Predictors of success of the modified maze procedure using radiofrequency device

Introduction

The Cox maze procedure has been considered the gold standard for surgical treatment of atrial fibrillation (AF) since the 1990s. ¹ This procedure involves a complex series of biatrial lesions made using a cut-and-sew technique. Because of the associated morbidity, other modalities, such as cryothermic ablation, microwave ablation, ultrasonic ablation, and radiofrequency (RF) ablation, have been tried as alternatives to the cut-and-sew technique. ^2–4 Recently, clamp-type bipolar RF devices have been introduced. We have been using an irrigated RF system (Cardioblate Surgical Ablation System; Medtronic, Inc., Minneapolis, MN, USA) for the modified Cox maze procedure since March 2005. This system consists of a bipolar clamp device and a monopolar pen device, which can produce most of the lesions of the original Cox maze III. Gaynor and colleagues ⁵ described the use of RF devices to make a similar lesion set to that of the original Cox maze III, which they termed Cox maze IV. Their lesion set initially consisted of pulmonary vein (PV) isolation on both right and left sides separately, and a single connecting inferior linear lesion performed with a bipolar device. However, it was found that adding another connecting superior lesion between the right and left PV isolation lines, to create a box lesion, improved the early results with a lower incidence of recurrent AF and reduced usage of antiarrhythmic drugs. ⁶ Since these data were presented in May 2007, we have added a connecting superior lesion to create a complete box lesion (Figure 1).

OPEN IN VIEWER

Figure 1.

Lesion set of the modified maze procedure. The superior connecting lesion (red dotted line) was added in patients in group 2. RF = radiofrequency.

In addition to isolating the PV from the left atrium, it is also important to interrupt conduction across the posteroinferior portion of the left atrium. ⁷ This can be accomplished by the combination of a transmural lesion on the isthmus portion of the left atrium towards the mitral valve, and a circumferential lesion at the coronary sinus (CS). This CS lesion is usually made by cryoablation, but many surgeons use a bipolar device, and this may not yield completely transmural lesions. Thus, we started using the monopolar pen to ablate the CS separately from the epicardial side (Figure 2). Because the ideal lesion set in the modified maze procedure using alternative energy sources remains unclear, we assessed the importance of the box lesion and CS ablation by comparing patients in whom these lesions were performed with another group without these additional lesions.

OPEN IN VIEWER

Figure 2.

Surgeon’s view of coronary sinus ablation with the monopolar pen device during a mitral valve procedure.

Patients and methods

From March 2005 to July 2009, 61 patients with history of paroxysmal, persistent, or permanent AF underwent the modified maze procedure using RF devices, concomitant with other cardiac operations (Table 1). Exclusion criteria were reoperations, thoracic aortic aneurysm surgery, and a left atrial diameter >70 mm. This study was approved by our institutional review board. Informed consent and permission for the release of medical information were obtained from all patients.

OPEN IN VIEWER

Table 1.

Concomitant cardiac procedures in 61 patients undergoing the maze procedure

Procedures	No. of patients
Mitral repair ± tricuspid annuloplasty	23
Mitral valve replacement ± tricuspid annuloplasty	12
Aortic valve replacement + mitral repair/replacement ± tricuspid annuloplasty	10
Aortic valve replacement ± tricuspid annuloplasty	7
Coronary artery bypass + mitral repair ± tricuspid annuloplasty	5
Coronary artery bypass	3
Patent foramen ovale closure + tricuspid annuloplasty	1

A median sternotomy was performed, and cardiopulmonary bypass was established with ascending aortic and bicaval cannulation. The lesion set on the left atrium consisted of bilateral PV isolation and a connecting lesion at the inferior side of the PV isolation lines through a right-sided left atriotomy incision. In all cases, we resected the left atrial appendage and made a connecting ablation line from the edge of the remnant of the left atrial appendage towards the left PV isolation line. The left atrial isthmus was ablated towards the posterior mitral annulus, using a bipolar device with the aid of a monopolar device from the endocardial side. On the right atrium, after resecting the edge of the right atrial appendage, a transverse incision was made as well as a longitudinal ablation line at the lateral aspect of the right atrium beyond the superior and inferior vena caval junction. The right atrial isthmus was ablated in the direction of the tricuspid annulus, using both bipolar and monopolar devices on the superior and anterior aspects. The atrial septum was also ablated with a bipolar device, just up to the fossa ovalis. The bipolar clamp device was usually applied twice and sometimes 3 times on each lesion, to insure that it was completely transmural. The initial 30 patients (group 1) underwent the modified maze procedure with these lesions only. From May 2007, we added 2 additional lesions: a connecting superior linear lesion between the left and right PV isolation lines, which completed the encirclement of all 4 PV from the left atrium, giving a box lesion (Figure 1); and direct CS ablation from the epicardial side was performed with the monopolar device (Figure 2), to ensure the complete conduction block of the posteroinferior portion of the left atrium. Those 2 minor modifications were applied in 31 patients (group 2).

All patients were given oral beta blockers postoperatively, and anticoagulated with warfarin with a target international normalized ratio of 2.0–3.0 for at least 3 months, unless contraindicated. If the patient remained in sinus rhythm (SR) for 3 months, anticoagulation and antiarrhythmic medications could be discontinued. It is well known that many maze patients develop atrial tachyarrhythmia, including AF, during the postoperative recovery period. ⁸ Our treatment of choice for postoperative AF included the routine use of oral beta blockers, unless contraindicated, and aggressive use of oral or intravenous amiodarone, combined with electrical cardioversion if necessary. Some patients with AF were discharged if they tolerated their AF and had an acceptable heart rate. Such patients received anticoagulation and remained on oral amiodarone for at least 8 weeks after surgery. Cardioversion was considered in an outpatient setting if the patient remained in AF at 2–3 months after surgery. All patients were carefully examined at 1, 2, 3, and 6 months postoperatively, and every 3 months thereafter. This examination included a 12-lead electrocardiogram and 24-h Holter monitoring in addition to echocardiography. Success of the maze procedure in this study was primarily defined as maintenance of SR at 6 months post-surgery on 12-lead electrocardiogram and 24-h Holter monitoring. Following the criteria set forth in the Heart Rhythm Society, European Heart Rhythm Association, and European Cardiac Arrhythmia Society Expert Consensus Statement, we also assessed the maintenance of SR without antiarrhythmic medications at 6 months post-surgery and at the last follow-up. ⁹

Data were collected from the prospective data base and by retrospective chart review. Continuous variables are expressed as mean ± standard deviation. Univariate analysis using the chi-squared test or Fisher’s exact test was performed for categoric variables, and the unpaired t test was used for continuous variables. All variables found to be significant (p < 0.05) on univariate analysis were entered into multivariate analysis along with other variables, using logistic regression to assess the independent predictors of operative success. All analyses were performed with JMP 8.0 software (SAS Institute, Cary, NC, USA).

Results

Patient details for each group are shown in Table 2. There were no significant differences between the 2 groups in terms of age, sex, and type of AF. Although the left ventricular ejection fraction was better in group 2, there were no significant differences in the diameters of the left atrium and left ventricle, and intraoperative variables were similar in the 2 groups. There was no operative death. Two patients in group 1 required permanent pacemaker implantation for junctional bradycardia. The rate of maintenance of SR at the time of hospital discharge was 73% in group 1 and 90% in group 2 (p = 0.08). At 6 months postoperatively, maintenance of SR with and without antiarrhythmic medications was 70% and 63%, respectively, in group 1, and 94% and 90%, in group 2 (p < 0.01). Maintenance of SR at the last follow-up was 70% in group 1 at a mean follow-up of 31 ± 20 months, and significantly higher at 90% in group 2 at a mean follow-up of 13 ± 10 months (Table 3). The success of the maze procedure was assessed after risk factors such as age, type of AF, left ventricular ejection fraction, left atrial diameter, a box lesion with CS ablation, were adjusted using multivariate logistic regression analysis. Type of AF (paroxysmal) and the presence of a box lesion and CS ablation were independent predictors of the maintenance of SR at 6 months post-surgery (p = 0.027 and 0.011, respectively).

OPEN IN VIEWER

Table 2.

Characteristics of 61 patients undergoing the maze procedure

Variable	Group 1	Group 2	p Value
Mean age (years)	67.5 ± 9.1	68.0 ± 8.9	0.82
Male sex	67%	71%	0.72
Paroxysmal atrial fibrillation	17%	23%	0.56
Cardiothoracic ratio	58.3% ± 6.7%	56.2% ± 6.1%	0.22
Left ventricular ejection fraction	48% ± 16%	57% ± 15%	0.016
Left atrial diameter (mm)	50.6 ± 7.5	50.2 ± 7.8	0.86
LVEDD (mm)	53.2 ± 10.8	52.7 ± 8.0	0.82
Mitral valve procedure	90%	71%	0.060
Aortic valve procedure	27%	39%	0.32
Coronary artery bypass	25%	26%	0.59
Aortic crossclamping (min)	99 ± 30	112 ± 28	0.077
Cardiopulmonary bypass (min)	147 ± 38	144 ± 36	0.73

OPEN IN VIEWER

Table 3.

Maintenance of sinus rhythm after the maze procedure in 61 patients

				Last	Follow-Up (months)
Group		Discharge	6 Months	Follow-Up
1	All	73%	70%	70%	31 ± 20
	Drug-free		63%	53%
2	All	90%	94%	90%	13 ± 10
	Drug-free		90%	71%
p value		0.08	0.003	0.042

Discussion

Since the development of alternatives to replace the original cut-and-sew technique of the Cox maze procedure, it has become the standard surgical treatment for AF, particularly in patients undergoing other cardiac operations such as mitral valve repair. Recently, an RF bipolar clamp device has become popular due to its ease and speed of use, and high rate of transmural completeness. ^3,4 Furthermore, some surgeons have modified the Cox maze III procedure by eliminating certain lesions to simplify the lesion set. ^10,11 However, it remains unclear which lesions can be eliminated without decreasing the success rate of the operation. A few studies comparing the success rates of the cut-and-sew technique with RF ablation have been published recently. A report from the Mayo Clinic demonstrated that the cut-and-sew technique was superior to bipolar RF ablation. ¹² In contrast, a study at Washington University found no difference in success rates between the Cox maze III and IV procedures, leading to the adoption of Cox maze IV. ¹³

In terms of the transmural completeness of RF ablation, there is no definitive evidence that the ablated lesion completely blocks electrical conduction. However, blockage can be assessed intraoperatively by pacing the PV to make sure that PV stimulus is not transferred to the left atrium. In addition, intraoperative mapping of the ganglionic plexi at the PV, which are thought to play an important role in inducing AF, can be carried out to ensure the efficacy of the ablation. ¹⁴ This intraoperative assessment is only feasible for confirming PV isolation or ganglionic ablation on a beating heart; there is no practical way to assess intraoperatively the conduction block of other lesions such as those of the left atrial isthmus, coronary sinus, or right atrium. It is important to apply the bipolar clamp device 2 or 3 times for each lesion, to insure complete transmural ablation. With regard to ablation of the isthmus, Castella and colleagues ¹⁵ demonstrated in an anatomical study on the atrioventricular junction that bipolar clamps are not sufficient to achieve complete ablation. They concluded that an additional monopolar ablation device or the cut-and-sew technique was required to complete the isthmus ablation. Another recent multislice computed tomography study by Tops and colleagues ¹⁶ revealed significant variability in the position of the CS in relation to the mitral annulus. Their study indicated the possibility that a bipolar clamp device might not even reach the CS because it may lie at the same level as the mitral annulus. Therefore, the combination of a bipolar clamp and a monopolar pen to complete the Cox maze III lesion set seems to be a reasonable option.

The importance of achieving complete PV separation from the left atrial wall (box lesion) has been demonstrated by a few electrophysiological studies. Oral and colleagues ¹⁷ showed that encircling the entire PV was more effective than segmental PV ostial ablation for treating AF. By completely encircling the PV, the ablation lines may eliminate anchor points for rotors or mother waves that drive AF. Likewise, such an ablation lesion isolates the vein of Marshall, which has a left atrial insertion in close proximity to the left superior PV, and may be a source of triggers for AF. Furthermore, the connecting line may eliminate sources of AF arising from the posterior wall of the left atrium, and 25%–30% of the left atrial myocardium is excluded by the encircling lesion, limiting the area available for circulating wavelets that may perpetuate AF.

Cox and Ad ⁷ described the importance of CS ablation during the maze procedure after they performed it along with the original cut-and-sew technique for more than a decade. They found that the lack of complete conduction block at the CS may cause a high rate of recurrence of AF. Another anatomical and pathological study by Chauvin and colleagues ¹⁸ demonstrated that the venous wall of the CS is surrounded by a cuff of striated muscle extending 40 ± 8 mm from the ostium. Thus, they emphasized the need for surgical dissection or catheter ablation in or around the CS to eliminate these muscular connections for complete conduction block. A recent study at the Cleveland Clinic, including electrophysiological mapping and additional catheter ablation for patients with recurrent atrial arrhythmias after the maze procedure, demonstrated that 35% had recurrent AF resulting from recovered conduction around the PV isolation lines, and 13% of these patients had a focus mapped to the CS. ¹⁹ Although the reasons for failure of the maze procedure seemed to be multifactorial, the CS lesion was definitely thought to be an important factor in the restoration of SR.

It is difficult to discriminate the effectiveness of either the box-lesion or CS ablation contributing more to the better success. Voeller and colleagues ⁶ demonstrated that the box lesion significantly decreased the incidence of early atrial tachyarrhythmia at 1 and 3 months, and requirements for antiarrhythmic drugs at 3 and 6 months postoperatively. However, there were no significant differences in the freedom from AF recurrence at 6 and 12 months in patients with and without a box lesion. Our improved early results seem to be related to the additional box lesion. On the other hand, the additional CS ablation may have contributed to our improved late results.

Several limitations of this study should be noted. All of the data were collected from a prospective data base, but patients were not randomized into each group. There might have been a selection bias, as is usual in most retrospective studies. Also, there is a possibility that the patients in group 2 (who underwent surgery later than those in group 1) had better results due to our increased surgical experience. Because the follow-up duration was longer in group 1 than group 2, the long-term effectiveness of our modification remains uncertain. Furthermore, the number of patients in this study was relatively small; thus multivariate analysis gave less powerful statistics. Nevertheless, it was concluded that it is important to perform the box lesion and CS ablation to obtain a better success rate for the modified maze procedure using RF devices. Therefore, these 2 lesions should not be eliminated from the modified maze procedure.

A part of this study was presented at the 17 ^th Annual Congress of the Asian Society for Cardiovascular and Thoracic Surgery, Taipei, March 7, 2009.