|Home | About | Journals | Submit | Contact Us | Français|
This study evaluates whether non-inducibility of atrial fibrillation (AF) after achieving bi-directional electrical pulmonary vein (PV) isolation is a useful predictor of freedom from AF recurrence.
This study included 102 consecutive patients who underwent PV isolation for symptomatic paroxysmal (59%), persistent (32%), or permanent (9%) AF. Patients were followed for 16 ± 10 months. Complete isolation of all four PVs was confirmed by demonstration of bi-directional block, defined by both loss of PV potentials and failure to capture the LA by pacing (at 10 mA) 10–14 bipolar pairs of electrodes on a circumferential catheter positioned at the entrance of the PV. Induction of AF by burst pacing on/off isoproterenol was attempted after PV isolation. Freedom from recurrent symptomatic or asymptomatic AF was present in 70% of patients at 6 months and 62% of patients at 12 months. In multi-variable analysis, non-inducibility post-PV isolation (OR = 3.84, P = 0.047) and paroxysmal AF (OR = 4.80, P = 0.012) predicted freedom from AF at 12 months.
Non-inducibility of AF after bi-directional PV isolation predicts maintenance of sinus rhythm. This finding suggests that routine extensive left atrial ablation may be unnecessary.
The recent trend in catheter ablation of atrial fibrillation (AF) has been a movement from electrophysiological to anatomical approaches. The initial approach of ablating active pulmonary vein (PV) foci1,2 was replaced by segmental ostial ablation to isolate all PVs because the latter strategy decreased the rate of PV stenosis and the need for repeat procedures.3–5 More recently, an anatomical approach of circumferential PV ablation (CPVA)6 has gained popularity and replaced segmental PV isolation at some centres.7,8
A potentially useful electrophysiological predictor of long-term procedural success is non-inducibility of AF by burst atrial pacing after ablation. A recent study found that non-inducibility of AF after CPVA with additional posterior and mitral isthmus lines predicted freedom from symptomatic AF at 6 months.9 Two studies of segmental ostial ablation with PV isolation defined by entrance block reported that non-inducibility of AF post-ablation predicted freedom from AF.10,11 Although exit block of inducible PV arrhythmias has been recently proposed as a useful endpoint of ablation,12 the significance of inducible atrial arrhythmias in patients with bi-directional block demonstrated in all PVs in the absence of PV arrhythmias has not been tested.
The objective of this study is to determine whether non-inducibility of AF after achieving bi-directional electrical PV isolation predicts freedom from AF recurrence.
We selected 102 consecutive patients with symptomatic paroxysmal (59%), persistent (32%), or permanent (9%) AF, who underwent PV isolation by a single operator. All patients were included in this study. Duration of follow-up varied across patients according to time since procedure. Of the 102 patients, 86 had >6 months of follow-up and 60 had >12 months of follow-up. The median duration of follow-up was 15 months (interquartile range 8–23 months). AF was considered persistent if it was not self-terminating within 7 days or was terminated electrically or pharmacologically and considered permanent if it lasted >6 months and cardioversion failed or was not attempted.13 All patients with paroxysmal and persistent AF had recurrent episodes in the year prior to PV isolation. Patients had failed a mean of 1.5 ± 1.1 anti-arrhythmic agents prior to ablation. A single repeat PV isolation procedure was performed in six (5.9%) of the patients in the study cohort, at a median of 5 months (interquartile range 2–7 months) after the initial PV isolation procedure. Prior to ablation, all patients underwent transoesophageal echocardiography (TEE) or intracardiac echocardiography (ICE) and cardiac magnetic resonance imaging (MRI) or cardiac computed tomography (CT) to assess PV size, anatomy, and flow velocities. Baseline patient characteristics are presented in Table 1.
This study was approved by the Beth Israel Deaconess Medical Centre committee on clinical investigations. All patients provided informed consent prior to the procedure.
All procedures were performed under general anaesthesia except for the first 17 patients. A decapolar 5 mm spaced electrode catheter was positioned in the coronary sinus via a femoral venous approach and a second electrode catheter was placed in the right atrium. Left atrial access was obtained by two transseptal punctures. TEE (before and after actual ablation) or ICE (throughout the procedure) was used to identify PV anatomy and velocities before and after ablation, to identify catheter position with reference to the vein ostia, to determine the presence or absence of pericardial effusion, and in some cases to guide transseptal puncture. Following transseptal puncture, patients received intravenous heparin to maintain a serum activated clotting time > 250 s.
Three-dimensional electroanatomic mapping of the left atrium and PVs was performed using a non-irrigated NaviStar™ catheter (Biosense Webster) and CARTO™ (Biosense Webster) and/or EnSite NavX™ (Endocardial Solutions) recording systems. Electrograms were recorded during sinus rhythm, coronary sinus pacing, or AF at the ostia of the PVs with a 10–14 pole circumferential catheter with distal ring configuration (Biosense Webster or Bard). Radiofrequency ablation was performed outside the PV ostium near sites with the earliest PV electrograms. Ablation was performed for 20–60 s with a target temperature of 52°C. Temperature was considered adequate if it reached 45°C. The process was repeated until complete bi-directional electrical PV isolation was achieved, defined by both entrance block as demonstrated by loss of PV potentials and exit block demonstrated by failure to capture the left atrium during sinus rhythm by pacing (at 10 mA and 2 ms pulse width) each of the bipolar pairs of electrodes of the circumferential catheter positioned at the entrance of the PV. All PVs were routinely isolated for all patients. A repeat three-dimensional electroanatomic mapping was performed following PV isolation in the majority of patients.
After PV isolation, induction of AF was attempted by burst pacing at 200 ms for 5 s from the RA and twice from the CS. AF that did not terminate spontaneously was electrically cardioverted. Isolation of PVs was re-assessed in all patients, and if reconnection was observed, the vein was re-isolated. Isoproterenol (up to 10 μg/min) was then administered to assess triggers and reconnection, following which attempt at induction of AF by burst atrial pacing was repeated. If reconnection was observed with isoproterenol, the vein was re-isolated. If AF was inducible, the effectiveness of PV isolation was again re-evaluated with the circumferential catheter, and any observed electrical reconnection was ablated. Re-induction of AF was attempted in patients who underwent further ablation to re-isolate reconnected PVs. Inducibility was defined as induction of AF or left atrial tachycardia (after the final ablation to achieve PV isolation) lasting >10 s and was evaluated in this study as a predictor of long-term AF recurrence (defined as any AF episode lasting >10 s).
No additional left atrial lines were performed in patients without inducible AF or left atrial tachycardia after PV isolation. Overall, 52 of the 102 patients were inducible, with 22 of these patients inducible only during isoproterenol infusion. Left atrial ablation lines (mitral isthmus line and/or posterior left atrial line) were performed in 15 patients with inducible sustained left atrial tachycardia and in six patients with inducible sustained AF. Mitral isthmus lines were attempted in patients with inducible sustained AF only if sequential activation in the CS catheter suggested isthmus participation. No attempt was made to assess the completeness of the lines, although in the majority of instances, they were incomplete as judged by the absence of split potentials or presence of voltages >0.1 mV along the lesion line. Ablation within the CS was avoided, and irrigated ablation catheters were not available.
A right atrial isthmus line (tricuspid valve to inferior vena cava) was performed in patients with a history of or inducible right atrial isthmus dependent flutter (52% of the study population).
Patients were treated with warfarin to achieve an international normalized ratio of 2.0 to 3.0 for at least 6 months, as well as acetylsalicylic acid for at least 1 month. Anti-arrhythmic drugs were continued post-procedure for patients with a history of persistent or permanent AF and were re-initiated in patients with early (<30 days) recurrences of AF. Anti-arrhythmic drugs were used in 53% of patients post-procedure. During last follow-up, 28% of patients who remained free of recurrent AF were taking anti-arrhythmic drugs.
Evaluation of symptomatic or asymptomatic AF was performed using a 2 week transtelephonic event recorder and 24 h Holter monitor at 1, 3, 6, and 12 months. Additional monitoring was done for patients with symptoms. Recurrent AF was deemed present if an asymptomatic or symptomatic atrial tachyarrhythmia consistent with AF was documented to last >10 s. Freedom from AF was defined as no detected episode of symptomatic or asymptomatic AF beyond 30 days post-procedure, because >50% of patients with early recurrences (<30 days post-ablation) may be free of late recurrences of AF.14 Patients with recurrent AF were not considered free of AF even if they later responded to a change in anti-arrhythmic therapy. Repeat MRI or CT scan was performed 4–6 weeks post-procedure to assess for PV stenosis and repeated at 3 and/or 6 months, if significant narrowing was observed.
Continuous variables are expressed as mean ± standard deviation or as median (with interquartile range). Differences between groups of continuous variables were determined by analysis of variance (ANOVA). Dichotomous variables were compared using χ2 analysis. Cumulative AF-free survival was determined using Kaplan–Meier analysis, and differences in AF-free survival were evaluated using the log-rank test. We estimated that a sample size of 58 patients would be required to detect a relative risk of AF recurrence of 2 associated with AF inducibility, assuming a 25% event rate at 12 months in non-inducible patients, using a power of 0.8 and a type I error of 0.05 (two-sided). We estimated the association between non-inducibility and freedom from AF recurrence at 6 and 12 months using multi-variable logistic regression. We decided a priori to adjust for AF type because paroxysmal AF has been associated with freedom from AF in prior studies of PV isolation. We additionally evaluated four clinically meaningful variables as potential independent predictors, effect modifiers, or confounders; these variables were sex, age, hypertension, and valvular heart disease (VHD). The linearity assumption for age was assessed by dividing age into quartiles and plotting the logit for each quartile. Absence of modification of the effect of non-inducibility by other variables was verified using interaction terms. Starting from a model including all six variables, non-significant variables were sequentially removed if the resultant model Akaike Information Criterion statistic did not increase and there was little change in the magnitude of the odds ration (OR) for non-inducibility (indicating absence of confounding). The P-values were not corrected to adjust for multiple testing. The goodness-of-fit of the final models was validated using the Hosmer and Lemeshow test as well as Pearson and deviance goodness-of-fit statistics. A separate multi-variable logistic regression analysis was similarly performed using the same variables to explore whether any variables predicted which patients were inducible. All significance tests were two-sided, and a P-value <0.05 was considered statistically significant. All statistical analyses were performed using SAS Release 8.2 statistical software package (SAS Institute Inc., Cary, NC, USA).
AF was inducible after PV isolation in 52 (51%) of the 102 study patients; this included 26 (43%) of the 60 patients with paroxysmal AF and 26 (62%) of the 42 patients with persistent or permanent AF. In 22 (42%) of the 52 inducible patients, AF was only induced by burst pacing during isoproterenol infusion. No tachyarrhythmias were induced by isoproterenol in the absence of stimulation. Patients with persistent AF were more likely to have inducible AF after PV isolation. There were no other differences between patients with and without inducible AF post-PV isolation ( Table 1). In multi-variable analysis, there were no clinical predictors of inducibility. Patients who were non-inducible were more likely to be free of AF than inducible patients (82 vs. 59% at 6 months and 72 vs. 53% at 12 months; log-rank test, P = 0.040) ( Figure 1). There was no difference in AF-free survival between patients in whom AF was induced without isoproterenol and only after isoproterenol infusion (53 vs. 52% at 12 months; log-rank test, P = 0.718).
Freedom from asymptomatic or symptomatic AF was present in 70% of patients at 6 months and 62% of patients at 12 months. In multi-variable analysis, freedom from any AF at 6 months was predicted by non-inducibility post-PV isolation (OR = 4.29, P = 0.027), paroxysmal (as opposed to persistent or permanent) AF (OR = 3.24, P = 0.040), and absence of moderate-to-severe VHD (OR = 4.02, P = 0.050). Freedom from any AF at 12 months was predicted by non-inducibility post-PV isolation (OR = 3.84, P = 0.047) and paroxysmal AF (OR = 4.80, P = 0.012) ( Table 2).
AF was paroxysmal in 59% of the 102 patients in the study. Patients with paroxysmal AF were more likely than patients with persistent or permanent AF to be free of symptomatically or asymptomatically detected recurrences (81 vs. 54% at 6 months and 74 vs. 45% at 12 months; log-rank test, P < 0.001) ( Figure 2). Patients with paroxysmal AF in whom AF was non-inducible post-PV isolation had greater AF-free survival than patients with paroxysmal AF in whom AF was inducible (88 vs. 72% at 6 months and 81 vs. 64% at 12 months; log-rank test, P = 0.050). Patients with persistent AF also displayed a trend towards greater AF-free survival if AF was non-inducible (82 vs. 45% at 6 months and 65 vs. 41% at 12 months; log-rank test, P = 0.304).
Mitral isthmus and posterior left atrial ablation lines were performed only in patients with inducible left atrial arrhythmias after PV isolation. Left atrial lines were performed in 15 patients (mitral isthmus lines in 14 patients and posterior lines in eight patients) in whom inducible left AT was sustained and attributed to macro-re-entry. Empiric left atrial lines (mitral isthmus lines in five patients and posterior lines in three patients) were performed in six patients with sustained AF. Freedom from AF was not significantly better in patients with inducible AF who underwent additional left atrial lines compared with those who did not get additional ablation (57 vs. 50% at 12 months; log-rank test, P = 0.524). Left atrial tachycardias during long-term follow-up were observed in six patients (four of the 21 patients with left atrial lines and two of the 81 patients without left atrial lines).
Significant complications occurred in eight of the 102 patients who underwent PV isolation. Pericardial tamponade successfully managed by placement of a pericardial drain occurred in one patient and non-haemodynamically significant pericardial effusions occurred in two patients. A right frontal cerebrovascular accident with resultant transient right facial droop occurred in one patient 4 h post-procedure, with rapid resolution and no persistent residual deficits. Bleeding complicating femoral venous access that required blood transfusion occurred in four patients (it occurred in two patients after premature resumption of activity against advice). There were no PV stenoses (defined as >70% reduction in diameter), no atrio-oesophageal fistulae, and no deaths.
Non-inducibility of AF following PV isolation with confirmation of bi-directional conduction block predicted freedom from AF independent of clinical predictors of AF recurrence. Patients with paroxysmal AF experienced significantly better long-term AF-free survival than patients with persistent or permanent AF.
Bi-directional block is not commonly used as an endpoint of PV isolation. Although most reports consider entrance block indicative of segmental isolation,5 < 60% of PVs demonstrate exit block after entrance block is achieved; PVs without exit block only demonstrate conduction to the atrium when pacing from a median of two of the five bipoles of a circumferential catheter.15 Our experience has been similar, with ~40% of PVs demonstrating conduction from PV to atrium after entrance block, generally occurring with pacing from only one to three of the 10 bipoles tested. The finding that PVs without exit block demonstrate conduction to the atrium from only a minority of bipoles suggests that pacing from a single site is insufficient to confirm exit block. With anatomic CPVA, conduction block is not routinely assessed.6,8 Most patients undergoing repeat procedures for recurrent AF post-PV isolation have recovery of PV conduction.5,16–21 Segmental isolation using bi-directional block as an endpoint may increase long-term PV isolation and improve success rates. Although this strategy has not been compared with CPVA, two randomized studies have compared segmental PV isolation with entrance block alone with CPVA with routine left atrial lines; CPVA was superior in one study,8 but inferior in the other.22 Given that PV isolation with bi-directional block may be successful in many patients, it is important to identify procedural endpoints predictive of freedom from AF to determine whether and when to perform additional ablation.
In our study, non-inducibility of AF after bi-directional PV isolation predicted long-term freedom from AF. Patients with inducible AF had lower AF-free survival, with no difference whether AF was induced without vs. only after isoproterenol infusion. The similar predictive value and the fact that 42% of inducible patients were induced only after isoproterenol support the utility of isoproterenol in evaluating inducibility. Our finding that non-inducibility after bi-directional PV isolation predicts freedom from AF is consistent with prior studies evaluating non-inducibilty following segmental PV isolation guided by PV entrance block10,11 and non-inducibility after CPVA with routine additional left atrial lines.9 In the study of CPVA with routine additional left atrial lines in patients with paroxysmal AF, 85% of non-inducible patients were free of AF at 6 months and 20% were complicated by atrial tachycardia.9 In the current study, patients with paroxysmal AF who were not inducible after bi-directional PV isolation appeared to have better AF-free survival with less atrial tachycardia using less extensive ablation. Patients with paroxysmal AF may be good candidates for electrophysiologically guided segmental PV isolation without additional left atrial ablation, particularly if rendered non-inducible. A randomized study comparing this approach to CPVA with routine additional ablation lines would be useful to determine which method yields better results with the least complications.
The mechanism of inducible AF after PV isolation in this study is uncertain. We believe that it reflects a more diffuse substrate rather than inadequate PV isolation, because we assure PV isolation during induced AF in all patients. AF induction likely results from an interaction between the underlying atrial substrate, the stimulation protocol, and isoproterenol infusion. The fact that >50% of patients with inducible AF do not have spontaneous AF during follow-up suggests that in most patients, the atrial substrate is insufficient to initiate and sustain AF in the absence of potent triggers. In addition, AF recurrences are associated with PV reconnection and are effectively treated with repeat isolation.23 As such, we would not recommend additional ablation lines during the initial procedure because of the potential for complications with these lesion sets (given subsequently).
Anatomic CPVA is more likely to be complicated by left atrial tachycardia than segmental PV isolation, primarily due to the creation of macro-re-entrant circuits.24,25 Routine mitral isthmus and posterior lines have been added to CPVA to decrease the risk of atrial tachycardia.8,26 However, routine linear lesions have also been complicated by atrial tachycardia9,17,26,27 as well as potentially fatal atrio-oesophageal fistulae.28,29 The aim of catheter ablation procedures for AF should be to maximize AF-free survival and minimize the amount of ablation necessary. As only half of the patients with inducible AF will have recurrences, often associated with PV reconnection and effectively treatable by re-isolation,23 we believe that patients with symptomatic recurrences should generally be treated by re-isolation. Substrate-based ablation may be best reserved for patients with symptomatic recurrences despite persistence of bi-directional PV isolation. Future research should focus on defining which patients are likely to achieve long-term freedom from AF with PV isolation alone, so that extensive left atrial ablation and its associated risks may be avoided.
This is a prospective study of the predictive value of inducibility of AF post-bi-directional PV isolation. This approach was not compared with other approaches to AF ablation. The use of additional left atrial ablation in patients with inducible AF after PV isolation was not randomized, nor was conduction block by these additional lines verified. Our study was performed without irrigated tip ablation catheters that produce deeper lesions and are more likely to produce lines of conduction block. Although we attempted to identify asymptomatic recurrences of AF using event recorders and Holter monitors, some recurrences may have been missed. Nevertheless, we likely detected more episodes than most prior studies reporting only symptomatic recurrences, and our criterion for recurrence of 10 s of AF are also more rigorous than previous studies. Event recorders have been shown to detect a significant number of asymptomatic AF episodes in patients who otherwise would have been considered free of AF.30
The current study suggests that non-inducibility of AF after bi-directional electrical PV isolation is a useful endpoint that predicts freedom from AF, particularly in patients with paroxysmal AF. Studies are needed to determine the minimal lesion sets required to optimize the efficacy and safety of catheter ablation for selected patients. Research is required to further define clinical and procedural endpoints to determine when and what type of additional ablation is indicated.
Dr Essebag is the recipient of a Clinician Scientist Award from the Canadian Institutes of Health Research (CIHR). Dr Reynolds is the recipient of grant number 1K23HL077171-01 from the National Heart, Lung, and Blood Institute (NHLBI).
Conflict of interest: none declared.