- Email: [email protected]
Feasibility of catheter ablation of mitral annular flutter in patients with prior mitral valve surgery
Feasibility of catheter ablation of mitral annular flutter in patients with prior mitral valve surgery Stavros Mountantonakis, MD, David S. Frankel, MD, Mathew D. Hutchinson, MD, FHRS, Sanjay Dixit, MD, FHRS, Michael Riley, MD, PhD, David J. Callans, MD, FHRS, Fermin Garcia, MD, David Lin, MD, Wendy Tzou, MD, Rupa Bala, MD, Francis E. Marchlinski, MD, FHRS, Edward P. Gerstenfeld, MD From the Section of Cardiac Electrophysiology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania. BACKGROUND Mitral annular flutter (MAF) may occur after ablation of atrial fibrillation in patients with prior mitral valve (MV) replacement or repair. Percutaneous catheter ablation may be challenging owing to the presence of surgical scar and a prosthetic MV. OBJECTIVE We examined the feasibility of and outcome after mitral isthmus ablation in patients with prior MV surgery. METHODS Twenty-one consecutive patients (18 males, age 61 ⫾10 years) with a history of MV surgery (nine replacement, 12 repair with annuloplasty ring) underwent catheter ablation of clinical (n ⫽ 17) or easily inducible (n ⫽ 4) MAF (group 1). Patients were matched for age, gender, and ejection fraction, with 21 patients undergoing MAF ablation without prior MV surgery (group 2). Irrigated ablation was delivered endocardially in a linear fashion from the MV to the left inferior and/or to the right superior pulmonary vein and, when required, epicardially inside the coronary sinus. Isolation of all pulmonary veins was also performed.
Introduction Preoperative atrial fibrillation (AF) is common in patients undergoing mitral valve (MV) surgery1 and has been associated with increased incidence of stroke and all-cause mortality.2 Surgical ablation of AF at the time of MV surgery has become more popular based on evidence that it is associated with low morbidity and mortality3 and that it decreases the incidence of postoperative thromboembolic events and valve-related complications.4 – 6 The ablation procedure typically consists of epicardial pulmonary vein isolation using radiofrequency energy, amputation of the left atrial appendage (LAA), and linear ablation connecting the isolated pulmonary veins to the mitral annulus. Restoration of sinus rhythm has been achieved in 70%–96% of patients over an average follow-up of 5 years.7,8 However, gaps in the linear ablation lesions may lead to left atrial flutters after surgery. Address reprint requests and correspondence: Edward P. Gerstenfeld, M.D., 9 Founders Pavilion, 3400 Spruce Street, Philadelphia, Pennsylvania 19104. E-mail address: [email protected]. (Received November 1, 2010; accepted January 7, 2011.)
RESULTS There was no difference in termination of tachycardia during ablation (group 1 vs. group 2; 86% vs. 71%; P ⫽ .454), achieving mitral isthmus block (71% vs. 71%; P ⫽ 1.000), or need for epicardial ablation (43% vs. 62%; P ⫽ .354) between groups. No complications occurred in either group. After a mean follow-up of 7 ⫾ 4 months, 15 (71%) patients in group 1 and 14 (67%) in group 2 had no recurrence of atrial arrhythmias. CONCLUSIONS Percutaneous mitral isthmus ablation is feasible and safe in patients with prior MV replacement or repair and has comparable outcomes to patients without prior MV surgery. KEYWORDS Catheter ablation; Mitral annular flutter; Mitral valve surgery; Atrial fibrillation ABBREVIATIONS AF ⫽ atrial fibrillation; CS ⫽ coronary sinus; CT ⫽ computed tomography; INR ⫽ international normalized ratio; LA ⫽ left atrium; LAA ⫽ left atrial appendage; MAF ⫽ mitral annular flutter; MV ⫽ mitral valve (Heart Rhythm 2011;8:809 – 814) © 2011 Heart Rhythm Society. All rights reserved.
Management of patients with postoperative atrial flutter after combined AF and MV surgery is challenging. Patients can be quite symptomatic, and the ventricular rate is often difficult to control. Antiarrhythmic medications have limited efficacy in treating these incessant arrhythmias. Mitral annular flutter (MAF) may occur after surgical ablation, and catheter ablation may be technically challenging because of the presence of surgical scar and a prosthetic valve or annuloplasty ring. The MAF recurrence rate after linear ablation of the mitral isthmus without achieving bidirectional block is quite high.9,10 We examined the feasibility and outcome of mitral isthmus ablation in patients with MAF and a history of MV repair or replacement.
Methods Study population We studied 21 consecutive patients with a history of MV surgery who were referred to our institution for ablation of symptomatic left atrial flutter, alone or in addition to AF, between 2005 and 2010 (group 1). The diagnosis of MAF was confirmed during electrophysiology study with activa-
1547-5271/$ -see front matter © 2011 Heart Rhythm Society. All rights reserved.
doi:10.1016/j.hrthm.2011.01.019
810
Heart Rhythm, Vol 8, No 6, June 2011
Figure 1 Left panel: Recordings from the surface electrocardiogram, circular mapping catheter (Lasso), CS, and ablation catheter (Carto) during entrainment from the ablation catheter positioned at the septal mitral annulus. The postpacing interval (PPI) is similar to the tachycardia cycle length (245 ms), which confirms participation of the septal mitral annulus in the tachycardia circuit. Right panel: Activation map from earliest (pink) to latest (blue) during clockwise mitral flutter projected on CT. Note the image of the mitral prosthesis on the CT. RSPV: right superior pulmonary vein; LAA: left atrial appendage; LSPV: left superior pulmonary vein.
tion and entrainment mapping (Figure 1). These patients were matched for age, gender, left atrial size, and ejection fraction, with 21 patients undergoing MAF ablation without prior MV surgery (group 2).
Electrophysiology study and ablation Cardiac computed tomography (CT) was performed the night before the study in all patients to evaluate the anatomy of the left atrium (LA) and pulmonary veins. Antiarrhythmic agents were discontinued for at least 5 half-lives, except for amiodarone, which was stopped at least 2 weeks before the procedure. Ablation was typically performed on warfarin with a therapeutic international normalized ratio (INR). In cases in which the INR was subtherapeutic at any point during the 3 weeks before ablation, a transesophageal echocardiogram was performed before ablation to exclude LA thrombus. Two decapolar catheters with 5-mm electrodes and 2-mm interelectrode spacing were placed in the posterior right atrium and coronary sinus (CS), with the proximal poles located at the CS ostium. A phased-array diagnostic ultrasound catheter (5.5–10 MHz, 8 Fr, AcuNav, Siemens Medical, Mountain View, CA) was advanced to the level of the fossa ovalis in the right atrium. In cases in which atrial flutter was the presenting rhythm, overdrive atrial pacing was initially performed from the proximal and distal CS to determine their location relative to the tachycardia circuit and to exclude right atrial flutter. After excluding right atrial flutter, heparin was initiated to achieve an activated clotting time of ⬎350 seconds, and this was maintained throughout
LA access. The LA was accessed with two transseptal punctures; a 10-pole, 15- to 25-mm adjustable circular mapping catheter (Lasso, Biosense-Webster, Inc., Diamond Bar, CA) and 3.5-mm irrigated-tip ablation catheter (Thermocool, Biosense Webster) were used for mapping and radiofrequency ablation, respectively. Three-dimensional electroanatomic mapping was performed with the Carto (Biosense-Webster) or NavX (St. Jude Medical, Minneapolis, MN) system. Activation and entrainment mapping from at least three sites around the mitral annulus were performed to confirm the diagnosis of MAF. Ablation lesions were delivered endocardially in a linear fashion from the MV annulus to the left inferior pulmonary vein in most cases. On the basis of operator’s preference, a line to the right superior pulmonary vein was occasionally created initially or if the mitral isthmus line failed to achieve block. Epicardial ablation inside the CS was not performed as part of the line from the MV to the right superior pulmonary vein. Radiofrequency lesions were delivered during ongoing tachycardia at a power of 35–50 W, with temperature limited to 42°C. Beginning in 2008, a steerable sheath (Agilis, St. Jude Medical) was used to facilitate manipulation of the ablation catheter near the MV. If tachycardia persisted despite anatomic completion of the linear ablation between the MV and left inferior pulmonary vein, irrigated epicardial ablation (20 –30 W) inside the CS was performed opposite the endocardial line (Figure 2). If tachycardia persisted after epicardial ablation, the patient was electrically cardioverted to
Mountantonakis et al
Mitral Flutter after MVR
811
Figure 2 A: Endocardial ablation lesions (blue) between left inferior pulmonary vein and mitral annulus in a patient with a prosthetic MV and MAF. The CS is shown in green, and the LAA in red. White tags represent ablation lesions placed epicardially within the CS. The red tag indicates the site where mitral isthmus block was achieved. B: Ablation lesions (red) between the right superior pulmonary vein and mitral annulus in another patient with a prosthetic MV and mitral flutter. LCPV: left common pulmonary vein; RSPV: right superior pulmonary vein; LAA: left atrial appendage; RIPV: right inferior pulmonary vein.
sinus rhythm and ablation continued, reinforcing the previous line or creating a new line to another anatomical barrier until block was achieved or the operator ended the procedure. The presence of block across the mitral line was confirmed by the characteristic reversal of CS activation (proximal-to-distal) while pacing from the LAA with the distal poles of the CS catheter positioned just medial to the line (Figure 3).
In cases in which the LAA was ligated or excised, the ablation catheter was placed at the LAA stump and the same maneuver was performed. For patients whose mitral isthmus line was created to the right superior pulmonary vein, block was confirmed by documenting the latest activation time inferior to the line while pacing from the LAA and progressively earlier activation times and moving the ablation catheter inferiorly. In each case, isolation of all pulmo-
Figure 3 Left panels: Fluoroscopic views in the right anterior oblique (RAO) and left anterior oblique (LAO) projections demonstrating the circular mapping catheter (C Map) inside the left inferior pulmonary vein, the CS catheter, and the approximate location of the mitral isthmus line (red line). The MV prosthesis (MV) can be easily visualized on fluoroscopy. In this patient, a prosthetic aortic valve (AV) was also present. Middle panel: After termination of mitral flutter, pacing from the mapping catheter (Map D, catheter not shown) in the LAA demonstrates persistent mitral isthmus conduction with early activation of the distal CS electrogram. Right panel: Additional lesions placed epicardially inside the CS were required to achieve mitral isthmus block with reversal of CS activation. In this case, pacing is performed from the circular mapping catheter in the LAA. The mapping (Map D) catheter has been placed in the LA just inferior to the mitral line.
812 Table 1
Heart Rhythm, Vol 8, No 6, June 2011 Baseline characteristics
Characteristic
Group 1: history of MV surgery (n ⫽ 21)
Age Male gender, % Left ventricular ejection fraction, % LA size, cm No. of prior ablations (including MAZE)
61.7 85.7 53 5.2 1.3
⫾ 10.4
62.3 85.7 56 4.9 2.1
⫾ 11 ⫾ 1.1 ⫾ 1.1
nary veins with confirmation of entrance and exit block using the circular mapping catheter was also performed. For patients who presented with AF but had electrocardiogram evidence of LA flutter, we first performed isolation of all pulmonary veins; if sinus rhythm was not restored, then the patient was electrically cardioverted. Isoproterenol was then infused starting at 3 g/min and increased to a target of 20 g/min unless limited by nausea or hypotension. If no arrhythmia was induced, atrial programmed stimulation was performed in an attempt to induce MAF. In all group 1 patients, MAF either occurred spontaneously or was inducible atrial programmed stimulation with or without isoproterenol infusion. The primary ablation endpoint was the presence of conduction block across the ablated line. Secondary endpoints were termination of tachycardia during ablation and inability to reinduce MAF after creation of the mitral isthmus line.
Follow-up All patients underwent transthoracic echocardiography on the day after the procedure and telemetric monitoring for at least 24 hours postablation. All patients were discharged with a transtelephonic monitor for 4 weeks with instructions to transmit telemetric strips twice daily and with the occurrence of any symptoms. Since 2008, patients were given transtelephonic monitors with “autotrigger” capabilities to detect asymptomatic arrhythmias. All patients underwent follow-up clinic visits at 6 weeks, 6 months, and 1 year after ablation. Most patients were evaluated annually thereafter. Those patients not seen in our clinic were contacted by telephone to ascertain clinical status, and referring physician records were obtained to document arrhythmia recurrence.
Long-term endpoints The primary long-term endpoint was absence of MAF after a 2-month blanking period after ablation. The secondary endpoint was the absence of any atrial arrhythmias, including AF or atrial flutter.
Statistics Continuous variables are reported as mean ⫾ 1 standard deviation, and categorical variables are reported as proportions. Comparisons of continuous variables were performed using the Student’s t-test, and categorical variables using the Fisher’s exact test.
Group 2: controls (n ⫽ 21) ⫾ 10.7 ⫾ 10 ⫾ 0.8 ⫾ 1.2
P .861 1 .365 .216 .031
Results From January 2005 to January 2010, 67 patients with prior MV surgery were referred for catheter ablation. Forty-one patients had only AF, and 26 had a history of AF and/or atrial flutter. MAF was confirmed at electrophysiology study in 21 patients (31%).
Baseline characteristics Group 1 consisted of 21 patients (18 males) with a mean age of 61.7 ⫾ 10.4 years; nine had a history of MV replacement (St. Jude n ⫽ 7, bioprosthetic n ⫽ 2), and 12 had a history of MV repair with annuloplasty ring. MV surgery preceded ablation by a mean of 5.8 ⫾ 5.7 years (interquartile range 2.0 –9.5 years). Twelve patients underwent concomitant surgical AF ablation (LA MAZE), and 10 had undergone at least one catheter-based AF ablation after surgery. Patients in group 2 were matched for age, gender, left ventricular function, and LA size (Table 1). Patients in group 2 had more prior AF ablations than patients in group 1 (2.1 ⫾ 1.2 vs. 1.3 ⫾ 1.1 per patient; P ⫽ .031). A prior mitral annular “line,” either as part of surgical AF ablation (n ⫽ 5) or catheter ablation (n ⫽ 16), was performed in 14/21 group 1 patients and 7/21 group 2 patients (P ⫽ .063). Conduction block across the mitral line was not confirmed in any patient. There was a trend for the mitral flutter cycle length to be longer in group 1 compared with in group 2 patients (278 ⫾ 55 vs. 245 ⫾ 48 ms; P ⫽ .064).
Procedural characteristics The presenting rhythm was MAF in most patients (17 in group 1 and 13 in group 2). The remaining patients presented in AF but had either electrocardiogram evidence11 or easily inducible MAF after isolation of the pulmonary veins and restoration of sinus rhythm. In all cases, MAF was confirmed with entrainment maneuvers during the ablation procedure. A steerable sheath (Agilis, St. Jude Medical) was used in most patients (17 in group 1 and 15 in group 2) to assist with mapping and ablation. The mitral isthmus line was created from the MV annulus to the left inferior pulmonary vein in 34 patients (group 1 vs. group 2, 18 vs. 16 patients). In 30 of those patients (17 vs. 13; P ⫽ .229), flutter terminated during creation of the mitral isthmus line. Mitral isthmus block was achieved in 25 (14 vs. 13; P ⫽ .885), and epicardial CS lesions were required to achieve block in 18 (8 vs. 10; P ⫽ .229) patients. In eight patients (three in group 1 vs. five group 2), a line was ablated from the MV to the septal aspect of the right pulmonary vein: in
Mountantonakis et al Table 2
Mitral Flutter after MVR
813
Procedural characteristics
Procedural characteristics
Group 1: history of MV surgery (n ⫽ 21)
Group 2: controls (n ⫽ 21)
P
Mitral flutter cycle length, ms No. of mitral isthmus ablation lesions Ablation time, minutes Patients requiring epicardial (CS) ablationa (%) Tachycardia termination during ablation Mitral isthmus block
278 ⫾ 54 61 ⫾ 43 76 ⫾ 50 9 (43) 18 (86) 15 (71)
245 ⫾ 48 35 ⫾ 25 43 ⫾ 30 13 (68) 15 (71) 15 (71)
.064 .105 .087 .225 .454 1.000
a
Excluding two patients from group 2 who only had line to RSPV.
six after an MA to left inferior pulmonary vein line did not achieve block and in two as the initial ablation line. In the six patients without block after the MA to left inferior pulmonary vein line, flutter termination and block were achieved in two patients and neither flutter termination nor block in four patients. In the two patients in whom the septal line was the first line placed at the operator’s discretion, one of two patients had termination of flutter with documented block across the line. Overall, MAF terminated during endocardial mitral isthmus ablation in 18 patients in group 1 and 15 patients in group 2 (P ⫽ .454), and confirmed block across the line was achieved in 15 (71%) patients in group 1 and 15 (71%) patients in group 2 (P ⫽ 1.000). There was no difference in procedural outcomes between patients with MV repair versus replacement. There was a nonsignificant trend toward a longer ablation time in group 1 (76 ⫾ 50 vs. 43 ⫾ 30 minutes; P ⫽ .087). No acute or late complications were observed in either group. Specifically, there was no incidence of catheter entrapment in mitral prosthesis and no evidence of MV regurgitation or other dysfunction of mitral prosthesis as evaluated by intracardiac echocardiography. Procedural characteristics are shown in Table 2.
Follow-up After a mean follow-up of 6.1 ⫾ 3.5 months, 18 (86%) patients in group 1 and 18 (86%) patients in group 2 had no recurrence of MAF (Table 3). MAF recurred in six patients; medical management was unsuccessful in restoring and maintaining sinus rhythm. Notably, in all six patients MAF occurred during the blanking period and persisted into the monitoring period. In addition, mitral isthmus block was not achieved in any of these six patients, despite termination of tachycardia during ablation. Three patients from group 1 and four from group 2 experienced AF recurrence at the time of follow-up without any evidence of MAF on tranTable 3
stelephonic monitoring. Overall, 15 (71%) patients from group 1 and 14 (67%) patients from group 2 maintained sinus rhythm without evidence of arrhythmia recurrence. No difference in long-term outcome was found between the patients with MV replacement versus repair. There was no significant difference in the number, class, or dose of antiarrhythmic agents before or after ablation (Table 3).
Discussion We present the largest series of patients with prior MV surgery undergoing catheter ablation for MAF. MAF was present in 31% of patients referred for catheter ablation of atrial arrhythmias occurring after MV surgery. This high incidence may be explained by the presence of perimitral scar (either as a result of MV surgery or prior ablation) that facilitates slow conduction around the mitral annulus. Despite the presence of prior MV repair or replacement with a prosthetic valve, there was no difference in need for epicardial ablation or the ability to achieve conduction block across the mitral isthmus compared with in patients without prior MV surgery. Freedom from mitral flutter (86%) and any atrial arrhythmias (71%) during 6-month follow-up was excellent. Several factors make catheter ablation technically challenging in this population. First, it is generally believed that to achieve block across the mitral isthmus, ablation is required on the mitral annulus and occasionally even on the ventricular aspect of the annulus.12 Therefore, creation of a mitral isthmus line is often avoided in patients with mitral prostheses for fear of catheter entrapment in the prosthetic valve. Second, the presence of advanced atrial myopathy and scar in the LA, particularly around the mitral annulus, may create challenges in creating deep, contiguous lesions. Nevertheless, the presence of MAF is usually highly symptomatic and medical management is often unsuccessful.13 Therefore, catheter ablation remains the treatment of choice.
Follow-up outcomes
Follow-up data
Group 1: history of MV Surgery (n ⫽ 21)
Group 2: controls (n ⫽ 21)
P
Follow-up time, months Freedom from atrial flutter (%) Freedom from AF and flutter (%) Antiarrhythmics at follow-up (%)
7.1 ⫾ 5.0 18 (86) 15 (71) 12 (57)
6.9 ⫾ 3.9 18 (86) 14 (67) 15 (71)
.891 1.000 1.000 .520
814 In this study, we demonstrate that creation of a mitral isthmus line can be safely and efficiently performed in patients with a history of MV surgery, with comparable results to patients without prior MV surgery. We believe that a few procedural steps addressing the anatomic challenges in patients with MV surgery can be helpful in avoiding complications. For our transseptal punctures, we typically performed a posterior puncture and always advanced a catheter through the first sheath into the right superior pulmonary vein to avoid inadvertent movement of the sheath toward the mitral annulus while the second puncture was being performed. Use of a steerable sheath also allowed better control of the catheter when manipulating it near the MV. The circular mapping catheter was typically not used to create the LA geometry, as is our usual practice, to avoid entrapment in the MV. In general, the mechanical mitral prosthesis or ring provided an excellent fluoroscopic landmark to define the plane of mitral annulus and avoid catheter entrapment. Finally, all procedures were performed by experienced operators. Although the use of a circular mapping catheter may theoretically increase the risk of mitral prosthesis entrapment, we believe that it is extremely useful for confirming isolation of pulmonary veins. Rostock et al14 demonstrated that the most common site of mitral isthmus line conduction recovery was at the junction with the left inferior pulmonary vein. We therefore believe that confirmation of isolation of the left pulmonary veins is of paramount importance for achieving enduring block across the mitral isthmus and that achieving isolation of the remaining pulmonary veins is helpful in preventing triggering of MAF in case conduction resumes across the mitral isthmus. Interestingly, most patients in the MV group had undergone a prior attempt at performing a mitral line using either a surgical or catheter-based approach. In none of these patients was there documentation of conduction block across the mitral line. In our study, mitral isthmus block could not be achieved in six patients, despite termination of the arrhythmia. MAF recurred in all of these patients. This emphasizes the importance of achieving mitral isthmus conduction block in patients with MAF.
Prior work Creation of a mitral isthmus line of ablation in patients with MV surgery has been previously reported by Lang et al.15 In that series of 26 patients with MV prosthesis referred for ablation for AF, 24 (92%) had creation of additional lines, including a mitral isthmus line. These lines were created empirically in an effort to reduce the incidence of postablation tachycardias and not to address clinical or easily inducible MAF. In addition, the presence of block across the mitral isthmus line was not documented. The absence of block across the mitral isthmus has been demonstrated to be proarrhythmic in previous studies by facilitating slow conduction and reentry.9,10 Chae et al9 reported that the critical isthmus in the vast majority of postablation macroreentrant atrial tachycardia involves a
Heart Rhythm, Vol 8, No 6, June 2011 gap in a prior ablation line. Similarly, Matsuo et al10 showed that MAF occurs more commonly in patients with attempted isthmus ablation in the past. In a large series of patients undergoing mitral isthmus ablation in the presence of MV prosthesis, the postablation atrial tachycardia incidence was 23%,15 and all patients who had recurrence of MAF had received ablation across the mitral isthmus.
Limitations The sample size of our study includes the largest number of patients with prior MV surgery undergoing AF ablation that has been reported; however, these patients are still referred for ablation infrequently. Therefore, the study may be underpowered to detect some differences between groups. However, any difference between the major clinical outcomes of the study (freedom from atrial flutter 86% vs. 86%; freedom from AF and flutter 71 vs. 67%) were small and unlikely to be affected by a larger sample size.
Conclusion Mitral isthmus ablation is feasible and safe in patients with prior surgical MV replacement or repair. The short- and long-term freedom from atrial arrhythmias is comparable to that of patients without prior MV surgery.
References 1. Ad N, Cox JL. The significance of atrial fibrillation ablation in patients undergoing mitral valve surgery. Semin Thorac Cardiovasc Surg 2002;14:193–197. 2. Ngaage DL, Schaff HV, Mullany CJ, et al. Influence of preoperative atrial fibrillation on late results of mitral repair: is concomitant ablation justified? Ann Thorac Surg 2007;84:434, 442; discussion 442– 443. 3. Fujita T, Kobayashi J, Toda K, et al. Long-term outcome of combined valve repair and maze procedure for nonrheumatic mitral regurgitation. J Thorac Cardiovasc Surg 2010;140:1332–1337. 4. Bando K, Kobayashi J, Kosakai Y, et al. Impact of cox maze procedure on outcome in patients with atrial fibrillation and mitral valve disease. J Thorac Cardiovasc Surg 2002;124:575–583. 5. Stulak JM, Schaff HV, Dearani JA, Orszulak TA, Daly RC, Sundt TM, 3rd.Restoration of sinus rhythm by the maze procedure halts progression of tricuspid regurgitation after mitral surgery. Ann Thorac Surg 2008;86:40, 44; discussion 44 – 45. 6. Fukunaga S, Hori H, Ueda T, Takagi K, Tayama E, Aoyagi S. Effect of surgery for atrial fibrillation associated with mitral valve disease. Ann Thorac Surg 2008;86:1212–1217. 7. Gillinov AM, Sirak J, Blackstone EH, et al. The Cox maze procedure in mitral valve disease: predictors of recurrent atrial fibrillation. J Thorac Cardiovasc Surg 2005;130:1653–1660. 8. Gaynor SL, Schuessler RB, Bailey MS, et al. Surgical treatment of atrial fibrillation: predictors of late recurrence. J Thorac Cardiovasc Surg 2005;129:104–111. 9. Chae S, Oral H, Good E, et al. Atrial tachycardia after circumferential pulmonary vein ablation of atrial fibrillation: mechanistic insights, results of catheter ablation, and risk factors for recurrence. J Am Coll Cardiol 2007;50:1781–1787. 10. Matsuo S, Wright M, Knecht S, et al. Peri-mitral atrial flutter in patients with atrial fibrillation ablation. Heart Rhythm 2010;7:2– 8. 11. Gerstenfeld EP, Dixit S, Bala R, et al. Surface electrocardiogram characteristics of atrial tachycardias occurring after pulmonary vein isolation. Heart Rhythm 2007;4:1136 –1143 12. Jais P, Hocini M, Hsu LF, et al. Technique and results of linear ablation at the mitral isthmus. Circulation 2004;110:2996 –3002. 13. Gerstenfeld EP, Marchlinski FE. Mapping and ablation of left atrial tachycardias occurring after atrial fibrillation ablation. Heart Rhythm 2007;4(3 Suppl):S65–72. 14. Rostock T, O’Neill MD, Sanders P, et al. Characterization of conduction recovery across left atrial linear lesions in patients with paroxysmal and persistent atrial fibrillation. J Cardiovasc Electrophysiol 2006;17:1106 –1111. 15. Lang CC, Santinelli V, Augello G, et al. Transcatheter radiofrequency ablation of atrial fibrillation in patients with mitral valve prostheses and enlarged atria: safety, feasibility, and efficacy. J Am Coll Cardiol 2005;45:868 – 872.
Introduction Preoperative atrial fibrillation (AF) is common in patients undergoing mitral valve (MV) surgery1 and has been associated with increased incidence of stroke and all-cause mortality.2 Surgical ablation of AF at the time of MV surgery has become more popular based on evidence that it is associated with low morbidity and mortality3 and that it decreases the incidence of postoperative thromboembolic events and valve-related complications.4 – 6 The ablation procedure typically consists of epicardial pulmonary vein isolation using radiofrequency energy, amputation of the left atrial appendage (LAA), and linear ablation connecting the isolated pulmonary veins to the mitral annulus. Restoration of sinus rhythm has been achieved in 70%–96% of patients over an average follow-up of 5 years.7,8 However, gaps in the linear ablation lesions may lead to left atrial flutters after surgery. Address reprint requests and correspondence: Edward P. Gerstenfeld, M.D., 9 Founders Pavilion, 3400 Spruce Street, Philadelphia, Pennsylvania 19104. E-mail address: [email protected]. (Received November 1, 2010; accepted January 7, 2011.)
RESULTS There was no difference in termination of tachycardia during ablation (group 1 vs. group 2; 86% vs. 71%; P ⫽ .454), achieving mitral isthmus block (71% vs. 71%; P ⫽ 1.000), or need for epicardial ablation (43% vs. 62%; P ⫽ .354) between groups. No complications occurred in either group. After a mean follow-up of 7 ⫾ 4 months, 15 (71%) patients in group 1 and 14 (67%) in group 2 had no recurrence of atrial arrhythmias. CONCLUSIONS Percutaneous mitral isthmus ablation is feasible and safe in patients with prior MV replacement or repair and has comparable outcomes to patients without prior MV surgery. KEYWORDS Catheter ablation; Mitral annular flutter; Mitral valve surgery; Atrial fibrillation ABBREVIATIONS AF ⫽ atrial fibrillation; CS ⫽ coronary sinus; CT ⫽ computed tomography; INR ⫽ international normalized ratio; LA ⫽ left atrium; LAA ⫽ left atrial appendage; MAF ⫽ mitral annular flutter; MV ⫽ mitral valve (Heart Rhythm 2011;8:809 – 814) © 2011 Heart Rhythm Society. All rights reserved.
Management of patients with postoperative atrial flutter after combined AF and MV surgery is challenging. Patients can be quite symptomatic, and the ventricular rate is often difficult to control. Antiarrhythmic medications have limited efficacy in treating these incessant arrhythmias. Mitral annular flutter (MAF) may occur after surgical ablation, and catheter ablation may be technically challenging because of the presence of surgical scar and a prosthetic valve or annuloplasty ring. The MAF recurrence rate after linear ablation of the mitral isthmus without achieving bidirectional block is quite high.9,10 We examined the feasibility and outcome of mitral isthmus ablation in patients with MAF and a history of MV repair or replacement.
Methods Study population We studied 21 consecutive patients with a history of MV surgery who were referred to our institution for ablation of symptomatic left atrial flutter, alone or in addition to AF, between 2005 and 2010 (group 1). The diagnosis of MAF was confirmed during electrophysiology study with activa-
1547-5271/$ -see front matter © 2011 Heart Rhythm Society. All rights reserved.
doi:10.1016/j.hrthm.2011.01.019
810
Heart Rhythm, Vol 8, No 6, June 2011
Figure 1 Left panel: Recordings from the surface electrocardiogram, circular mapping catheter (Lasso), CS, and ablation catheter (Carto) during entrainment from the ablation catheter positioned at the septal mitral annulus. The postpacing interval (PPI) is similar to the tachycardia cycle length (245 ms), which confirms participation of the septal mitral annulus in the tachycardia circuit. Right panel: Activation map from earliest (pink) to latest (blue) during clockwise mitral flutter projected on CT. Note the image of the mitral prosthesis on the CT. RSPV: right superior pulmonary vein; LAA: left atrial appendage; LSPV: left superior pulmonary vein.
tion and entrainment mapping (Figure 1). These patients were matched for age, gender, left atrial size, and ejection fraction, with 21 patients undergoing MAF ablation without prior MV surgery (group 2).
Electrophysiology study and ablation Cardiac computed tomography (CT) was performed the night before the study in all patients to evaluate the anatomy of the left atrium (LA) and pulmonary veins. Antiarrhythmic agents were discontinued for at least 5 half-lives, except for amiodarone, which was stopped at least 2 weeks before the procedure. Ablation was typically performed on warfarin with a therapeutic international normalized ratio (INR). In cases in which the INR was subtherapeutic at any point during the 3 weeks before ablation, a transesophageal echocardiogram was performed before ablation to exclude LA thrombus. Two decapolar catheters with 5-mm electrodes and 2-mm interelectrode spacing were placed in the posterior right atrium and coronary sinus (CS), with the proximal poles located at the CS ostium. A phased-array diagnostic ultrasound catheter (5.5–10 MHz, 8 Fr, AcuNav, Siemens Medical, Mountain View, CA) was advanced to the level of the fossa ovalis in the right atrium. In cases in which atrial flutter was the presenting rhythm, overdrive atrial pacing was initially performed from the proximal and distal CS to determine their location relative to the tachycardia circuit and to exclude right atrial flutter. After excluding right atrial flutter, heparin was initiated to achieve an activated clotting time of ⬎350 seconds, and this was maintained throughout
LA access. The LA was accessed with two transseptal punctures; a 10-pole, 15- to 25-mm adjustable circular mapping catheter (Lasso, Biosense-Webster, Inc., Diamond Bar, CA) and 3.5-mm irrigated-tip ablation catheter (Thermocool, Biosense Webster) were used for mapping and radiofrequency ablation, respectively. Three-dimensional electroanatomic mapping was performed with the Carto (Biosense-Webster) or NavX (St. Jude Medical, Minneapolis, MN) system. Activation and entrainment mapping from at least three sites around the mitral annulus were performed to confirm the diagnosis of MAF. Ablation lesions were delivered endocardially in a linear fashion from the MV annulus to the left inferior pulmonary vein in most cases. On the basis of operator’s preference, a line to the right superior pulmonary vein was occasionally created initially or if the mitral isthmus line failed to achieve block. Epicardial ablation inside the CS was not performed as part of the line from the MV to the right superior pulmonary vein. Radiofrequency lesions were delivered during ongoing tachycardia at a power of 35–50 W, with temperature limited to 42°C. Beginning in 2008, a steerable sheath (Agilis, St. Jude Medical) was used to facilitate manipulation of the ablation catheter near the MV. If tachycardia persisted despite anatomic completion of the linear ablation between the MV and left inferior pulmonary vein, irrigated epicardial ablation (20 –30 W) inside the CS was performed opposite the endocardial line (Figure 2). If tachycardia persisted after epicardial ablation, the patient was electrically cardioverted to
Mountantonakis et al
Mitral Flutter after MVR
811
Figure 2 A: Endocardial ablation lesions (blue) between left inferior pulmonary vein and mitral annulus in a patient with a prosthetic MV and MAF. The CS is shown in green, and the LAA in red. White tags represent ablation lesions placed epicardially within the CS. The red tag indicates the site where mitral isthmus block was achieved. B: Ablation lesions (red) between the right superior pulmonary vein and mitral annulus in another patient with a prosthetic MV and mitral flutter. LCPV: left common pulmonary vein; RSPV: right superior pulmonary vein; LAA: left atrial appendage; RIPV: right inferior pulmonary vein.
sinus rhythm and ablation continued, reinforcing the previous line or creating a new line to another anatomical barrier until block was achieved or the operator ended the procedure. The presence of block across the mitral line was confirmed by the characteristic reversal of CS activation (proximal-to-distal) while pacing from the LAA with the distal poles of the CS catheter positioned just medial to the line (Figure 3).
In cases in which the LAA was ligated or excised, the ablation catheter was placed at the LAA stump and the same maneuver was performed. For patients whose mitral isthmus line was created to the right superior pulmonary vein, block was confirmed by documenting the latest activation time inferior to the line while pacing from the LAA and progressively earlier activation times and moving the ablation catheter inferiorly. In each case, isolation of all pulmo-
Figure 3 Left panels: Fluoroscopic views in the right anterior oblique (RAO) and left anterior oblique (LAO) projections demonstrating the circular mapping catheter (C Map) inside the left inferior pulmonary vein, the CS catheter, and the approximate location of the mitral isthmus line (red line). The MV prosthesis (MV) can be easily visualized on fluoroscopy. In this patient, a prosthetic aortic valve (AV) was also present. Middle panel: After termination of mitral flutter, pacing from the mapping catheter (Map D, catheter not shown) in the LAA demonstrates persistent mitral isthmus conduction with early activation of the distal CS electrogram. Right panel: Additional lesions placed epicardially inside the CS were required to achieve mitral isthmus block with reversal of CS activation. In this case, pacing is performed from the circular mapping catheter in the LAA. The mapping (Map D) catheter has been placed in the LA just inferior to the mitral line.
812 Table 1
Heart Rhythm, Vol 8, No 6, June 2011 Baseline characteristics
Characteristic
Group 1: history of MV surgery (n ⫽ 21)
Age Male gender, % Left ventricular ejection fraction, % LA size, cm No. of prior ablations (including MAZE)
61.7 85.7 53 5.2 1.3
⫾ 10.4
62.3 85.7 56 4.9 2.1
⫾ 11 ⫾ 1.1 ⫾ 1.1
nary veins with confirmation of entrance and exit block using the circular mapping catheter was also performed. For patients who presented with AF but had electrocardiogram evidence of LA flutter, we first performed isolation of all pulmonary veins; if sinus rhythm was not restored, then the patient was electrically cardioverted. Isoproterenol was then infused starting at 3 g/min and increased to a target of 20 g/min unless limited by nausea or hypotension. If no arrhythmia was induced, atrial programmed stimulation was performed in an attempt to induce MAF. In all group 1 patients, MAF either occurred spontaneously or was inducible atrial programmed stimulation with or without isoproterenol infusion. The primary ablation endpoint was the presence of conduction block across the ablated line. Secondary endpoints were termination of tachycardia during ablation and inability to reinduce MAF after creation of the mitral isthmus line.
Follow-up All patients underwent transthoracic echocardiography on the day after the procedure and telemetric monitoring for at least 24 hours postablation. All patients were discharged with a transtelephonic monitor for 4 weeks with instructions to transmit telemetric strips twice daily and with the occurrence of any symptoms. Since 2008, patients were given transtelephonic monitors with “autotrigger” capabilities to detect asymptomatic arrhythmias. All patients underwent follow-up clinic visits at 6 weeks, 6 months, and 1 year after ablation. Most patients were evaluated annually thereafter. Those patients not seen in our clinic were contacted by telephone to ascertain clinical status, and referring physician records were obtained to document arrhythmia recurrence.
Long-term endpoints The primary long-term endpoint was absence of MAF after a 2-month blanking period after ablation. The secondary endpoint was the absence of any atrial arrhythmias, including AF or atrial flutter.
Statistics Continuous variables are reported as mean ⫾ 1 standard deviation, and categorical variables are reported as proportions. Comparisons of continuous variables were performed using the Student’s t-test, and categorical variables using the Fisher’s exact test.
Group 2: controls (n ⫽ 21) ⫾ 10.7 ⫾ 10 ⫾ 0.8 ⫾ 1.2
P .861 1 .365 .216 .031
Results From January 2005 to January 2010, 67 patients with prior MV surgery were referred for catheter ablation. Forty-one patients had only AF, and 26 had a history of AF and/or atrial flutter. MAF was confirmed at electrophysiology study in 21 patients (31%).
Baseline characteristics Group 1 consisted of 21 patients (18 males) with a mean age of 61.7 ⫾ 10.4 years; nine had a history of MV replacement (St. Jude n ⫽ 7, bioprosthetic n ⫽ 2), and 12 had a history of MV repair with annuloplasty ring. MV surgery preceded ablation by a mean of 5.8 ⫾ 5.7 years (interquartile range 2.0 –9.5 years). Twelve patients underwent concomitant surgical AF ablation (LA MAZE), and 10 had undergone at least one catheter-based AF ablation after surgery. Patients in group 2 were matched for age, gender, left ventricular function, and LA size (Table 1). Patients in group 2 had more prior AF ablations than patients in group 1 (2.1 ⫾ 1.2 vs. 1.3 ⫾ 1.1 per patient; P ⫽ .031). A prior mitral annular “line,” either as part of surgical AF ablation (n ⫽ 5) or catheter ablation (n ⫽ 16), was performed in 14/21 group 1 patients and 7/21 group 2 patients (P ⫽ .063). Conduction block across the mitral line was not confirmed in any patient. There was a trend for the mitral flutter cycle length to be longer in group 1 compared with in group 2 patients (278 ⫾ 55 vs. 245 ⫾ 48 ms; P ⫽ .064).
Procedural characteristics The presenting rhythm was MAF in most patients (17 in group 1 and 13 in group 2). The remaining patients presented in AF but had either electrocardiogram evidence11 or easily inducible MAF after isolation of the pulmonary veins and restoration of sinus rhythm. In all cases, MAF was confirmed with entrainment maneuvers during the ablation procedure. A steerable sheath (Agilis, St. Jude Medical) was used in most patients (17 in group 1 and 15 in group 2) to assist with mapping and ablation. The mitral isthmus line was created from the MV annulus to the left inferior pulmonary vein in 34 patients (group 1 vs. group 2, 18 vs. 16 patients). In 30 of those patients (17 vs. 13; P ⫽ .229), flutter terminated during creation of the mitral isthmus line. Mitral isthmus block was achieved in 25 (14 vs. 13; P ⫽ .885), and epicardial CS lesions were required to achieve block in 18 (8 vs. 10; P ⫽ .229) patients. In eight patients (three in group 1 vs. five group 2), a line was ablated from the MV to the septal aspect of the right pulmonary vein: in
Mountantonakis et al Table 2
Mitral Flutter after MVR
813
Procedural characteristics
Procedural characteristics
Group 1: history of MV surgery (n ⫽ 21)
Group 2: controls (n ⫽ 21)
P
Mitral flutter cycle length, ms No. of mitral isthmus ablation lesions Ablation time, minutes Patients requiring epicardial (CS) ablationa (%) Tachycardia termination during ablation Mitral isthmus block
278 ⫾ 54 61 ⫾ 43 76 ⫾ 50 9 (43) 18 (86) 15 (71)
245 ⫾ 48 35 ⫾ 25 43 ⫾ 30 13 (68) 15 (71) 15 (71)
.064 .105 .087 .225 .454 1.000
a
Excluding two patients from group 2 who only had line to RSPV.
six after an MA to left inferior pulmonary vein line did not achieve block and in two as the initial ablation line. In the six patients without block after the MA to left inferior pulmonary vein line, flutter termination and block were achieved in two patients and neither flutter termination nor block in four patients. In the two patients in whom the septal line was the first line placed at the operator’s discretion, one of two patients had termination of flutter with documented block across the line. Overall, MAF terminated during endocardial mitral isthmus ablation in 18 patients in group 1 and 15 patients in group 2 (P ⫽ .454), and confirmed block across the line was achieved in 15 (71%) patients in group 1 and 15 (71%) patients in group 2 (P ⫽ 1.000). There was no difference in procedural outcomes between patients with MV repair versus replacement. There was a nonsignificant trend toward a longer ablation time in group 1 (76 ⫾ 50 vs. 43 ⫾ 30 minutes; P ⫽ .087). No acute or late complications were observed in either group. Specifically, there was no incidence of catheter entrapment in mitral prosthesis and no evidence of MV regurgitation or other dysfunction of mitral prosthesis as evaluated by intracardiac echocardiography. Procedural characteristics are shown in Table 2.
Follow-up After a mean follow-up of 6.1 ⫾ 3.5 months, 18 (86%) patients in group 1 and 18 (86%) patients in group 2 had no recurrence of MAF (Table 3). MAF recurred in six patients; medical management was unsuccessful in restoring and maintaining sinus rhythm. Notably, in all six patients MAF occurred during the blanking period and persisted into the monitoring period. In addition, mitral isthmus block was not achieved in any of these six patients, despite termination of tachycardia during ablation. Three patients from group 1 and four from group 2 experienced AF recurrence at the time of follow-up without any evidence of MAF on tranTable 3
stelephonic monitoring. Overall, 15 (71%) patients from group 1 and 14 (67%) patients from group 2 maintained sinus rhythm without evidence of arrhythmia recurrence. No difference in long-term outcome was found between the patients with MV replacement versus repair. There was no significant difference in the number, class, or dose of antiarrhythmic agents before or after ablation (Table 3).
Discussion We present the largest series of patients with prior MV surgery undergoing catheter ablation for MAF. MAF was present in 31% of patients referred for catheter ablation of atrial arrhythmias occurring after MV surgery. This high incidence may be explained by the presence of perimitral scar (either as a result of MV surgery or prior ablation) that facilitates slow conduction around the mitral annulus. Despite the presence of prior MV repair or replacement with a prosthetic valve, there was no difference in need for epicardial ablation or the ability to achieve conduction block across the mitral isthmus compared with in patients without prior MV surgery. Freedom from mitral flutter (86%) and any atrial arrhythmias (71%) during 6-month follow-up was excellent. Several factors make catheter ablation technically challenging in this population. First, it is generally believed that to achieve block across the mitral isthmus, ablation is required on the mitral annulus and occasionally even on the ventricular aspect of the annulus.12 Therefore, creation of a mitral isthmus line is often avoided in patients with mitral prostheses for fear of catheter entrapment in the prosthetic valve. Second, the presence of advanced atrial myopathy and scar in the LA, particularly around the mitral annulus, may create challenges in creating deep, contiguous lesions. Nevertheless, the presence of MAF is usually highly symptomatic and medical management is often unsuccessful.13 Therefore, catheter ablation remains the treatment of choice.
Follow-up outcomes
Follow-up data
Group 1: history of MV Surgery (n ⫽ 21)
Group 2: controls (n ⫽ 21)
P
Follow-up time, months Freedom from atrial flutter (%) Freedom from AF and flutter (%) Antiarrhythmics at follow-up (%)
7.1 ⫾ 5.0 18 (86) 15 (71) 12 (57)
6.9 ⫾ 3.9 18 (86) 14 (67) 15 (71)
.891 1.000 1.000 .520
814 In this study, we demonstrate that creation of a mitral isthmus line can be safely and efficiently performed in patients with a history of MV surgery, with comparable results to patients without prior MV surgery. We believe that a few procedural steps addressing the anatomic challenges in patients with MV surgery can be helpful in avoiding complications. For our transseptal punctures, we typically performed a posterior puncture and always advanced a catheter through the first sheath into the right superior pulmonary vein to avoid inadvertent movement of the sheath toward the mitral annulus while the second puncture was being performed. Use of a steerable sheath also allowed better control of the catheter when manipulating it near the MV. The circular mapping catheter was typically not used to create the LA geometry, as is our usual practice, to avoid entrapment in the MV. In general, the mechanical mitral prosthesis or ring provided an excellent fluoroscopic landmark to define the plane of mitral annulus and avoid catheter entrapment. Finally, all procedures were performed by experienced operators. Although the use of a circular mapping catheter may theoretically increase the risk of mitral prosthesis entrapment, we believe that it is extremely useful for confirming isolation of pulmonary veins. Rostock et al14 demonstrated that the most common site of mitral isthmus line conduction recovery was at the junction with the left inferior pulmonary vein. We therefore believe that confirmation of isolation of the left pulmonary veins is of paramount importance for achieving enduring block across the mitral isthmus and that achieving isolation of the remaining pulmonary veins is helpful in preventing triggering of MAF in case conduction resumes across the mitral isthmus. Interestingly, most patients in the MV group had undergone a prior attempt at performing a mitral line using either a surgical or catheter-based approach. In none of these patients was there documentation of conduction block across the mitral line. In our study, mitral isthmus block could not be achieved in six patients, despite termination of the arrhythmia. MAF recurred in all of these patients. This emphasizes the importance of achieving mitral isthmus conduction block in patients with MAF.
Prior work Creation of a mitral isthmus line of ablation in patients with MV surgery has been previously reported by Lang et al.15 In that series of 26 patients with MV prosthesis referred for ablation for AF, 24 (92%) had creation of additional lines, including a mitral isthmus line. These lines were created empirically in an effort to reduce the incidence of postablation tachycardias and not to address clinical or easily inducible MAF. In addition, the presence of block across the mitral isthmus line was not documented. The absence of block across the mitral isthmus has been demonstrated to be proarrhythmic in previous studies by facilitating slow conduction and reentry.9,10 Chae et al9 reported that the critical isthmus in the vast majority of postablation macroreentrant atrial tachycardia involves a
Heart Rhythm, Vol 8, No 6, June 2011 gap in a prior ablation line. Similarly, Matsuo et al10 showed that MAF occurs more commonly in patients with attempted isthmus ablation in the past. In a large series of patients undergoing mitral isthmus ablation in the presence of MV prosthesis, the postablation atrial tachycardia incidence was 23%,15 and all patients who had recurrence of MAF had received ablation across the mitral isthmus.
Limitations The sample size of our study includes the largest number of patients with prior MV surgery undergoing AF ablation that has been reported; however, these patients are still referred for ablation infrequently. Therefore, the study may be underpowered to detect some differences between groups. However, any difference between the major clinical outcomes of the study (freedom from atrial flutter 86% vs. 86%; freedom from AF and flutter 71 vs. 67%) were small and unlikely to be affected by a larger sample size.
Conclusion Mitral isthmus ablation is feasible and safe in patients with prior surgical MV replacement or repair. The short- and long-term freedom from atrial arrhythmias is comparable to that of patients without prior MV surgery.
References 1. Ad N, Cox JL. The significance of atrial fibrillation ablation in patients undergoing mitral valve surgery. Semin Thorac Cardiovasc Surg 2002;14:193–197. 2. Ngaage DL, Schaff HV, Mullany CJ, et al. Influence of preoperative atrial fibrillation on late results of mitral repair: is concomitant ablation justified? Ann Thorac Surg 2007;84:434, 442; discussion 442– 443. 3. Fujita T, Kobayashi J, Toda K, et al. Long-term outcome of combined valve repair and maze procedure for nonrheumatic mitral regurgitation. J Thorac Cardiovasc Surg 2010;140:1332–1337. 4. Bando K, Kobayashi J, Kosakai Y, et al. Impact of cox maze procedure on outcome in patients with atrial fibrillation and mitral valve disease. J Thorac Cardiovasc Surg 2002;124:575–583. 5. Stulak JM, Schaff HV, Dearani JA, Orszulak TA, Daly RC, Sundt TM, 3rd.Restoration of sinus rhythm by the maze procedure halts progression of tricuspid regurgitation after mitral surgery. Ann Thorac Surg 2008;86:40, 44; discussion 44 – 45. 6. Fukunaga S, Hori H, Ueda T, Takagi K, Tayama E, Aoyagi S. Effect of surgery for atrial fibrillation associated with mitral valve disease. Ann Thorac Surg 2008;86:1212–1217. 7. Gillinov AM, Sirak J, Blackstone EH, et al. The Cox maze procedure in mitral valve disease: predictors of recurrent atrial fibrillation. J Thorac Cardiovasc Surg 2005;130:1653–1660. 8. Gaynor SL, Schuessler RB, Bailey MS, et al. Surgical treatment of atrial fibrillation: predictors of late recurrence. J Thorac Cardiovasc Surg 2005;129:104–111. 9. Chae S, Oral H, Good E, et al. Atrial tachycardia after circumferential pulmonary vein ablation of atrial fibrillation: mechanistic insights, results of catheter ablation, and risk factors for recurrence. J Am Coll Cardiol 2007;50:1781–1787. 10. Matsuo S, Wright M, Knecht S, et al. Peri-mitral atrial flutter in patients with atrial fibrillation ablation. Heart Rhythm 2010;7:2– 8. 11. Gerstenfeld EP, Dixit S, Bala R, et al. Surface electrocardiogram characteristics of atrial tachycardias occurring after pulmonary vein isolation. Heart Rhythm 2007;4:1136 –1143 12. Jais P, Hocini M, Hsu LF, et al. Technique and results of linear ablation at the mitral isthmus. Circulation 2004;110:2996 –3002. 13. Gerstenfeld EP, Marchlinski FE. Mapping and ablation of left atrial tachycardias occurring after atrial fibrillation ablation. Heart Rhythm 2007;4(3 Suppl):S65–72. 14. Rostock T, O’Neill MD, Sanders P, et al. Characterization of conduction recovery across left atrial linear lesions in patients with paroxysmal and persistent atrial fibrillation. J Cardiovasc Electrophysiol 2006;17:1106 –1111. 15. Lang CC, Santinelli V, Augello G, et al. Transcatheter radiofrequency ablation of atrial fibrillation in patients with mitral valve prostheses and enlarged atria: safety, feasibility, and efficacy. J Am Coll Cardiol 2005;45:868 – 872.