Author + information
- Received November 7, 2016
- Revision received January 17, 2017
- Accepted February 9, 2017
- Published online August 21, 2017.
- Rakesh Gopinathannair, MD, MAa,
- Philip L. Mar, MD, PharmDa,
- Muhammad Rizwan Afzal, MDb,
- Luigi Di Biase, MDc,d,
- Yixi Tu, MDa,
- Thanmay Lakkireddye,
- Jaimin R. Trivedi, MD, MPHa,
- Mark S. Slaughter, MDa,
- Andrea Natale, MDc,d and
- Dhanunjaya Lakkireddy, MDe,∗ ()
- aUniversity of Louisville, Louisville, Kentucky
- bOhio State University, Columbus, Ohio
- cTexas Cardiac Arrhythmia Institute at St. David’s Medical Center, Austin, Texas
- dAlbert Einstein College of Medicine at Montefiore Hospital, Bronx, New York
- eUniversity of Kansas Medical Center, Kansas City, Kansas
- ↵∗Address for correspondence:
Dr. Dhanunjaya Lakkireddy, Cardiovascular Research Institute, Cardiac Electrophysiology, University of Kansas Hospital, 3901 Rainbow Boulevard, Suite G600, Kansas City, Kansas 66195.
Objectives The clinical characteristics, electrophysiological mechanisms, and ablation outcomes of post-surgical atrial fibrillation ablation (SAFA) atrial tachycardias (ATs) have not been studied in a large, multicenter cohort.
Background ATs are often seen following SAFA.
Methods Analysis was performed on 137 patients (age, 62 ± 10 years; 74% male) who underwent catheter ablation for symptomatic post-SAFA AT from 2004 to 2013 at 3 high-volume institutions in the United States.
Results A total of 137 patients had 149 ATs that were mapped; 103 (69%) had a left atrial (LA) origin and 46 (31%) had a right atrial origin. Of the 149, a total of 44 (30%) had a focal mechanism, with 29 (66%) having an LA origin, with 53% localized to LA posterior wall. Of the 105 re-entrant ATs, 74 (71%) were of LA origin. The predominant circuits were cavotricuspid isthmus (n = 25), perimitral (n = 19), LA roof (n = 17), left pulmonary veins (n = 13), right pulmonary vein/LA septum (n = 12), and LA appendage (n = 7). A total of 93% of patients had ≥1 pulmonary vein reconnection requiring reisolation. Catheter ablation resulted in termination and noninducibility of 97% of right atrial and 93% of LA ATs. Over a 12-month follow-up, 80% of patients were free of any AT or AF.
Conclusions In this large multicenter cohort of post-SAFA ATs, most were of LA origin, with macro–re-entry being the most common arrhythmia mechanism. Wide variability in location of AT circuits was seen in both right atrial and LA and likely reflects underlying arrhythmogenic substrate and differences in modified SAFA techniques. Catheter ablation was highly successful in eliminating the culprit AT with favorable long-term outcomes.
- atrial flutter
- atrial tachycardia
- catheter ablation
- MAZE procedure
- surgical atrial fibrillation ablation
The original MAZE procedure was developed to treat atrial fibrillation (AF) and atrial flutter (AFL) refractory to medical therapy (1). The procedure has undergone many modifications with several variations in lesion location and technique, with the original “cut and sew” lesions largely being replaced by intraoperative ablation using energy sources, such as radiofrequency, cryoablation, and high-frequency ultrasound, now collectively referred to as surgical AF ablation (SAFA) (2–5). Wide variability in SAFA technique, with respect to energy source and lesion sets used, exist from institution to institution and from operator to operator. No specific gold standard technique exists. Consequently, in addition to recurrent AF, atrial tachycardias (ATs) are not uncommonly seen following SAFA and if persistent, can result in significant symptoms, poor quality of life, and in some cases can result in tachycardia-induced cardiomyopathy and congestive heart failure (6). The overall incidence of post-SAFA AT in large single-center studies is significant and is estimated to range from 5% to 15% (4,7–9).
Previously, single-center studies have been conducted to elucidate the clinical characteristics and electrophysiological (EP) mechanisms (4,10). These studies have shown that most of these ATs originate in the left atrium (LA) and are likely a result of incomplete ablation at the time of SAFA (4,10). However, surgical techniques and expertise can vary significantly among institutions, and a large multicenter study has not been conducted to understand the frequency and mechanisms of these ATs. Such a study can also help broadly generalize the arrhythmia characteristics and locations of post-SAFA ATS. The purpose of the present study was to evaluate the clinical characteristics, EP mechanisms, and ablation success and outcomes of post-SAFA ATs in a large, multicenter, US cohort. We also sought to assess predictors of development of left-sided AT and predictors of recurrence following catheter ablation of post-SAFA ATs.
We retrospectively studied 137 patients who were referred for EP study and catheter ablation for symptomatic AT recurrence >90 days following SAFA between 2004 and 2013 at 3 high-volume institutions in the United States: University of Louisville, Louisville, Kentucky; University of Kansas Hospital, Kansas City, Kansas; and Texas Cardiac Arrhythmia Institute, Austin, Texas. All patients had electrocardiographically documented AT. All study patients underwent SAFA using either radiofrequency energy, cryoablation, high-frequency ultrasound, or a combination. Patients who underwent “cut and sew” operations were excluded. This research was carried out in accordance with the Declaration of Helsinki (208) of the World Medical Association and the study was approved by the institutional review board at each of the respective sites.
For each subject, demographics (age, gender, race), transthoracic echocardiogram findings (left ventricular ejection fraction [LVEF], LA diameter, grade of mitral regurgitation), type of concomitant surgery (valve repair and/or replacement or coronary artery bypass grafting [CABG]) with SAFA, AF type (paroxysmal, persistent, long-standing persistent) before SAFA, comorbid conditions, CHADS2 score (11), and medications were recorded.
Cardiac electrophysiology procedure
Electrophysiology study was performed in a fasting state using the standard protocol at each site. All antiarrhythmic agents, except amiodarone, were held for at least 5 half-lives. Intracardiac thrombus was excluded by either transesophageal echocardiography or intracardiac echo. Conscious sedation or general anesthesia was used based on operator preference. Briefly, catheters were placed in the coronary sinus and the right atrium (RA). For patients who were in a regular AT at baseline, three-dimensional (3D) activation mapping of the AT was performed using either the Ensite NAVX (St. Jude Medical, Minneapolis, Minnesota) or the CARTO (Biosense Webster, Diamond Bar, California) electroanatomic mapping system in both RA and LA. Patients who were in sinus rhythm at the beginning of the procedure underwent programmed atrial stimulation and/or burst pacing for arrhythmia induction. A stable coronary sinus electrogram was assigned as reference for mapping. A bipolar voltage <0.05 mV was assigned as scar. Findings from the activation map were confirmed using entrainment mapping for macro–re-entrant rhythms. Ablation was performed using an externally irrigated ablation catheter using power settings of the operator’s choice. For cavotricuspid isthmus dependent flutters the ablation line extended from the tricuspid annulus to the inferior vena cava with bi-directional block across the ablation line as the endpoint. For other macro–re-entrant circuits, ablation lines were extended, whenever possible, to the nearest anatomic obstacle or scar border to create a line a block. Block across linear lesions was tested using differential pacing. Following ablation of clinical AT, a circular deca- or duo-decapolar catheter was used to assess pulmonary vein (PV) conduction and reconnected veins were electrically isolated by ablation. Further substrate-based or linear ablation was left to the discretion of the operator. Programmed electrical stimulation was repeated and any inducible ATs with a stable cycle length were mapped and ablated.
Acute procedural success was defined as elimination of culprit AT with noninducibility of any ATs by programmed atrial stimulation and/or atrial burst pacing in the presence of isoproterenol.
Definitions of arrhythmia mechanisms
ATs were defined based on the classification scheme proposed by the Working Group of Arrhythmias of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology (12).
Focal AT was defined as a discrete focal source of early activation by conventional or electroanatomic activation mapping, from which there is centrifugal spread to both atria. The presence of a clearly defined isoelectric baseline between p waves in all leads of a 12-lead electrocardiogram also characterized focal AT.
Macro–re-entrant AT (MRT) was suggested when color-coded 3D electroanatomic activation mapping showed continuous activation (early meets late) in the chamber of interest that approximated the AT cycle length. Entrainment mapping showing definite evidence of progressive fusion with faster pacing cycle lengths confirmed MRT. Micro–re-entrant circuits showed continuous activity localized to a small area spanning a large portion of AT cycle length and post-pacing intervals similar to AT cycle lengths.
Patients who were on antiarrhythmic agents had them discontinued following successful ablation, either immediately following the ablation or by the end of the 3-month blanking period. All patients were followed for a median of 12 months and were evaluated at 3, 6, and 12 months using 12-lead electrocardiogram. After the 3-month blanking period, most patients also received continuous, either 3- or 7-day Holter monitors at 3-, 6-, and 12-month intervals to assess for recurrence of atrial arrhythmias.
Descriptive statistics for continuous variables are reported as mean ± SD and categorical variables as number and percentages. Predictors of left-sided AT were analyzed using a multivariate regression model with age, gender, LVEF, LA size, type of AF at baseline (paroxysmal, persistent, long-standing persistent), and presence of reconnected PVs as covariates. A similar regression analysis was performed to identify predictors of recurrence following catheter ablation, with presence of left-sided AT as an additional covariate. Statistical analysis was carried out using SPSS version 22 for Windows (IBM Corp., Armonk, New York).
Of the 137 patients (mean age, 62 ± 10 years; 74% male; 90% white), 64% had SAFA + CABG, 24% underwent SAFA + valve repair or replacement, and 6% had standalone SAFA (Figure 1). Baseline patient characteristics are shown in Table 1. Of the patients undergoing SAFA, 56% had paroxysmal AF, 41% had persistent AF, and 3% had long-standing persistent AF. Mean AF duration before SAFA was 45 ± 58 months. At baseline, 68 (50%) had hypertension, 45 (33%) had valvular heart disease, 31 (23%) had coronary artery disease, and 24 (18%) had obstructive sleep apnea.
Surgical AF: Technique and energy sources
The SAFA procedures were done using cryoablation in 64% of patients, radiofrequency energy in 27% of patients, and high-frequency ultrasound in 9% (Figure 2). The SAFA lesion set involved LA ablation only in 78% of patients and biatrial ablation in 22%. All the patients who underwent SAFA had bilateral PV isolation. However, only 28% had PV isolation as the only lesion set. Posterior wall ablation was common (posterior line in 53.3% and LA roof line in 51.8%). A mitral isthmus line was performed in 57% of SAFA patients. The other linear lesion sets included an LA septal line (6.6%), coronary sinus (3.6%), tricuspid isthmus (14.5%), superior to inferior vena cava line (9.5%), and RA septal line (0.7%). Left trial appendage closure or excision was performed at the time of SAFA in 50.4% of patients.
Atrial arrhythmia recurrence following SAFA involved AT only in 36 (26%) and both AT and AF in 101 (74%). At the time of EP study, patients had failed a mean of 1.4 ± 1.1 antiarrhythmic agents; 72% were on an atrioventricular nodal blocking agent with 60% on a beta-blocker. Mean CHADS2 score was 1 ± 1 and 89% (n = 126) were on oral anticoagulants, with 114 (83%) on warfarin. Mean LVEF was 56 ± 11% and mean LA size was 4.9 ± 2.8 cm.
AT characteristics during electrophysiology study
A total of 149 ATs were fully mapped, of which 103 (69%) had an LA origin and 46 (31%) had an RA origin. A mean of 1.1 ± 0.4 ATs were identified per patient during the EP study, with 1 AT noted in 127 (92.7%) patients, 2 ATs in 9 (6.6%) patients, and 4 ATs in 1 (0.7%) patient. Of the 149 ATs, 44 (30%) had a focal mechanism, with 29 (66%) having an LA origin and 15 (34%) having an RA origin. A total of 53% of focal LA ATs were located to the posterior wall. An example of a focal LA tachycardia is shown in Figures 3A to 3C. No significant association was noted between the type of surgical procedure (SAFA only, CABG + SAFA, valve + SAFA, and CABG + valve + SAFA) and the incidence of focal AT.
Of the 105 re-entrant ATs, 74 (71%) were of LA origin and 31(29%) were of RA origin. The predominant LA MRT circuits were mitral isthmus dependent (n = 19) (Figures 4A to 4E), LA roof dependent (n = 18), re-entry around the left PVs (n = 13), re-entry around the right PVs and LA septum (n = 12), circuit involving the LA posterior wall (n = 6), and circuit around the LA appendage (n = 7). Of the 19 (26%) patients who developed mitral isthmus–dependent flutter, 14 (74%) had a mitral isthmus line during SAFA. In comparison, 64 of 117 (54.7%) of patients without mitral isthmus–dependent flutter had a mitral isthmus line. The odds ratio for having perimitral flutter following a mitral isthmus line (vs. no mitral isthmus line) during SAFA was 2.48 (95% confidence interval: 0.840 to 7.342; p = 0.093).
Among the RA re-entrant ATs, cavotricuspid isthmus–dependent flutter (n = 25) was most common followed by RA MRT around an incisional scar (n = 3). An example of an RA incisional AT is shown in Figure 5. The other RA re-entrant ATs included upper loop flutter (n = 1), lower crista flutter (n = 1), and RA posterior wall flutter (n = 1). The location and mechanisms of post-SAFA ATs during EP evaluation are shown in Figure 6.
During EP study, >1 PV reconnection was noted in 127 (92.7%) patients and all underwent successful reisolation. Catheter ablation resulted in termination and noninducibility of 97% of all RA and 93% of all LA ATs. Overall, acute procedural success was 93.4% (139 of 149). The following procedure-related complications were noted: transient ischemic attack in 1 (0.7%), pericardial effusion in 1 (0.7%), femoral pseudoaneurysm in 1, groin bleed in 6 (4.4%), gut dysmotility and gastroparesis in 1 (0.7%), diaphragmatic paralysis followed by full recovery in 1 (0.7%), and severe pericarditis followed by effusion requiring drainage in 1 patient 8 days following ablation. Mean total procedure time was 190 ± 91 min, mean fluoroscopy time was 47 ± 26 min, and mean ablation time was 46 ± 28 min (Table 2).
Over a 12-month follow-up, 80% of patients were free of any atrial arrhythmias. Multivariate regression analysis of predictors of left-sided AT was performed with age, gender, study center, LVEF, LA size, type of AF, and PV reconnection as covariates. Age was the only variable that predicted left-sided AT in a multivariate analysis (odds ratio: 1.052; 95% confidence interval: 1.004 to 1.102; p = 0.033). A similar regression analysis was performed, using baseline parameters (age, gender, LVEF, LA size, type of AF, sleep apnea, renal insufficiency) and presence of left-sided AT, to assess predictors of atrial arrhythmia recurrence following successful catheter ablation of post-SAFA AT. None of the selected baseline variables and presence of left-sided AT predicted recurrence following catheter ablation.
To our knowledge, this is the largest and only multicenter study describing the clinical and EP characteristics of post-SAFA AT. Previously, several single-center studies have described the characteristics of post-SAFA AT (4,8–10). The inclusion of multiple centers, with multiple SAFA operators with a wide range of experience using different surgical techniques, permits greater generalizability of the data presented.
AT characteristics in comparison with prior studies
There are several similarities between our results and those of previously published single-center studies. With the exception of the study by Deneke et al. (9), the distribution of post-SAFA MRT is remarkably constant, with left-sided MRT accounting for roughly two-thirds of all MRT in previous studies (67% , 63% , 67% ). This is consistent with our findings where 71% of MRTs were left sided. This distribution of LA MRT also seems to apply to SAFA using the “cut and sew” technique where a small study showed that 60% of MRTs were left sided (14).
Of the MRTs that were left sided, the predominant circuit in our study was perimitral AFL, which is similar to the prior report from Huo et al. (10), but they noted that 59% of LA MRTs were perimitral compared with 26% in our study. The second most common LA MRT in our study was LA roof-dependent (23%) followed by re-entry around left PVs (18%), and were comparable with previously reported data (10).
The predominance of perimitral flutter may be related to incomplete ablation at the mitral isthmus during SAFA (15). The mitral isthmus stretches from the lateral commissure of the mitral valve and runs up the posteroinferior wall of the LA up to the orifice of the left inferior PV, spanning roughly 3.5 cm (15). The thickness of the LA widely varies between 3.8 ± 0.9 mm at its thickest portion and then tapers at both ends, with a myocardial depth of 1.5 ± 0.7 mm at the vestibule (16). This makes creation of the complete mitral isthmus block during SAFA technically challenging because one runs the danger of inadequate ablation near the thicker portions of the mitral isthmus and then possible perforation near the vestibule or around the left inferior PV when the thickness decreases dramatically. During endocardial ablation, most patients require ablation within in the coronary sinus to achieve complete mitral isthmus block (17,18). Furthermore, proximity to the left circumflex coronary artery can add to the difficulty during epicardial ablation of the mitral isthmus (18,19). Our data showed that 74% (14 out of 19) of patients who developed perimitral flutter had a mitral isthmus line at the time of SAFA. Although there was a trend toward higher incidence of perimitral flutter following creation of a mitral isthmus line during SAFA, it was not statistically significant (odds ratio: 2.48; 95% confidence interval: 0.840 to 7.342; p = 0.093), possibly because of smaller sample size, and needs to be tested further in larger studies.
As suggested previously, one-third of all MRT were right sided based on our results and previous studies (4,10,13). Most of the right-sided MRTs were cavotricuspid isthmus dependent, with prevalence ranging from 80% in our study to 96% in the study by Huo et al. (10). It may be pertinent to consider a cavotricuspid isthmus lesion during SAFA. In most SAFA cases, cavotricuspid isthmus lesion is typically not offered unless it is a complete biatrial MAZE. This is partly because of the anatomic issues and inability to access the cavotricuspid isthmus safely when a minimally invasive approach is used or when the RA is improperly exposed.
Several differences in AT characteristics were also noted when comparing our results with previous studies. The proportion of focal AT was considerably greater (30%) in our study compared with single-center studies where the rate of focal AT ranged from 1% to 16% (4,10). The most likely explanation for this observation is the use of different surgical techniques and lesion sets between studies. However, 53% of all true focal LA tachycardias in our study originated from the LA posterior wall, which is consistent with previous single-center studies, even though focal LA ATs overall were rare in those reports (4,10). Age was the only factor that predicted LA ATs in our study, likely reflecting advanced atrial substrate/remodeling, and associated comorbidities.
In our study, 92.7% of patients were noted to have reconnection of >1 PVs requiring further ablation for isolation. This is higher than the 79% PV reconnection rate reported by Huo et al. (10), where SAFA was done using a standard protocol, and could represent variations in operator experience, SAFA technique, lesion sets, and energy source. Catheter ablation, using a combination of 3D activation mapping and entrainment techniques, was highly successful in terminating the culprit AT with acceptable complication rates. Most patients who underwent catheter ablation were free of atrial arrhythmias at the end of follow-up. This once again highlights the limitation of the lesion set. Often a surgical lesion that is farther off from the PV ostia is prone for greater gaps and incomplete isolation. Although we did not find any statistically significant association between SAFA lesion sets and a specific type of AT mechanism/location, it is very important to create a good 3D map of the target chamber and assess the voltage and activation pattern to precisely identify the location of gaps and presence of scar. The long-term impact of different energy sources on gaps and SAFA-related scar pattern is also poorly understood.
Our report is limited by its retrospective, observational nature. However, the large sample size and multicenter experience adds credibility and broad generalizability to the data presented. All the included patients were referred to the 3 tertiary centers for EP study and ablation of post-SAFA AT. Thus the true incidence of post-SAFA AT could not be ascertained with accuracy. Smaller, single-center studies have estimated the incidence of post-SAFA AT to be 5% to 15%, but this is likely dependent on operator experience, technique, energy modality used for ablation, and preference of lesion sets (4,8,10). This statistic, therefore, may not be applicable to the larger population receiving SAFA. To determine the true incidence of post-SAFA AT, multicenter prospective studies with long-term follow-up are needed.
In this large multicenter cohort of post-SAFA ATs, most were of LA origin, with macro–re-entry being the most common arrhythmia mechanism. Wide variability in location of AT circuits was seen in both RA and LA and likely reflects underlying arrhythmogenic substrate and differences in SAFA techniques. Most post-SAFA patients had 1 or more reconnected PVs requiring repeat isolation. Catheter ablation was highly successful in eliminating the culprit AT with favorable long-term outcomes.
COMPETENCY IN MEDICAL KNOWLEDGE: Most ATs following SAFA are left-sided and macro–re-entrant in nature, and can be effectively and safely treated by catheter ablation with very good long-term outcomes.
TRANSLATIONAL OUTLOOK: The relationship between underlying arrhythmogenic substrate, ablation energy source, and incomplete SAFA lines on the mechanism and location of post-SAFA atrial tachycardias is unclear and requires further study.
Dr. Gopinathannair is a consultant for St. Jude Medical, Boston Scientific, and HealthTrust PG; and is on the speaker’s bureau for American Heart Association, Zoll Medical, AltaThera Pharmaceuticals, Pfizer, and Bristol-Myers Squibb. Dr. Di Biase is a consultant for Stereoataxis, Biosense Webster, Boston Scientific, and St. Jude Medical; and has received speaker honoraria/travel fees from Medtronic, Biotronik, Boston Scientific, EpiEp, Pfizer, and Janssen. Dr. Slaughter has served on the Oregon Heart Medical Advisory Board (no compensation); the Carmat Scientific Advisory Board (no compensation); and has received a HeartWare, Inc., training and education grant. Dr. Natale is a consultant for Stereotaxis, Biosense Webster, St. Jude Medica, and Janssen; and has received speaker honoraria from Biotronik, Medtronic, Boston Scientific, Janssen, Pfizer, Epi EP, Biosense Webster, and St. Jude Medical. Dr. Lakkireddy is a speaker and consultant to Janssen, Pfizer, SJM, and Biosense Webster; and is on the advisory board clinical study for SentreHeart. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- atrial fibrillation
- atrial flutter
- atrial tachycardia
- coronary artery bypass grafting
- left atrium/atrial
- left ventricular ejection fraction
- macro–re-entrant atrial tachycardia
- pulmonary vein
- right atrium/atrial
- surgical atrial fibrillation ablation using radiofrequency, cryoenergy, or high frequency ultrasound
- Received November 7, 2016.
- Revision received January 17, 2017.
- Accepted February 9, 2017.
- 2017 American College of Cardiology Foundation
- Cox J.L.,
- Jaquiss R.D.,
- Schuessler R.B.,
- Boineau J.P.
- Saoudi N.,
- Cosio F.,
- Waldo A.,
- et al.
- Wazni O.M.,
- Saliba W.,
- Fahmy T.,
- et al.
- Ho S.Y.,
- Cabrera J.A.,
- Sanchez-Quintana D.
- Miyazaki S.,
- Kusa S.,
- Hachiya H.,
- Iesaka Y.