 Comparison of Epicardial and Endocardial Linear Ablation Using Handheld Probes Stuart P. Thomas, PhD, Duncan J. R. Guy, MBBS, Anita C. Boyd, BMedSc, Vicki E. Eipper, David L. Ross, MBBS, and Richard B. Chard, MBBS Departments of Cardiology and Cardiothoracic Surgery, Westmead Hospital, Westmead, New South Wales, Australia Background. The optimal technique for producing linear radiofrequency thermal lesions in myocardial tissue is unclear. We compared epicardial ablation on the beating heart with endocardial ablation after cardioplegia. Methods. Radiofrequency lesions were produced using a multielectrode malleable handheld probe in ovine myocardium with three wall thicknesses. Detailed analysis of lesion dimensions was used to assess the effects of site of ablation, muscle thickness, and duration of ablation. Results. After epicardial atrial ablation, myocardial lesions were detected in all sections without macroscopically visible epicardial fat (n 10), but only 43% (6/14) of sections with epicardial fat. Three of 24 atrial epicardial sections (13%) and 92% (23/25) of endocardial atrial lesion sections were clearly transmural. In thicker tissues lesion depth was independent of endocardial (right ventricle: mm, left ventricle: mm) or epicardial (right ventricle: mm, left ventricle: mm) ablation site. Epicardial lesions are less deep in thinner areas of myocardium (p 0.003). Lesions were all wider than they were deep. There was no significant increase in lesion depth with the increase in ablation duration from 1 to 2 minutes. Conclusions. Lesions were unlikely to be transmural with either technique when the wall thickness was greater than about 4 mm. Epicardial fat has an important negative effect on epicardial lesion formation. Where epicardial fat is absent epicardially produced lesions penetrate less deeply when the wall thickness is small, possibly due to endocardial cooling by circulating blood. Prolongation of the duration of ablation from 1 to 2 minutes does not significantly increase lesion depth. (Ann Thorac Surg 2003;75:543 8) 2003 by The Society of Thoracic Surgeons Several groups have reported the use of linear lesions formed with radiofrequency energy to treat atrial fibrillation [1 6]. The results of these procedures have been encouraging, but fall short of the success rates reported for the Cox maze procedure in which the atria are segmented by incisions rather than radiofrequency lesions [7, 8]. An endocardial approach was used in early studies of linear radiofrequency ablation for atrial fibrillation. A single or multielectrode handheld probe was inserted through an atriotomy to produce lesions by passing radiofrequency current between the electrodes and a large indifferent electrode placed on the skin or between For editorial comment see page 329 the endocardial and an epicardial electrode [6]. Recently some investigators have produced linear lesions with the ablation probe positioned on the epicardial surfaces of the atria [1, 5]. This technique may simplify the procedure by allowing the lesions to be formed on a beating heart. The effects of this alternative approach on the effectiveness of lesion formation are unknown. Several Accepted for publication Aug 19, Address reprint requests to Dr Thomas, Department of Cardiology, Westmead Hospital, Westmead, New South Wales 2145, Australia; stuartpt@yahoo.com. factors including the structure of the epicardium, the presence of blood perfusing the myocardium, and the presence of blood cooling the endocardial surface during ablation may alter lesion formation. The purpose of this study was to compare the site of ablation, endocardial versus epicardial, on radiofrequency energy induced lesion formation in myocardium. We also examined the effect of myocardial wall thickness and duration of ablation on lesion formation. A sheep model was used to compare lesions formed either epicardially on the beating heart or on the endocardial surface after cardioplegia. The atria of sheep are very thin and structurally different from those in humans. The wall thickness of the sheep right ventricle is closer to that of diseased human atria. The prominent trabeculation of the ovine right ventricle also mimics that of the human trabeculated atrium. Therefore, atrial and right ventricular lesions were produced. Lesions were also produced in left ventricular tissue. The thick left ventricular tissue allows comparison of lesion size unobstructed by substrate boundaries and independent of the endocardial blood cooling effect. Material and Methods Ten Merino wether sheep weighing 35 2 kg were used for the studies. After a premedication of intramuscular xylazine (40 mg), anesthesia was induced with intravenous thiopental (500 mg). Endotracheal intubation was 2003 by The Society of Thoracic Surgeons /03/$30.00 Published by Elsevier Science Inc PII S (02)04314-X

2 544 THOMAS ET AL Ann Thorac Surg RF ABLATION USING HANDHELD PROBES 2003;75:543 8 performed and intermittent positive pressure ventilation maintained with a Harvard Respirator (Harvard Apparatus, Holliston, MA). Anesthesia was maintained with isoflurane (2% to 4%). Heart rate, oxygen saturation, expired CO 2 levels and cardiac rhythm were continuously monitored throughout the experiments. In all animals a thoracotomy was performed at the fourth left intercostal space to expose the heart. In 5 sheep, ablation was performed directly on the epicardial surface of the beating heart. Two long linear lesions were possible on the epicardial surface of the atria (right and left), two lesions were produced in the right ventricle, and a further two lesions were produced in the left ventricle of each animal. In the remaining 5 sheep the aorta was cross-clamped and cardioplegia solution (Solution A, Baxter, Deerfield, IL) administered antegradely at a temperature of less than 5 C through an aortic cannula until electromechanical arrest was achieved. The cardiac chambers were opened and vented where appropriate to allow access to the endocardial surface. Two long linear lesions were possible in the atria (left and right), two lesions were produced in the right ventricle and a further two lesions were produced in the left ventricle of each animal. Care was taken in all cases to ensure a large area of viable tissue separated the linear lesions. The positioning of lesions was consistent throughout but the ablation durations were alternated at each site to prevent systematic bias. Lesions were approximately 50 mm long. Ablation Procedure A multielectrode handheld ablation device was used to produce all lesions (Cobra, Boston Scientific, Boston Scientific-EPT, San Jose, CA). This probe had seven electrodes including six 12.5 mm coil electrodes wound around a 7F circular shaft with an interelectrode distance of 2 mm. The distal electrode was 8 mm long. Simultaneous, in-phase, unipolar ablation was performed between selected electrodes and a large surface electrode. Radiofrequency current was provided by a 150-W generator (Boston Scientific). The electrode temperatures during ablation were controlled using two thermocouples attached to the margins of each electrode. A target temperature of 85 C was used for each energy delivery. The duration of radiofrequency energy application was either 1 or 2 minutes. Lesion Analysis After completion of the ablation protocol the animals were killed and the hearts excised. Lesions were examined macroscopically and microscopically. Sections were excised in a plane perpendicular to the long axis of the lesions. The spacing of sections was 5 to 10 mm. Sections of each lesion ( sections/lesion) were incubated in a phosphate-buffered solution of blue tetrazolium containing succinate at 37 C for 30 to 45 minutes. This technique results in blue staining of viable myocardium. The sections were digitally photographed. Analysis of lesion width and depth were performed with computerized image analysis software (Scion Image; Scion, Frederick, MD) by two blinded observers. Where there was doubt about the extent of the lesion the sections were placed in formalin and sectioned further for microscopic analysis. Eight lesions were also examined microscopically to characterize the histopathological features of the lesions. Four sections were made from each lesion. Gomori s Trichrome stain was used to identify the lesion margins. The width and depth of lesions were measured. The study was approved by the Western Sydney Area Health Service Animal Ethics Committee and conducted in a manner conforming to the ethical and scientific principles set out by the National Health and Medical Research Council of Australia. All animals received humane care in accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health publication 85-23, revised 1985). Statistics Sections were averaged for each lesion before statistical analysis of lesion sizes. Expressions of transmurality describe individual sections. Analysis of variance was used to assess the effects of lesions site and duration of ablation on lesion volume. Differences were considered statistically significant if the p value was less than Tissue thickness comparisons were performed using Student s t test. Continuous variables are expressed as mean standard deviation. Results Identification of Lesions Endocardial or epicardial changes were clearly visible in most cases after application of radiofrequency energy. Lesion formation was characterized by blanching of the region adjacent to the electrode. Staining with tetrazolium blue allowed clear identification of lesion margins. In a selection of cases the tissue was subsequently fixed and histologic sections were produced. The size of lesions identified on the Gomori Trichrome stained sections corresponded to the lesion size identified by tetrazolium blue. Lesions were pale compared with the surrounding tissue after staining with tetrazolium blue. Microscopic examination showed loss of cell definition, separation of the fibers by edema, loss of nuclei and cross-striations, and formation of contraction bands (Fig 1). Extravascular red blood cells were present at the margins of the lesions. The borders of the lesions were readily identifiable with the Gomori Trichrome stain (Fig 2). Lesion Formation in Thin Myocardium: Ovine Atria The overall wall thickness of the myocardium in each section was measured. The mean muscle thickness of the left and right atria (not including any epicardial fat) was mm and mm, respectively. Macroscopically visible fat was present on the epicardial surface in 65% (17/26) of left atrial sections and 26% (6/23) of right atrial sections. Where fat was present the thickness of the fat layer was mm in the left atrium and mm in the right atrium.

3 Ann Thorac Surg THOMAS ET AL 2003;75:543 8 RF ABLATION USING HANDHELD PROBES 545 Fig 1. Microscopic appearance of lesion ( 40 before 56% reduction). The lesions are characterized by loss of cell margins, separation of cells due to edema, loss of nuclei, and contraction banding. Typically there is a hemorrhagic border zone (not shown). The presence of epicardial fat modified lesion formation during epicardial application of radiofrequency energy. Fourteen of the 24 (58%) sections analyzed after application of epicardial radiofrequency current contained macroscopically visible epicardial fat. The thickness of the fat layer in these sections was mm. Myocardial lesions were detected in only six of the sections containing fat (43%). Lesions were detected in all sections in which no epicardial fat was present. Only three of 24 atrial epicardial lesions were transmural (13%). Where lesions were present the maximum lesion depth and width were and mm, respectively. In contrast to the findings after epicardial ablation 92% (23/25) of endocardial atrial lesion sections were clearly transmural (Fig 2B). The two nontransmural sections were in unusually thick left atrial myocardium ( mm). The lesion depth in those cases was mm. The width of the endocardial lesions was mm. Lesion Formation in Myocardium of Moderate Wall Thickness: Ovine Right Ventricle No fat was noted on the epicardial surface at the ventricular lesion sites. The wall thickness of the right ventricles was mm. The epicardial lesion depth in right ventricular tissue ( mm) was smaller than the endocardial lesion depth in the right ventricle ( mm, n 20) but this difference was not statistically significant. The epicardial lesion depth in the right ventricle was significantly smaller than the epicardial lesion depth in the left ventricle ( mm, p 0.02, n 20). There was a significant proportional relationship between right and left ventricular wall thickness and lesion depth during epicardial ablation (p 0.003). This relationship was not observed in endocardial lesions. This finding suggests the cooling effect of endocardial blood flow may reduce the depth of lesions in thinner tissues. Fig 2. (A) Macroscope image showing the typical appearance of radiofrequency lesions (Gomori Trichrome). The sections are cut in a plane perpendicular to the long axis of the lesion. The ablated area (Abl) contains a greater concentration of blue tones. Viable tissue (Normal) is red. This example shows a nontransmural lesion in thick atrial muscle. The lesion was produced by epicardial ablation for 120 seconds. (B) A section stained with Gomori Trichrome of the posterior free wall of the left atrium showing a transmural lesion produced by endocardial ablation. The width of the lesion in this example is more than 10 mm. The coronary sinus is marked (Cs). No significant difference was noted in the width of lesions produced by the epicardial ( mm) or endocardial ( mm) approach. Lesion Formation in Very Thick Tissue: Ovine Left Ventricle The thickness of the left ventricular myocardium was mm. In left ventricular tissue, the depth of lesions produced by endocardial ablation was not significantly different to those produced by epicardial ablation. The endocardial and epicardial lesion depths were and mm, respectively. A trend was noted toward a reduction in lesion depth for endocardial le-

4 546 THOMAS ET AL Ann Thorac Surg RF ABLATION USING HANDHELD PROBES 2003;75:543 8 statistically independent of the endocardial or epicardial placement of the lesions or the position of the lesion in the right or left ventricular myocardium. Fig 3. Lesion expansion with increasing duration of ablation. The error bars represent the standard deviations. (Top) Depth. (Bottom) Width. sions, but this difference was not statistically significant. This trend may be due to the lower baseline tissue temperature during endocardial ablation. The lower temperature during ablation is due to the prior administration of cold cardioplegia and the lack of blood perfusion. The widths of endocardial and epicardial lesions in the left ventricle were and mm, respectively (p NS). Effect of Duration of Ablation on Lesion Size The small wall thickness of myocardium and the confounding influence of epicardial fat prevented the analysis of duration of ablation in the ovine atrial tissue. Therefore, analysis of the effects of wall thickness and duration of ablation on lesion development was performed using measurements of lesions from the right and left ventricles. The target electrode temperatures were achieved in all radiofrequency energy applications. In all cases there was a small overshoot in the temperature. No significant increase was noted in lesion depth or width with the increase in ablation duration from 1 to 2 minutes. These results are illustrated in Figure 3. This finding was Comment Determinants of Lesion Formation in Thin Myocardium: Epicardial Fat and Endocardial Flow A major finding of this study was that epicardial application of radiofrequency energy in the beating heart is less likely to produce transmural lesions than a similar application on the endocardial surface after cold cardioplegia. One factor contributing to this outcome was the presence of epicardial fat. Epicardial fat has a detrimental effect on lesion formation. Epicardially placed lesions were not formed or were small at sites where epicardial fat was present. This finding has important implications for epicardial ablation in humans. Linear radiofrequency lesions extending to the atrioventricular junction, as required by the Cox maze patterns, are unlikely to be transmural if significant epicardial fat is present. Even when no epicardial fat was present, the epicardial lesions in atrial tissue were still smaller than those produced by application of radiofrequency energy on the endocardial surface and frequently not transmural. Epicardial radiofrequency application also produced smaller lesions in the ovine right ventricles, but the difference in lesion depth was less marked and did not achieve statistical significance. Endocardial and epicardial lesions were similar in size in the thicker left ventricular tissue, but epicardial lesions were smaller in the right ventricle than similar lesions in the left ventricle. These findings suggest the relative reduction in lesion size with epicardial ablation is inversely related to the myocardial thickness. This relationship may be due to endocardial cooling in the thinner myocardium. The cooling effect of circulating blood in the underlying cavity is likely to reduce lesion size during epicardial ablation. In the human atrial myocardium this effect may prevent lesions from penetrating the full thickness of myocardial tissue and create discontinuities in lines of ablation. Duration of Ablation Increasing the duration of ablation from 60 to 120 seconds did not significantly increase lesion size. This finding is consistent with previous studies of lesion formation using radiofrequency energy in the myocardium [9 11]. These studies showed that most of the lesion depth is achieved in the first 15 to 30 seconds and that increases in duration of ablation beyond 60 seconds produce only a small increment in lesion size. Lesion Width The optimal lesions for linear ablation are deep and narrow. Such lesions would provide electrical disruption along the line of ablation without excessive damage to contractile myocardium. We have previously shown that linear lesions similar to those produced in this study reduce atrial contraction [12]. All lesions in the present

5 Ann Thorac Surg THOMAS ET AL 2003;75:543 8 RF ABLATION USING HANDHELD PROBES 547 study were wider than they were deep. Width of lesions was independent of the thickness of tissue and the duration of ablation. Optimal Lesion Depth The Cox maze procedures septate the atria causing electrical isolation of adjacent areas of myocardium. When radiofrequency ablation was initially used to replace the incisions of the Cox operations, it was assumed that continuous transmural lesions would be required. The present study demonstrates that radiofrequency ablation applied through the handheld probe electrodes to the endocardial surface during cardioplegia produces lesions mm in depth in the right ventricle and mm in the left ventricle. Where lesions were not transmural in atrial tissue because the atrial wall was thick, lesion depth was identical to lesion depths in ventricular tissue. These findings indicate when radiofrequency energy is applied to the endocardial surface the probe is capable of consistently producing lesions that would be transmural in thin parts of the atria. Most of the human atrium is less than 5 mm thick, but structures such as the crista terminalis, the posterior left atrial free wall, and the trabeculations of the right atrial free wall in hypertrophic hearts are up to 6 mm thick [13, 14]. The results of the present study indicate that lesions in the thicker parts of the human atria are unlikely to be transmural even when produced from the endocardial surface. Patterns of linear radiofrequency lesions should be designed with the regional differences in atrial wall thickness in mind to reduce the risk of discontinuities in lines of ablation. Even where there was no epicardial fat, lesions produced by epicardial ablation were often not transmural in the atria. Previous studies have demonstrated that conduction can proceed over a small surviving isthmus of tissue [15, 16]. Most linear ablation procedures for atrial fibrillation incorporate lesions designed to isolate the pulmonary veins electrically. Early experience with transvenous catheter techniques for pulmonary vein isolation has demonstrated that failure to isolate the veins adequately often leads to arrhythmia recurrence. This principle is also likely to apply to the surgical isolation of the pulmonary veins. Furthermore, gaps in linear lesions may also lead to the development of fixed reentrant circuits that manifest clinically as atrial flutter [2]. Implications for Linear Ablation in Humans The lesions created in this study were in normal ovine myocardium. Long-standing atrial fibrillation is characterized histopathologically by diffuse interstitial fibrosis. Scarring in the myocardium may alter the biophysics of radiofrequency ablation. However, a study of radiofrequency ablation in normal and scarred myocardium performed using the percutaneous catheter technique suggested this effect is small and unlikely to be clinically significant [17]. Another important difference between ovine and human atrial myocardium is the prominence of endocardial fibrosis in diseased human atria. The fibrosis may reduce the penetration of lesions formed from the endocardial surface. The limited depth of lesions produced by radiofrequency energy restricts the patterns of lesions that may be used and may limit the use of this device for epicardial ablation. Probes capable of producing deeper lesions are necessary to provide greater versatility with lesion placement. Devices incorporating irrigated electrodes produce deeper lesions in the setting of percutaneous transvenous catheter ablation [18]. Further studies are required to confirm that this finding also applies to intraoperative ablation and determine whether the increased lesion depth results in improved outcomes. Study Limitations The myocardial tissue was not reperfused before sacrifice of the animals in this study. The major determinant of lesion formation is direct thermal injury but damage to vessels coursing through the lesion and supplying adjacent areas of nonablated myocardium may produce ischemia and extension of the lesion on reperfusion. Nath and colleagues [19] demonstrated evidence of relative ischemia in the region around the limits of the pathologic scar produced by radiofrequency ablation. However, the effect of this relative ischemia on lesion size is unclear. Another study by the same group comparing lesion size in perfused and nonperfused myocardium showed lesion formation was independent of flow [11]. Large areas of infarction outside the expected limits of ablation have not been observed in chronic animal studies of linear ablation. Therefore any effect is likely to be small. Accurate determination of the influence of this mechanism on lesion size would require further study. Conclusions Lesions produced in this study were limited in depth and unlikely to be transmural in the thickest parts of the atria. Epicardial fat has an important negative effect on lesion formation. Where fat is absent, epicardially delivered radiofrequency energy still results in smaller lesions, possibly due to endocardial cooling by circulating blood. Radiofrequency lesions were wider than they were deep, so deep penetration of lesions into the myocardium can only be achieved by producing broad lesions with large volumes. Prolongation of the duration of ablation from 1 to 2 minutes does not significantly increase lesion depth. Patterns of lesions need to be designed that avoid areas of the atrial myocardium with a wall thickness of more than 3.5 mm and regions where epicardial fat is present. This work was supported by a grant from the National Health and Medical Research Council of Australia (980411). The authors thank Craig Campbell and Nicholas Kang for their technical assistance. The ablation devices were provided by Boston Scientific. References 1. Benussi S, Pappone C, Nascimbene S. A simple way to treat chronic atrial fibrillation during mitral valve surgery: the

6 548 THOMAS ET AL Ann Thorac Surg RF ABLATION USING HANDHELD PROBES 2003;75:543 8 epicardial radiofrequency approach. Eur J Cardiothorac Surg 2000;17: Thomas SP, Nunn GR, Nicholson IA. Mechanism, localization, and cure of atrial arrhythmias occurring after a new intraoperative endocardial radiofrequency ablation procedure for atrial fibrillation. J Am Coll Cardiol 2000;35: Kottkamp H, Hindricks G, Hammel D. Intraoperative radiofrequency ablation of chronic atrial fibrillation: a left atrial curative approach by elimination of anatomic anchor reentrant circuits. J Cardiovasc Electrophysiol 1999;10: Imai K, Sueda T, Orihashi K, Watari M, Matsuura Y. Clinical analysis of results of a simple left atrial procedure for chronic atrial fibrillation. Ann Thorac Surg 2001;71: Melo J, Adragao P, Neves J. 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. Intraoperative radiofrequency microbipolar coagulation to replace incisions of the maze III procedure for correcting atrial fibrillation in patients with rheumatic valvular heart disease. Eur J Cardiothoracic Surg 1997;12: Cox JL, Schuessler RB, D Agostino HJ Jr. 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: Wittkampf FHM, Hauer RNW, Robles de Medina EO. Control of radiofrequency lesion size by power regulation. Circulation 1989;80: Bardy GH, Sawyer PL, Johnson GW, Reichenbach DD. Radio-frequency ablation: effect of voltage and pulse duration on canine myocardium. Am J Physiol 1990;258:H Haines DE, Watson DD. Tissue heating during radiofrequency catheter ablation: a thermodynamic model and observations in isolated perfused and superfused canine right ventricular free wall. Pacing Clin Electrophysiol 1989;12: Thomas SP, Nicholson IA, Nunn GR, et al. Effect of atrial radiofrequency ablation designed to cure atrial fibrillation on atrial mechanical function. J Cardiovasc Electrophysiol 2000;11: Jensen DN, Wagner BK, Rose AG, Mehra R, Edwards JE. Human atrial dimensions relevant to intracardiac, ablation maze procedures for atrial fibrillation [Abstract]. Pacing Clin Electrophysiol 1997;20: Ho SY, Sanchez-Quintana D, Cabrera JA, Anderson RH. Anatomy of the left atrium: implications for radiofrequency ablation of atrial fibrillation. J Cardiovasc Electrophysiol 1999;10: Thomas SP, Wallace EM, Ross DL. The effect of a residual isthmus of surviving tissue on conduction after linear ablation in atrial myocardium. J Interv Card Electrophysiol 2000;4: Cabo C, Pertsov AM, Baxter WT, Davidenko JM, Gray RA, Jalife J. Wave-front curvature as a cause of slow conduction and block in isolated cardiac muscle. Circ Res 1994;75: Kottkamp H, Hindricks G, Horst E. Subendocardial and intramural temperature response during radiofrequency catheter ablation in chronic myocardial infarction and normal myocardium. Circulation 1997;95: Nakagawa H, Yamanashi WS, Pitha JV. Comparison of in vivo tissue temperature profile and lesion geometry for radiofrequency ablation with a saline-irrigated electrode versus temperature control in a canine thigh muscle preparation. Circulation 1995;91: Nath S, Whayne JG, Kaul S, Goodman NC, Jayaweera AR, Haines DE. Effects of radiofrequency catheter ablation on regional myocardial blood flow. Possible mechanism for late electrophysiological outcome. Circulation 1994;89: