Abstract
Catheter ablation is increasingly widely used to treat atrial fibrillation (AF) and constantly evolving techniques and protocols have led to improved success rates and lower risks of complications in a broad range of patients. However, long-term clinical trial data are still limited and as a result many questions remain unanswered. Particularly in patients with symptomatic paroxysmal AF, high success rates in restoring sinus rhythm (SR) with a low risk of complications are reported by experienced centres. Nevertheless, the continued suppression of AF over time, particularly with regard to the recurrence of asymptomatic episodes of AF which could influence stroke risk, is not yet adequately documented and therefore the safety of discontinuing oral anticoagulation remains unclear. The risk of silent or subclinical complications associated with ablation procedures, the likelihood of autonomic modulation and the long-term impact of ablation on left atrial mechanical function have not yet been fully determined. Limited clinical trial data suggest that catheter ablation may be particularly beneficial for patients suffering from heart failure (HF) secondary to AF, indicating improved left ventricular function and quality of life with the restoration of SR. However, the processes underlying HF and the co-existing morbidities vary from one patient to another and the factors predicting a successful outcome of ablation have not yet been fully defined. Similarly, the consistently higher rates of AF suppression achieved with catheter ablation vs. antiarrhythmic drug therapy shown in comparative clinical trials encourage consideration of this treatment as a potential first-line treatment in certain patients, but whether this is currently justifiable and if so, in which patient subsets, are open questions. This article reviews current approaches to catheter ablation and attempts to address some of the controversies regarding its use.
Introduction
A major problem in the treatment of atrial fibrillation (AF) is that it is not a single disease, but a wide spectrum of diseases with heterogeneity in clinical presentation and mechanisms, as well as therapeutic options and targets. Although patients are generally classified using a relatively simple scheme based on the chronological features of the arrhythmia, in clinical practice, treatment is also based on pattern recognition, a more difficult concept to express objectively. Increasingly, ablation procedures are modulated in accordance with the individual patient substrate and the type of AF, creation of the minimum lesion set consistent with likelihood of a successful clinical outcome being a major concern.
However, optimal tools are still lacking, patients may exhibit a delayed response to the procedure, electrophysiological and clinical outcomes may not fully coincide, and partial responses are very frequent, as is the necessity for multiple procedures, which remains a frustrating problem. Hybrid therapy combining ablation with antiarrhythmic drug therapy is also very common and objective assessment of clinical outcomes can be challenging, particularly with regard to the documentation of asymptomatic AF recurrence. Although current guidelines recommend catheter ablation as a second-line treatment option after failure of antiarrhythmic drug therapy, many centres are starting to consider it as a potential first-line option in selected patients, particularly younger individuals with highly symptomatic paroxysmal AF. However, the subsets of patients for whom this approach offers the most benefit with fewest risks remain to be defined by prospective clinical trials.
Rationale for catheter ablation
Risks of atrial fibrillation
There is considerable evidence that AF is associated with increased morbidity and mortality,1–4 as well as healthcare costs, in the long term. AF is particularly likely to increase mortality risk in patients with hypertension5 or heart failure (HF).6 Analysis of data from the AFFIRM study revealed no difference in either mortality rate (primary endpoint) or the composite secondary endpoint (death, disabling stroke, disabling anoxic encephalopathy, major nervous system haemorrhage or cardiac arrest) between patients presenting with symptomatic and asymptomatic AF, respectively, after adjustment for baseline inequalities in terms of coronary artery disease, congestive heart disease, and left ventricular ejection fraction (LVEF).7
An increased risk of stroke in patients with AF has been documented in several studies.1,8–10 Data from the Framingham study1 and, most recently, preliminary findings from the TRENDS study presented at the American College of Cardiology (ACC) 57th Annual Scientific Session11 have provided evidence that AF is an independent risk factor for stroke. The results of the TRENDS study showed a more than two-fold increase in stroke risk in patients with 5.5 h or more of device-detected atrial tachycardia (AT)/AF on any given day during the past 30 days, compared with those experiencing no episodes, independent of other stroke risk factors. Analysis of data from the European Heart Survey corroborated the current guideline that the type of AF should not be taken into account when deciding on oral anticoagulation, showing a higher risk of stroke in patients with paroxysmal AF than in those with persistent AF 1 year after cardioversion, although oral anticoagulation is paradoxically more often discontinued in these patients.12,13 Population studies have not revealed any difference in thrombo-embolic complications of AF associated with symptomatic or asymptomatic forms.14
The results of the Rotterdam study indicated that AF is also significantly associated with an increased prevalence of dementia and cognitive impairment in the elderly, in the absence of any history of overt stroke, this association remaining after adjustment for underlying disease factors such as myocardial infarction, hypertension, peripheral atherosclerosis, and diabetes mellitus.15
Absence of optimal pharmacological therapies
Several trials have indicated that restoration of sinus rhythm (SR) is associated with improvement of symptoms and quality of life,16–18 and possibly a reduction in mortality.19,20 However, as shown by the results of the AFFIRM trial, although SR is associated with better survival, antiarrhythmic drugs have limited efficacy in achieving this and may actually increase mortality20 (Figure 1).
Predictors of mortality in the AFFIRM trial. Based on data from the AFFIRM trial (the AFFIRM Investigators20).
A recent meta-analysis of the randomized trials evaluating various antiarrhythmic drugs indicated an increment in efficacy compared with placebo of only 22%, 33%, and 17% for class IA, IC, and III antiarrhythmic drugs.21 Similarly, in a substudy of the large AFFIRM trial, only amiodarone showed a success rate exceeding that of placebo (60% compared with 23–38% for other antiarrhythmic drugs and 35% for placebo.20 In the Rate Control vs. Electrical Cardioversion for AF (RACE) trial, only 36% of patients in the rhythm-control group were in SR at the end of follow-up,22 and in the Strategies for Treatment of AF (STAF), the rate was only 23%.23
The limited efficacy and often poorly tolerated side-effects of the currently available antiarrhythmic drugs reduce compliance with treatment, overall discontinuation rates in the Canadian Trial of Atrial Fibrillation (CTAF), for example, ranging from 34% among patients assigned to amiodarone to 46% of those assigned to sotalol or propafenone.24 The percentage of cross-overs from rhythm control to rate control in AFFIRM reached 38% after 5 years,25 and trials comparing antiarrhythmic drug therapy to catheter ablation have similarly reported high rates of cross-over to ablation among patients experiencing recurrent AF: 51% (after the protocol-defined 1 year follow-up) in the trial reported by Wazni et al.,26 42% in the APAF study,27 77% in the trial comparing antiarrhythmic drug therapy alone to drug therapy plus ablation reported by Oral et al.28 The increased mortality associated with the use of antiarrhythmic drugs demonstrated by the AFFIRM substudy20 was also shown in previous trials, including CAST,29–31 SWORD,32 and SPAF.33
Current status of catheter ablation for atrial fibrillation
There is now convincing data that relief of symptoms and quality of life can be achieved by experienced operators in a substantial majority of patients undergoing catheter ablation for paroxysmal AF and a significant percentage of patients undergoing more extensive ablation procedures for persistent AF. However, apparent success rates in terms of AF recurrence are highly dependent on the mode and intensity of rhythm monitoring,34–36 and trial-based information on the attainment of hard endpoints like delayed onset of HF and reduced stroke, mortality and hospitalization rates, as well as lower cost of care, are still lacking at present.
The guidelines published in 2006 by the American College of Cardiology (ACC), the American Heart Association (AHA), and the European Society of Cardiology (ESC) list catheter ablation as a second-line option for practically all subsets of patients with AF37 (Figure 2). However, while there is general agreement on the use of catheter ablation as second-line, or even sometimes as first-line therapy, in patients with symptomatic AF and no structural heart disease or minimal heart disease, the optimal approach to the treatment of other patient categories is less well-defined. Which patients are likely to be the best candidates for catheter ablation, at what stage to propose the procedure, and what are the likely outcomes that should be presented to the patient are still open questions with no clear answers based on large-scale clinical trial data. The techniques used for catheter ablation are still evolving, most patients undergoing ablation procedures have tended to be relatively young compared with the patients included in the AFFIRM trial, for example, and the overall population with AF, and long-term data on the efficacy and safety of this approach in large populations are still limited.
American College of Cardiology (ACC)/the American Heart Association (AHA)/European Society of Cardiology (ESC) 2006 Guidelines for the treatment of atrial fibrillation. (Adapted from Fuster et al., J Am Coll Cardiol 2006;48:e149–e246.)
The expert consensus statement on catheter and surgical ablation of AF issued jointly by the Heart Rhythm Society (HRS), the European Heart Rhythm Association (EHRA), and the European Cardiac Arrhythmia Society (ECAS)38 emphasizes that although catheter ablation of AF is now a commonly performed procedure in major hospitals, many questions remain unresolved and cannot be satisfactorily answered on the basis of the data currently available. Both this consensus statement and the Venice Chart international consensus document on AF ablation39 stress the need for sufficiently powered multicentre, randomized trials to address these crucial issues, including in particular, long-term mortality and morbidity outcomes, but also the safety and efficacy of different ablation techniques, patient management before, during, and after ablation, prevention and treatment of complications, the likely benefit of catheter ablation in patients with different types of AF and different types of underlying cardiac and non-cardiac disease and the cost-effectiveness of this procedure compared with alternative therapeutic strategies.
AF ablation procedures are unique in their complexity and the HRS/EHRA/ECAS consensus statement emphasizes the crucial importance of specialized training and extensive procedural experience to ensure patient safety and optimal outcomes. Besides mastery of a wide range of sophisticated technical skills, including trans-septal puncture, precise catheter manipulation for mapping and ablation, adjustment of energy application and appropriate use of fluoroscopy, safe and successful ablation of AF demands a particularly high level of anatomical and electrophysiological expertise. This should encompass detailed knowledge of the anatomy of the left atrium and adjacent structures, ability to interpret intracardiac electrograms to recognize pulmonary vein (PV) potentials and determine when electrical isolation of a PV has been achieved, understanding the role of coronary sinus pacing in the differentiation of far-field electrograms from PV potentials and ability to identify other supraventricular tachycardias that may trigger AF, as well as awareness of potential complications and ways of preventing, recognizing, and managing these. The consensus statement strongly advocates additional training beyond the standard fellowship, including observation and supervised practice in experienced centres.38
Targets for catheter ablation
The complexity and diversity of ablation techniques used for the treatment of AF reflect the multiple electrophysiological and autonomic mechanisms underlying this arrhythmia. Strategies and techniques of catheter ablation have evolved in parallel with improved understanding of these mechanisms, targeting either the drivers triggering AF, predominantly located within the PVs, or features of the atrial substrate that perpetuate the arrhythmia, such as arcs of intra-atrial conduction block, areas of slow conduction and rotors reflecting areas of re-entry within the atria.
Current approaches to catheter ablation
Catheter ablation is currently based on four main strategies:40
Isolation of the triggers and perpetuating re-entrant circuits located in the pulmonary veins;
Disruption of the substrate for perpetuating rotors in the antra of the pulmonary veins;
Disruption of putative dominant rotors in the left and right atria, recognized by high-frequency complex fractionated electrograms during mapping of AF.
Targeted ablation of ganglionated autonomic plexi in the epicardial fat pads.
Ostial and circumferential pulmonary vein isolation
Initial attempts to suppress AF by creating long linear lesions in the atria, replicating the surgical MAZE procedure,41 led to the detection of focal electrical activity triggering AF in the sleeves of atrial tissue extending into the PV,42,43 providing an alternative target for catheter ablation. However, ablation of focal triggers within the PVs was associated with a high incidence of PV stenosis and occlusion, leading to the development of two strategies designed to isolate the PV from the atrial substrate: segmental ostial PV ablation44 and circumferential PV ablation (CPVA)45 or intracardiac echography-guided PV antrum isolation (PVAI).46
The first approach involves ablation of electrical triggers at the PV ostia, guided by a multipolar circumferential mapping catheter and fluoroscopy, with an endpoint of electrical isolation of all four PV, manifested by elimination of all distal PV potentials. This endpoint was found to correlate more closely with clinical success than acute suppression of arrhythmias.44
The second approach involves empirical, anatomical isolation of the PV, continuous circular lesions being made around each PV or around each pair of ipsilateral PV at a distance from the ostia, with an endpoint of delay in conduction from the left atrium to the PV.45,47 The optimal distance from the ostia for creating circumferential lesions remains controversial. With increasing distance from the ostia, more lesions are required to achieve PV isolation and the likelihood of incomplete isolation is increased, but conversely the risk of PV stenosis is reduced and the inclusion of a larger area of atrial tissue within the lesion line may result in a greater impact on atrial rotors responsible for the maintenance of AF.40
In a study measuring the prevalence of electrical isolation of PVs after the creation of coalescent wide circumferential lesions around each PV, residual conduction was evident in 45% of the PVs.48 However, the initial study performed using this procedure in patients with paroxysmal (54%) or permanent (46%) AF showed no correlation between AF recurrence and the number of PVs with incomplete circumferential lesions. Overall, 85% of patients were free of AF at a mean follow-up of 9 months, 62% in the absence of antiarrhythmic drug therapy.45 In a further refinement of the CPVA approach, intracardiac echocardiography (ICE) is used to guide ablation around the funnel-shaped antrum of each PV extending posteriorly far from the ostium and blending into the posterior atrial wall at an oblique angle.46
Using either ICE or angiographic circular mapping to guide ablation in 323 patients with drug-refractory symptomatic paroxysmal (54%), persistent (11%), or permanent (35%) AF aged from 18 to 79 years, the success rate was over 80% after the first procedure with a minimum follow-up of 180 days. A successful outcome was defined as the absence of both AF and atrial flutter in the absence of antiarrhythmic drugs after a blanking period of 2 months post-ablation. Success rates were highest in patients presenting with paroxysmal AF, but did not differ significantly between age groups (<50 years, 50–60 years, >60 years) irrespective of the type of AF at baseline.49 Another retrospective analysis performed by the same centre compared the outcome of 315 consecutive patients with symptomatic paroxysmal (51%), persistent (13%), or permanent (36%) AF, according to the use of ICE or conventional circular mapping and angiography. This analysis showed a significantly lower recurrence rate with ICE-guided ablation at a mean follow-up of 417 ± 145 days (12.7 vs. 19.6%, P = 0.01) and a significantly reduced procedure time (P < 0.05). Additional monitoring of energy delivery using ICE to track microbubble formation further diminished AF recurrence rate and reduced the incidence of severe PV stenosis and embolic events to zero.50
Mansour et al. reported early experience with PV isolation close to the ostium or at a distance from this, in a cohort of 80 consecutive patients of whom 40 underwent segmental ostial PV ablation and 40 underwent CPVA, with a single encirclement of each pair of ipsilateral veins and an endpoint of PV isolation rather than LA to PV conduction delay.51 Structural heart disease was present in 68% and 63% of patients, respectively, in the segmental ostial PVA and CPVA groups; 18% and 20%, respectively, presented persistent AF and 13% of patients in each group had a LVEF of <40%. Success rates (freedom from AF at follow-up, Kaplan–Meier analysis) were 60% with segmental ostial ablation and 75% with CPVA, the difference being statistically non-significant. The incidence of procedural complications was similar in both the groups.
A later prospective, randomized study compared the efficacy and safety of segmental ostial PV isolation with those of left atrial catheter ablation (LACA), comprising CPVA plus two additional linear ablation lines, one along the posterior left atrium connecting the circumferential ablation lines around the right and left PVs and one along the mitral isthmus connecting the circumferential ablation line around the left PVs to the lateral mitral valve annulus, in 80 consecutive patients with symptomatic paroxysmal AF. The results showed the LACA approach to be more effective in preventing symptomatic recurrence of AF at 6-month follow-up, 88 vs. 67% of patients presenting no symptomatic recurrence in the absence of antiarrhythmic drug therapy (P = 0.02). Complications were limited to the development of left atrial flutter in one patient undergoing LACA, abolished by additional ablation in the mitral isthmus.52
The benefit of adding linear mitral isthmus ablation to PV isolation with linear ablation of the cavotricuspid isthmus was assessed in two groups of 100 consecutive patients with drug-refractory symptomatic paroxysmal AF.53 Significantly fewer patients undergoing additional linear mitral isthmus ablation required a second procedure as a result of AF recurrence (32 vs. 49%; P = 0.02) and a significantly higher percentage of the patients in this group were free of arrhythmia in the absence of antiarrhythmic drug therapy at 1 year following the last procedure (87 vs. 69%, P = 0.002).
Whether complete electrical isolation of all four PV is necessary for clinical success in all the patients and whether linear lesions within the atria in addition to PV isolation provide further clinical benefit in all patients are still open questions.54,55 In view of the increased procedural risk and higher incidence of incisional macro-re-entry propagating through incomplete linear lesions, Jaïs et al.54 recommended deploying linear atrial lesions only in patients exhibiting persistent or inducible sustained AF after PV isolation (around 50% of patients undergoing ablation for AF at this centre). Although complete PV isolation may not be essential in the CPVA approach,45,56 recovered PV conduction through incomplete circular lesions was found to be the dominant mechanism underlying recurrence of AT after this procedure in the study reported by Ouyang et al.57
Ablation of areas of complex fractionated atrial electrograms
The third approach, pioneered by Nadamanee et al.,58 involves mapping and ablating areas of the left and right atria characterized by temporally and spatially stable complex fractionated atrial electrograms (CFAEs), postulated to be the pivot points for re-entrant wavelets. CFAEs are defined as electrograms that are fractionated and exhibit two or more deflections and/or perturbation of the baseline with continuous deflections from a prolonged activation complex, or have a very short cycle length (≤120 ms) with or without multiple potentials, compared with those recorded from other areas of the atria. They are frequently located in the intra-atrial septum, the PVs, and the roof of the left atrium. CFAEs are invariably mapped during AF, either spontaneous or induced. The procedural endpoints are complete ablation of the areas with CFAEs or conversion of AF to SR.58,59 In a study on 121 patients with drug-refractory paroxysmal (47%), persistent (22%), or permanent (31%) AF, termination of AF occurred during ablation in 95% of patients and at 12-month follow-up, 70% of patients were free of AF in the absence of antiarrhythmic drug therapy having undergone a single ablation procedure.58
Ablation of cardiac ganglionic plexi
The fourth approach involves the use of high-frequency stimulation to identify ganglionic plexi as target sites for ablation, on the hypothesis that local cardiac autonomic ganglia clustered in the fat pads at the margins of the PV antra innervate PV myocardial sleeves and may play a critical role in the initiation and maintenance of AF.60 Local cardiac denervation by radiofrequency ablation at the PV–atrial junctions in 26 consecutive patients with persistent or chronic AF achieved non-inducibility of AF without PV isolation in 89% of cases with a small number of lesions (a mean of 12 in the first 10 patients, 7 in the following 10 patients, and 4 in the last six patients). At a median follow-up of 6 months, 84% of the patients remained in SR, 30% in the absence of antiarrhythmic drug therapy, and 54% on the previously ineffective antiarrhythmic drugs.61 In a cohort of 297 patients experiencing frequent episodes of symptomatic and asymptomatic AF (31% with associated cardiovascular disease), vagal reflexes could be elicited and abolished by targeted ablation during CPVA in 34% of patients.62 The predominant sites of vagal reflex around the PV ostia were the cranial junction between the left superior PV and the LA (95% of patients) and the septal or anterior junction between the right superior PV and the LA (25% of patients). Among patients undergoing CPVA with complete PV denervation, 99% were free of AF recurrence at 12 months compared with 85% of patients in whom vagal reflexes could not be elicited and who therefore underwent CPVA only (P = 0.0002).
Definition of success after catheter ablation
Although abolition of AF is the primary endpoint of catheter ablation in clinical trials, successful outcome of ablation in clinical practice is more likely to be defined in terms of symptom relief and improved quality of life—in other words a ‘happy patient’. It is increasingly recognized that partial success in eliminating AF may be enough to limit the deleterious consequences of this arrhythmia and that the benefits of complete abolition of AF need to be balanced against the procedural risks and possible adverse long-term consequences of more extensive ablation lesions. However, the relationship between the magnitude of AF burden and the risk of morbidity and mortality is unclear and the cut-off point for defining a ‘low risk’ AF burden has not yet been defined.46 In particular, it is not fully known to what extent paroxysmal AF is associated with a risk of stroke. The probability that paroxysmal AF will progress to persistent AF, with an increased risk of atrial remodelling, is also unclear; many patients never progress beyond the paroxysmal form.
Several authors have advocated a tailored approach for catheter ablation of paroxysmal AF, on the basis that the pathogenesis of AF is multifactorial and the use of a standardized lesion set may involve more ablation than necessary in some patients and not enough in others. Clinical trial data indicate that complete electrical isolation of the PV results in freedom from AF in approximately 70% of patients with paroxysmal AF and in about 25% of those with persistent AF in which the PV may play only a minor role.63 However, achievement of this electrophysiological endpoint may not be essential for a successful clinical outcome in all the patients. The challenge is to recognize the various subtypes of AF so that the ablation procedure can be tailored accordingly.
Oral et al.64 evaluated a staged approach to catheter ablation in 153 consecutive patients (mean age 56 ± 11 years) with symptomatic paroxysmal AF, based on mapping of the PV, left atrium, and superior vena cava during spontaneous or induced AF. The endpoint of ablation was termination and non-inducibility of AF. Complete PV isolation was not a required endpoint. In the first instance, arrhythmogenic PVs were isolated and encircled and if AF was still present or inducible, areas generating complex electrograms in the left atrium, coronary sinus, and superior vena cava were targeted, avoiding routine energy applications near the oesophagus.
In 33% of the patients, both the right and left PVs were encircled, in 10% only the right PVs and in 3% only the left PVs. Focal ostial ablations to eliminate PV tachycardias were performed at a left superior PV in 24% of patients, right superior PV in 19%, left inferior PV in 20%, and right inferior PV in 12%. After elimination of PV tachycardias, ablation was performed in the left atrium in 69% of patients, targeting driver tachycardias and/or complex electrograms, in the coronary sinus in 46% and in the superior vena cava in 6%. During ablation, AF was converted to SR in 45% of the patients and to atrial tachycardia or flutter in 27%, corresponding to a successful outcome in a total of 72% of patients. AF was rendered non-inducible in 58% of the patients. Left atrial flutter developed after ablation in 19% of patients, resolving spontaneously in half of these by 12 weeks of follow-up.
At a mean follow-up of 8 ± 2 months after the first procedure, 28 patients (18%) underwent a repeat ablation procedure for recurrent AF (13%) or atypical atrial flutter (5%). Among the 20 patients with recurrent AF, arrhythmogenic activity in PVs not originally targeted was seen in 12 patients, recovery of conduction in previously targeted PVs in three patients, and arrhythmogenic activity in the superior vena cava in two patients. In one patient, showing no arrhythmogenic activity in either the PVs or the superior vena cava, sites generating residual complex electrograms were ablated. The remaining two patients had experienced immediate AF recurrences originating in an initially targeted left PV located very close to the oesophagus for effective energy applications to be delivered. These sites could be effectively targeted in the repeat procedure as the oesophagus had moved. Total complications comprised pericardial tamponade in two patients (1%) and transient neurological events in two patients (1%).
During a mean follow-up of 11±4 months after the last ablation procedure, 77% of the patients were free of AF or atrial flutter in the absence of antiarrhythmic drug therapy. Seven patients (5%) had atypical atrial flutter and 28 patients (18%) still had paroxysmal AF, although four of them showed a more than 50% reduction in the duration and frequency of their symptoms in the absence of antiarrhythmic drug therapy. In patients whose AF was rendered non-inducible, the long-term efficacy of ablation was significantly higher than when this was not achieved. However, over 60% of patients in whom AF was still inducible after ablation remained free from recurrent AF during follow-up.
Jaïs et al.65 prospectively evaluated a step-wise, individualized ablation approach guided by non-inducibility of AF in 74 patients (mean age 53 ± 8 years) with paroxysmal AF. In the first instance, PV isolation was performed during induced or spontaneous AF. If AF was still inducible, one or two additional linear lesions were made at the mitral isthmus and/or atrial roof (avoiding lesions close to the posterior wall), with the endpoint of non-inducibility of AF or atrial flutter. In 57% of patients, PV isolation restored SR and rendered AF non-inducible. In 27%, the addition of a single linear lesion achieved non-inducibility of AF, and a further 14% attained non-inducibility of AF and restoration of SR after a second linear lesion. A repeat procedure was necessary in 23 patients (31%) who presented AF (15%) or organized atrial arrhythmias (16%) at the end of the 1-month blanking period after the index procedure. In 15 patients (20%), mapping revealed that the recurrent arrhythmia was related to the recovery of a previously targeted site (PVs in 10 patients, linear lesions in three, both in two) and re-ablation at these sites restored SR. In the remaining eight patients (11%), additional linear lesions were required.
At a mean follow-up of 18 ± 4 months after the final procedure, 67 patients (91%) were free of AF and atrial flutter without the use of antiarrhythmic drugs. The only complication was asymptomatic severe stenosis of the left inferior PV in one patient, detected by CT scan at 1 year post-ablation. This staged approach is consistent with the aim of achieving the best efficacy/safety ratio in each patient by performing the minimal set of lesions associated with a successful outcome. A long-term success rate of 91% was achieved, while restricting delivery of linear lesions to <50% of the patients. As the two patients in whom non-inducibility of AF was not achieved nevertheless remained free of arrhythmia during follow-up, the authors suggest that non-inducibility of AF is probably a more valuable endpoint for stopping ablation than persistent AF-inducibility is for indicating the need to pursue ablation. From this standpoint, an alternative strategy might be to limit the initial procedure to PV isolation and await clinical recurrence of AF before adding linear lesions.
This approach was also advocated by Cheema et al. on the basis of a study including 64 consecutive patients with paroxysmal (45%), persistent (29%), or permanent (25%) AF, 18% of the total patient population presenting structural heart disease.66 Rates of freedom from AF recurrence in the absence of antiarrhythmic drug therapy after a single PV isolation procedure and a mean follow-up of 13 ± 1 months did not differ significantly in patients with paroxysmal, persistent, and permanent AF (48%, 36%, and 50%, respectively; P = 0.90), suggesting that irrespective of the type of AF, the need for additional linear lesions or point substrate ablations in the initial procedure is questionable.
Tamborero et al.67 evaluated an individually tailored approach in 131 consecutive patients with drug-refractory paroxysmal (76%) or persistent (24%) AF, assigning patients to either selective segmental ostial PV isolation targeting only PVs with electrical potentials44,68 (n = 50), or electroanatomically guided CPVA47,69 (n = 81) with additional linear lesions at the posterior atrial wall and mitral isthmus to control left atrial flutter, based on the clinical characteristics of each patient. The ablation procedure chosen comprised selective segmental ostial PV isolation in patients with a left atrial diameter no larger than 40 mm and frequent runs of atrial tachycardia of more than 10 beats during Holter monitoring, assumed to have a predominantly focal mechanism of arrhythmia, and CPVA with additional linear lesions to modify the atrial substrate in the other patients. The electrophysiological endpoint of the CPVA procedure was voltage abatement within the encircled areas; conduction block across the linear lesions was not assessed. AF recurrence was assessed routinely by 24-Holter monitoring at 1, 4, and 7 months, then every 6 months if the patient remained asymptomatic, patients being encouraged to report any symptoms suggestive of atrial arrhythmia between visits and undergo additional ECG recordings if symptoms occurred.
The CPVA group included a greater percentage of patients with persistent AF (32 vs. 12%) and with hypertension (36 vs. 22%; P = 0.04) and the mean age of the patients in the CPVA group was higher (52 ± 11 vs. 47 ± 13; P = 0.02). Other characteristics did not differ significantly between the two groups, except for the procedure assignment criterion of left atrial diameter (mean 43 ± 5 in the CPVA group vs. 37±4 mm in the selective segmental ostial PV isolation group; P < 0.001). Freedom from atrial arrhythmias during a mean follow-up of 21.5 ± 15.2 months was achieved by 70% of patients undergoing selective segmental ostial PV isolation (mean number of PVs disconnected per patient: 1.8 ± 0.7) and by 72% in those undergoing CPVA with additional linear lesions (NS). In the absence of antiarrhythmic drugs, 58% and 59% of patients, respectively, remained free of arrhythmia recurrence. With respect to the 99 patients (76%) with paroxysmal AF at baseline (44 assigned to selective segmental ostial PV isolation and 55 to CPVA), the two procedures showed similar efficacy, 77% and 76% of patients, respectively, remaining free of arrhythmia during a mean follow-up of 22.2 ± 15.8 months (NS). In contrast, in accordance with the results of other studies,51,70,71 CPVA was more effective than selective segmental ostial PV isolation in patients with persistent AF, freedom from arrhythmia during a mean follow-up of 19.3 ± 13.3 months being achieved by 64% of patients vs. 17%, respectively (P = 0.02).
Complications associated with selective segmental ostial PV isolation comprised transient inferior myocardial ischaemia in two patients (4%), post-procedural pericarditis in one patient (2%), and femoral arterio-venous fistula, necessitating surgical correction, in one patient (2%). In the CPVA group, one patient (1%) suffered transient cerebrovascular ischaemia with dyplopia, six (7%) presented post-procedural pericarditis, of whom two developed a complete Dressler syndrome, and cardiac tamponade related to trans-septal puncture occurred in three patients (4%). Magnetic resonance angiography performed in the last 74 patients of the series revealed PV stenosis in 18% of the patients having undergone selective segmental ostial PV isolation and in no patient in the CPVA group. No symptoms associated with PV stenosis were observed during follow-up.
A second ablation procedure was performed in four patients (8%) in the segmental ostial PV isolation group owing to AF recurrences, recovery of conduction in at least one of the initially targeted PVs being seen in all these patients. Three of these patients subsequently remained arrhythmia-free during follow-up. In the CPVA group, 13 patients (16%) underwent a second ablation procedure prompted by persistence of AF or development of incessant left atrial flutter. Eight of these patients subsequently remained free of arrhythmia recurrences during follow-up. In both groups, the remaining patients experiencing arrhythmia recurrence after ablation presented a lower frequency of AF episodes and reported significant clinical improvement, so no further procedure was performed.
The results of these studies demonstrate the feasibility of a tailored approach to catheter ablation of AF, but at the same time illustrate the difficulty of predicting clinical outcome on the basis of electrophysiological endpoints. Clinicians would like to be able to reassure patients that they are unlikely to need a repeat ablation procedure, but this is generally not yet the case.
In a study on 674 high-risk patients (mean age: 67 ± 12 years, median: 69 years; 49% with hypertension, 21% with coronary disease, 23% with an ejection fraction below 40%, 10% having experienced a previous stroke) undergoing catheter ablation guided by CFAE mapping for paroxysmal (40%), persistent (23%), or permanent (37%) AF, 52% of the patients required only one procedure, 32% two procedures, 13% three procedures, and 3% four procedures.72 The necessity of a repeat ablation procedure in approximately one-third of the patients in the study by the Bordeaux group evaluating a stepwise ablation strategy65 corresponds to our experience in the Massachusetts General Hospital. In most centres, major problems relate to PV reconnection and the development of various types of left atrial flutters and tachycardias after the initial procedure. Cheema et al.73 reported a single-procedure success rate of <40% after a mean follow-up of 26 ± 11 months in 200 consecutive patients with paroxysmal (46%) or persistent/permanent (54%) AF undergoing either segmental PV isolation with a linear lesion in the cavotricuspid isthmus, or CPVA with additional linear lesions in the cavotricuspid isthmus, mitral isthmus, and posterior left atrium. Success was defined as freedom from symptomatic AF in the absence of antiarrhythmic drug therapy during the 6 months after the previous follow-up visit. Multivariate logistic regression analysis showed non-paroxysmal AF and segmental ablation to be independent predictors of single procedure failure (P < 0.01 and P = 0.03, respectively), but not age, gender, AF duration, left atrial size, ejection fraction, or structural heart disease.
The tailored approach comprising a minimal ablation procedure in the first instance may even tend to increase the need for repeat procedures if the goal is complete freedom from AF recurrence. Acceptance of a clinically relevant outcome comprising reduced frequency and intensity of AF episodes rather than complete restoration of SR might limit multiple procedures and procedural risks, while minimizing damage to the atrial myocardium and possible adverse long-term consequences.
Catheter ablation as a first-line therapeutic option
The concept that SR is a desirable outcome, but one that cannot be achieved effectively and safely with the existing pharmacological strategies, is the basis for considering catheter ablation as a possible first-line treatment option. This approach, in contrast to rate control or the use of antiarrhythmic drugs, also potentially offers the prospect of a lasting cure, at least for paroxysmal AF, although long-term data on efficacy and safety are still limited. The overt complications associated with ablation procedures are well known and the development of techniques and protocols is progressing rapidly. In contrast, the impact of the procedure with regard to silent embolic events, particularly those affecting the brain, subclinical PV stenosis, left atrial mechanical function, autonomic modulation, and recurrent asymptomatic episodes of AF is still poorly documented and warrants close monitoring over the long-term.
Considering catheter ablation as a first-line therapy for certain patient subsets will depend above all on its evidence-based short- and long-term efficacy and safety in these populations and confidence in its superior benefit–risk ratio compared with other treatment options. The relative cost and availability of this approach are also relevant factors. The initial cost of an ablation procedure is very high, but if the clinical outcome is superior to that achieved by other alternative treatments, this may be outweighed by subsequent savings in healthcare costs within a few years. Catheter ablation for AF is still performed only in relatively few centres and long waiting lists for ablation, resulting in treatment delays from several months to over a year, may be a problem in Europe and Canada, but for the moment not in the USA.
A wide range of patients with AF are now potential candidates for catheter ablation, although clinical trial data indicate that the probability of a successful outcome with a minimum lesion set is highest in patients with paroxysmal AF and no or minimal structural heart disease. The complexity of the procedure means that operator experience is a major factor in determining the outcome and clinical practice surveys indicate lower success rates than those reported in clinical trials, with wide variations in both successful outcome and complication rates.
A worldwide survey of 181 centres performing a total of 11 762 AF ablation procedures in 9370 patients reported mean success rates without antiarrhythmic drug therapy of 53% in 65 centres operating only patients with paroxysmal AF, 49% in 17 centres operating patients with paroxysmal or persistent AF, and 57% in eight centres operating on patients with all forms of AF.74 The overall rate of major complications was 6%, but fatal or disability-engendering complications were uncommon [death: 0.05%, stroke: 0.28%, transient ischaemic attack (TIA): 0.66%]. A survey of 80 European centres based on a total of 6759 AF ablation procedures reported a mean success rate (freedom from AF recurrence without antiarrhythmic drug therapy) of 51 ± 22%, with no mortalities and an average stroke incidence of 1 ± 2% (range 0–10%) and an incidence of over 50% PV stenosis of 1.85 ± 3% (range 0–11%).75 In a US survey including 92 physicians performing AF ablation, the mean 1-year success rate of 1431 procedures performed in 2002 was 58% and the projected mean 1-year success rate of the 2575 procedures performed in 2003 was 66%.76 Complications were not reported. An approximately four-fold increase was noted in the average number of AF ablations performed annually between 2000 and 2003 overall and an approximately eight-fold increase in non-academic centres. Nevertheless, academic centres still accounted for 64% of ablation procedures for AF in 2003 and reported significantly better short- and long-term success rates.
Efficacy and safety of catheter ablation vs. antiarrhythmic drugs
Several trials have compared the efficacy and safety of catheter ablation and antiarrhythmic drug therapy for the treatment of AF, including two randomized trials in patients with paroxysmal AF,26,27 two randomized trials comparing antiarrhythmic therapy alone and in combination with catheter ablation,28,77 and one long-term non-randomized retrospective study in which close to one-third of the patients presented chronic AF.69
Prospective randomized comparisons of ablation and antiarrhythmic drug therapy
The Radiofrequency Ablation vs. Antiarrhythmic drugs as First-line Treatment of Symptomatic Atrial Fibrillation Trial (RAAFT) randomized 70 patients with predominantly paroxysmal AF to catheter ablation (n = 33, mean age 53 ± 8 years, 97% paroxysmal AF) or antiarrhythmic drug therapy (n = 37, mean age 54 ± 8 years, 95% paroxysmal AF) with a follow-up of 1 year.26 At baseline, the two groups were similar with regard to mean duration of AF (5 years), mean LVEF, mean left atrial size, presence of structural heart disease and hypertension, use of beta-blockers, and quality of life. Patients in the ablation group underwent PV antral isolation using intracardiac echography to guide catheter positioning, targeting of energy application and energy titration,46,50 with an endpoint of complete electrical disconnection of all four PV antra from the left atrium. In the antiarrhythmic drug therapy group, the recommended treatment comprised flecainide, propafenone, or sotalol at maximum tolerated doses, and in third line, amiodarone. Continuation of beta-blocker therapy was at the investigator's discretion in both the groups. AF recurrence was monitored by a loop event-recorder worn for 1 month during the first month, at 3 months and subsequently in patients experiencing symptoms, with recordings at least once a day, by 24 h Holter monitoring before hospital discharge and at 3, 6, and 12 months post-enrolment, and by monthly telephone interviews with the patients.
During the first 2 months of follow-up, 54% of patients in the antiarrhythmic drug therapy group experienced recurrence of AF, resulting in 26 hospitalizations for cardioversion or medication adjustment, compared with 27% in the ablation group with no hospitalizations. No thrombo-embolic events occurred in either group. Excluding events during this 2-month period, 63% of patients in the drug therapy group experienced recurrence of symptomatic AF during 1 year follow-up vs. 13% in the ablation group (P < 0.001). Asymptomatic AF was documented in 16% and 2% of patients, respectively. In both groups, the mean number of episodes of non-sustained AF episodes recorded by Holter monitoring decreased significantly between baseline and study completion. During follow-up, 54% of patients receiving drug therapy and 9% of those undergoing ablation were hospitalized (P < 0.001). There were no repeat procedures in the ablation group. Quality of life improved to a significantly greater extent in patients undergoing ablation than in those receiving antiarrhythmic drug therapy with respect to SF-36 scores for general health, physical functioning, physical role, bodily pain, and social functioning.
The incidence of complications during follow-up was low in both groups. No thrombo-embolic events occurred in either group and bleeding rates were similar in the two groups. Bradycardia was documented in 8.6% of patients in the drug therapy group and in none of the ablation group. No severe (>70%) PV stenosis occurred in the ablation group, one patient experiencing asymptomatic moderate (50–70%) stenosis of one vein.
The Ablation for Paroxysmal Atrial Fibrillation (APAF) trial randomized 198 patients (mean age 56 ± 10 years) with drug-refractory paroxysmal AF (mean duration 6 ± 5 years, mean AF episodes 3.4 per month) to ablation or to maximum tolerated doses of another antiarrhythmic drug.27 The ablation procedure comprised CPVA with additional linear lesions at the mitral isthmus and cavotricuspid isthmus.62,69,78,79 Patients in the ablation group were treated with antiarrhythmic drugs for 6 weeks after ablation (blanking period) to reduce the probability of early AF recurrence that could interfere with reverse remodelling. Patients in the drug therapy group received flecainide, sotalol, or amiodarone. In the event of failure of the first antiarrhythmic drug assigned, the patient was switched to one of the other drugs or to a combination of two of the three drugs authorized. After two unsuccessful trials, the patient could be considered for cross-over to CPVA. The primary endpoint was freedom from documented recurrent AT (lasting at least 30 s) during the 1 year follow-up. Recurrence was monitored by event-monitor recordings one to three times daily (with additional recordings in the event of symptoms suggestive of AF), as well as 12-lead ECGs and 48 h Holter recordings at 3, 6, and 12 months post-randomization.
In the ablation group, 86% of the patients remained free of AT during follow-up in the absence of antiarrhythmic drug therapy. Among the 11 patients with recurrent AF at the end of the blanking period, five were controlled by continuing antiarrhythmic drug therapy and six underwent a repeat procedure, five of the latter patients subsequently remaining free of recurrence during a mean post-repeat procedure follow-up of 6 months. Mean left atrial size decreased significantly (P < 0.01) between baseline and the end of follow-up in this group. In the drug therapy group, only 24% of patients remained free of recurrent AF during follow-up under treatment with the first antiarrhythmic assigned (36% of those treated with amiodarone, 21% of those receiving flecainide, and 15% of those receiving sotalol), a further 11% experiencing no further recurrences after a switch to a combination of flecainide and amiodarone. By Kaplan–Meier analysis, 86% of patients randomized to CPVA and undergoing a single procedure were free of AT at the end of follow-up without any need for continuing antiarrhythmic drug therapy, vs. 22% of patients randomized to antiarrhythmic drug therapy who did not require a second drug trial (P < 0.001).
Despite intensive transtelephonic monitoring, asymptomatic AF was recorded post-ablation in only three patients, who also reported symptomatic recurrences. In the drug therapy group, asymptomatic recurrences were recorded in 20 of the 75 patients with recurrent AF (27%). Forty-two patients with continued AF recurrence after the second drug trial crossed over to CPVA and at a mean of 6.2 months after cross-over, 36 (86%) were free of recurrent AF in the absence of antiarrhythmic drug therapy. A total of 24 hospital admissions for cardiovascular reasons were recorded in the CPVA group, including admissions for repeat procedures, compared with 167 admissions in the drug therapy group, excluding those for cross-over to CPVA (<0.001).
Procedural complications in the ablation group comprised a TIA, resolving within seconds, in one patient and a small pericardial effusion in one patient. Three patients developed post-ablation AT necessitating a repeat procedure, after which no recurrence of AT was detected. No serious complications occurred. In the drug therapy group, significant adverse events necessitating permanent drug withdrawal occurred in 23 patients (23%). These included pro-arrhythmia with flecainide in three patients, thyroid dysfunction with amiodarone in seven patients, and sexual dysfunction with sotalol in 11 patients.
Randomized comparison of ablation plus antiarrhythmic drug therapy vs. antiarrhythmic therapy alone
Stabile et al.77 investigated the benefit of catheter ablation of AF in addition to antiarrhythmic drug therapy in 137 patients intolerant of antiarrhythmic drugs or having failed at least two drug trials, randomized to the combined therapy (n = 68) or to antiarrhythmic drug treatment only (n = 69; control group). In the combined therapy group, the mean age was 62 ± 9 years, 62% were male and 68% presented paroxysmal AF, the corresponding values for the control group being 62 ± 11 years, 64% and 73%. In both groups, more than 60% of the patients had structural heart disease, but mean LVEF was over 57%. The duration of AF was significantly longer in the control group (7 ± 6 vs. 5 ± 4 years; P = 0.02), but otherwise the groups were well-balanced.
In both the groups, antiarrhythmic drug therapy principally comprised amiodarone (66% of the patients undergoing ablation, 62% of the control group), flecainide (25% and 26%, respectively), or propafenone (7% and 10%, respectively), the remaining patients receiving sotalol or disopyramide. The antiarrhythmic drug assigned was changed only in the event of atrial arrhythmia recurrence. All the patients assigned to ablation underwent a single procedure comprising CPVA with a linear lesion at the mitral isthmus and an additional linear lesion at the cavotricuspid isthmus if conduction was present. AF recurrence was assessed by daily transtelephonic monitoring during 3 months after the initial 1-month blanking period and by standard ECG and Holter monitoring at 1, 4, 7, 10, and 13 months.
Atrial arrhythmia recurrence was documented in the 12 months following the blanking period in 44% of the combined therapy group (AF in 38% of the patients, atrial flutter in 6%) compared with 91% of the control group (P < 0.001). Among the 63 patients in the control group who experienced AF recurrence, 36 (57%) underwent CPVA while continuing antiarrhythmic drug therapy, of whom 22 experienced no further recurrence of atrial arrhythmia. Overall, 30 patients (22%) of patients presented an asymptomatic recurrence of AF documented during the 3-month period of transtelephonic monitoring. In the combined therapy group, major complications occurred in three patients (4.4%) during ablation (stroke, transient phrenic paralysis, and pericardial effusion), the patient suffering from stroke dying 9 months later from a brain haemorrhage. No PV stenosis was reported. In the control group, there was one sudden death, two patients had cancer (of whom one died) and one patient experienced a TIA.
Oral et al.28 randomized 146 patients with chronic AF (mean age 57 ± 9 years) to amiodarone (200 mg/day) and up to two cardioversions during the first 3 months, alone (control group; n = 69) or in combination with CPVA, comprising electroanatomically-guided encirclement of the right and left PV with linear lesions in the posterior left atrium or roof and along the mitral isthmus (n = 77). The study population was predominantly male (88%) and included few patients with structural heart disease (8%). The mean duration of AF was 4 ± 4 years in the control group and 5 ± 4 years in the combined therapy group. All patients had failed to respond to at least two antiarrhythmic drugs and had undergone at least one cardioversion before the trial. Cardiac rhythm was assessed by transtelephonic event-monitoring with 3-min recordings at least 5 days a week and whenever symptoms were experienced throughout the 12-month follow-up, an unprecedented intensity of long-term monitoring in a trial evaluating catheter ablation of AF. Transmissions were available in all the patients for 85 ± 8% of the days of follow-up. The primary endpoint was freedom from AF or atrial flutter lasting more than 3 s during 1 year post-ablation or post-cardioversion in the control group.
In the CPVA group, AF was terminated during the procedure in 16% of patients. In the remaining patients SR was restored after ablation by administration of ibutilide or by transthoracic cardioversion. Cavotricuspid isthmus ablation to prevent atrial flutter was performed in 71% of patients. A mean of 78 ± 45 days post-ablation, transthoracic cardioversion was performed in 23% of patients because of recurrent AF in 18% and atrial flutter in 5%. The ablation procedure was repeated in 26% of patients owing to recurrent AF and in 6% because of atrial flutter, a mean of 204 ± 82 days after the first procedure. In the control group (n = 69), four patients (4%) converted to SR after the start of amiodarone treatment and in the remaining patients SR was restored by cardioversion. In total, 97% of the patients experienced recurrence of AF during the following 2 months and underwent a second cardioversion, a mean of 72 ± 25 days after the first one.
According to the intention-to-treat analysis, 74% of the patients in the CPVA group were free of AF and atrial flutter at 12 months, in the absence of antiarrhythmic drug therapy, compared with 58% in the control group (P = 0.05). However, in the control group, 77% of the patients experiencing recurrent AF crossed over to CPVA a mean of 128 ± 57 days after cardioversion. At 12 months, 70% of these patients were in SR without antiarrhythmic drug therapy. Only three out of the 69 patients in the control group (4%) were free of recurrent AF in the absence of antiarrhythmic drug therapy or CPVA (P < 0.001 for the comparison with the group that underwent CPVA). Patients who achieved freedom from recurrent AF following CPVA showed a significant reduction in mean left atrial diameter at 12 months (from 45 ± 6 mm at baseline to 40 ± 6 mm; P<0.001), as well as a significant increase in mean LVEF (from 55 ± 6% to 62 ± 8%; P < 0.001). No significant change in these parameters was seen in patients with recurrent AF.
Among patients undergoing CPVA, symptom severity score decreased significantly both in patients remaining in SR (from 17 ± 4 at baseline to 6 ± 2 points 12 months after ablation; P < 0.001) and also, to a lesser extent, in those experiencing recurrent AF or atrial flutter (from 17 ± 4 at baseline to 12 ± 4 points 12 months after ablation; P = 0.02). The improvement in symptom severity score was significantly greater in patients remaining in SR (P = 0.002), but the improvement seen in patients still experiencing recurrent AF or atrial flutter suggests partial clinical success even in that group. Apart from atypical atrial flutters in the CPVA group, no complications were observed in either treatment group.
Long-term non-randomized comparison of ablation and antiarrhythmic drug therapy
Pappone et al.69 analysed the outcomes of 1171 consecutive patients, of whom 589 underwent CPVA and 582 received antiarrhythmic drug therapy, during a median follow-up of 900 days (range 161–1508 days). The choice of therapy was determined according to patient preference or the electrophysiologist's judgement based on the number of previously ineffective drug trials and AF-related hospital admissions and the duration of antiarrhythmic drug treatment. Freedom from AF recurrence (symptomatic episodes lasting more than 10 min) during follow-up was assessed on the basis of serial 12-lead ECGs and 24 h Holter monitoring performed when symptoms occurred and routinely at 1, 3, 6, 9, and 12 months, and then every 6 months. Patients were also asked to keep a log of symptoms suggestive of AF and transtelephonic monitoring was considered to document their cause. Quality of life was measured using the SF-36 survey in 109 patients in the CPVA group and 102 of the medically treated group. In the medically treated group, all the patients were discharged from hospital in SR, achieved either spontaneously or by pharmacological or electrical cardioversion. Prophylactic antiarrhythmic therapy at discharge comprised amiodarone in 33% of patients, propafenone in 17%, flecainide in 15%, sotalol in 13%, quinidine in 9%, disopyramide in 6%, and more than one antiarrhythmic drug in 7%.
The mean age of the patients (65 ± 9 and 65 ± 10, respectively, in the ablation and antiarrhythmic drug therapy groups) was higher than in most studies investigating catheter ablation of AF and the percentage of male patients was lower (58% and 59%, respectively). Sixty-nine percent of the patients undergoing ablation and 71% of those in the antiarrhythmic drug therapy group presented with paroxysmal AF. The mean duration of AF, the mean number of drug trials, and the mean number of AF-related hospitalizations were greater in the CPVA group, but otherwise the groups were well-balanced.
The overall mortality rate was 6% in the CPVA group vs. 14% in the medically treated group; Kaplan–Meier analysis indicating a significantly higher probability of survival among patients undergoing CPVA (P < 0.001), comparable with that of healthy individuals matched for age and gender. The difference in mortality rates resulted from a lower incidence of cardiovascular deaths in the CPVA group, non-cardiovascular mortality rates being comparable between the two groups. Concurrent AF at the time of death was documented in 75% of patients dying of cardiovascular causes. In particular, 74% and 61% of deaths owing to HF and myocardial infarction, respectively, occurred in subjects with AF. Cox proportional hazards analysis revealed a hazard ratio with ablation of 0.46 (95%CI 0.31–0.68; P < 0.001) for all-cause mortality, 0.45 (95%CI 0.31–0.64; P < 0.001) for morbidities, mainly because of HF and ischaemic cerebrovascular events, and 0.30 (95%CI 0.24–0.37; P < 0.001) for AF recurrence. Both the number of adverse events and the number of patients experiencing an adverse event were lower in patients undergoing CPVA than in those receiving antiarrhythmic drug therapy.
Kaplan–Meier estimates of the percentages of patients remaining free of AF recurrence at 1, 2, and 3 years were 84%, 79%, and 78%, respectively, in the CPVA group and 61%, 47%, and 37%, respectively, in the medically treated group—the difference between the two groups increasing with time. Among patients presenting with AF recurrence, the rate of development of persistent AF was lower in the CPVA group (26 vs. 65%). Maintenance of SR, as a time-dependent variable, was associated with significantly lower mortality and adverse event rates both in the study population as a whole and in each treatment group. Quality of life (SF-36 physical and mental component scores) improved over time (P = 0.007) only in patients undergoing CPVA, reaching normative levels at 6 months. AF recurrence was independently associated with impaired psychological well-being in the CPVA group and with physical and mental functioning scores in the medically treated group (P < 0.01).
Meta-analysis of trials comparing catheter ablation with antiarrhythmic drug therapy for atrial fibrillation
Noheria et al.80 performed a meta-analysis on published randomized clinical trials comparing the outcome of patients undergoing catheter ablation for AF with that of patients treated solely with antiarrhythmic drug therapy and permitting analysis of survival free of AT recurrence in the 12 months following catheter ablation or the start of antiarrhythmic drug therapy, the minimum follow-up duration recommended in the HRS/EHRA/ECAS expert consensus statement.38 Four trials meeting these selection criteria were identified, namely those published by Wazni et al.,26 Stabile et al.,77 and Pappone et al.27 described above and the study reported by Krittayaphong et al.,81 comparing amiodarone treatment and radiofrequency ablation (PV isolation with additional linear lesions) in 30 patients with chronic, drug-refractory AF. In this study, the probability of AF-free survival at 1 year was 79% in the ablation group vs. 40% in the amiodarone group (P = 0.018). A significant beneficial effect of treatment on symptoms and quality of life was seen in the ablation group, but not in the amiodarone group. One patient undergoing ablation suffered an ischaemic stroke, adverse events being reported by 47% of patients in the amiodarone group and prompting discontinuation of treatment in one patient.
Analysis of the combined trial data showed that of the 214 patients undergoing catheter ablation, 162 (76%) achieved AT recurrence-free survival during the 12-month follow-up period, compared with 41 (19%) of the 218 patients receiving antiarrhythmic drug therapy. The random-effects pooled estimate for the risk ratio for AT recurrence-free survival was 3.73 (95%CI: 2.47–5.63; P < 0.001). The statistical test for heterogeneity between the studies was not significant (P = 0.13). The rate of adverse events was significantly higher in the antiarrhythmic drug therapy group than in the catheter ablation group (P = 0.02). However, some of the adverse events occurring in the latter group, e.g. stroke, were considerably more severe than those experienced by patients in the antiarrhythmic drug therapy group. Two patients died during the 12-month follow-up, one in each treatment group.
Despite the statistically significantly better AT recurrence-free survival with CPVA compared with antiarrhythmic drug therapy revealed by this meta-analysis, the authors highlighted the paucity of clinical trials comparing catheter ablation with antiarrhythmic drug therapy and the need for larger randomized studies to evaluate the strategy of proposing CPVA as a first-line therapy for carefully selected patients with AF.
Comparative costs of catheter ablation and pharmacological therapy for atrial fibrillation
Retrospective analysis of comparative treatment costs for paroxysmal atrial fibrillation in France
Weerasooriya et al.82 performed a retrospective cost comparison of catheter ablation vs. drug therapy for paroxysmal AF based on the experience of a pioneering French ablation centre. The study population comprised 118 consecutive patients with symptomatic, drug-refractory paroxysmal AF (mean age 52 ± 18 years; 70% men) undergoing ablation in the authors' institution.
The cost estimate of the ablation procedure included hospitalization (5 days) and procedural costs, obtained from the hospital billing data. Medical treatment costs were assessed by reviewing the records of the last 20 consecutive patients undergoing ablation and contacting these patients to ascertain the antiarrhythmic treatment used just before ablation, the frequency of symptoms before ablation, and the frequency of emergency room visits, doctor's office visits and hospital admissions. The cost of drug treatment was estimated on the basis of the antiarrhythmic medication(s) used immediately prior to ablation, assuming maintenance of a fixed dose and combination of medications during the 12 months preceding ablation. Hospitalization costs were estimated from the daily bed charge at the authors' institution and consultation costs from the tariffs and reimbursement rates of the French national health insurance system. It was assumed that patients with an unsuccessful outcome of catheter ablation had the same event rates after the procedure as before and incurred the same long-term drug treatment costs. In the sensitivity analysis, the long-term pharmacological treatment costs (for patients whose AF was treated medically and for those with an unsuccessful outcome of ablation) were those calculated in the French ‘Cost of Care in Atrial Fibrillation’ (COCAF) study.83 Projected costs following ablation were calculated by adding the initial procedural cost to the annual cost of care discounted at 5% per year. Projected costs of drug treatment were similarly discounted.
The patients studied underwent a mean of 1.52 ± 0.71 ablation procedures (range 1–4), the mean duration of the procedure being 182 ± 33 min. No significant complications occurred and all PV were successfully isolated in all the patients. During a follow-up of 32 ± 15 weeks, 72% of the patients remained free of clinical recurrence in the absence of antiarrhythmic drugs. The initial cost of ablation, based on resource use at 2001 values, was €4715 and the follow-up cost was €445 per year, giving a total cost during the first year of €5160 compared with €1590 for pharmacological treatment. The comparison favoured ablation over pharmacological treatment starting from 5 years after the procedure onwards, the projected cumulative costs of the two treatment strategies being €6730 for ablation and €7194 for medical therapy at 5 years and diverging increasingly thereafter. The cost of ablation was 3.24 times that of medical therapy in the first year, but by 5 years the projected cost ratio had decreased to 0.93 and by 10 years to 0.65. The sensitivity analysis using the COCAF calculation of pharmacological therapy costs produced similar results, the cumulative treatment costs favouring ablation from the fourth year onwards.
The authors note that a limitation of this study is the absence of long-term data on the success rate of the ablation procedure, the 72% success rate used in the cost estimates being based on a mean follow-up of 32 weeks and assumed to continue unchanged during the following years. However, as the patients included in the study had at least monthly symptoms prior to ablation, despite multiple antiarrhythmic drugs, they considered that disappearance of symptoms during the 32-week follow-up was a reasonable measure of long-term success in this population.
Simulation of comparative treatment costs for paroxysmal atrial fibrillation in Canada
Khaykin et al.84 compared the costs related to catheter ablation and medical therapy for paroxysmal AF in Canada (Ontario) based on data from the Canadian Registry of Atrial Fibrillation (CARAF), government fee schedules and published reports, and standard practices for ablation in the electrophysiology centres in Ontario.
The costs taken into account for catheter ablation comprised the cost of ablation tools, hospital and physician bills, anticoagulation and costs associated with periprocedural medical care and complications (tamponade, stroke, TIA, and PV stenosis). The costs of ablation tools were based on the use of the procedures described by Pappone et al.45 and Verma et al.46
In accordance with the reported success rates obtained with these approaches, a sensitivity analysis assuming success rates ranging from 50% to 75% after the initial ablation procedure, 60–85% after two procedures, and 70–90% after three procedures was performed. This analysis also took into account late success attrition rates from 1% to 5% per year projected for 5 years after the first procedure and a range of catheter costs based on different ablation techniques. It was assumed that all the patients would receive anticoagulant therapy for 1 month before and 3 months after ablation, according to local practice, irrespective of whether they would otherwise require anticoagulation according to current guidelines. It was also assumed that the cost of follow-up in successfully ablated patients would equal that of medically treated patients for 3 months following ablation (the usual ‘blanking period’) and would then be limited to one AF-related physician visit per year. Costs related to medical therapy included anticoagulation, rate- and rhythm-control medications, non-invasive testing, physician follow-up visits, hospital admissions and costs incurred as a result of complications. The prevalence of chronic anticoagulation both post-ablation and in the context of medical rate- or rhythm-control therapy was assumed to equal its prevalence in the overall population with AF according to CARAF data (56%).
The total per-patient cost estimate for AF ablation ranged from $16 278 to $21 294 (median $18 151), with follow-up costs ranging from $1597 to $2132 per year. The annual per-patient cost of medical therapy was estimated to range from $4176 to $5060 (median $4840). Using these estimates, the costs for catheter ablation and chronic medical therapy for AF would meet at 3.2–8.4 years, which the authors considered would make catheter ablation for AF a fiscally reasonable alternative to continuing medical therapy. They noted, however, that many physicians do not refer eligible patients for this procedure in view of the 6- to 12-month waiting list in most centres in Ontario.
Decision-model analysis of comparative treatment costs according to stroke risk in the USA
Chan et al.85 developed a Markov decision analysis model to evaluate the cost-effectiveness of the LACA procedure, comprising encirclement of the PV 1–2 cm from the ostia with additional linear lesions in the left atrium63,70 in comparison with rhythm-control with amiodarone and medical rate-control therapy. The model was applied to three hypothetical cohorts: patients aged 65 years with moderate- and low-stroke risks, respectively, and patients aged 55 years with a moderate-stroke risk. Patients with moderate-stroke risk were defined as having one risk factor [hypertension, diabetes mellitus, coronary artery disease, or congestive heart failure (CHF)], whereas patients at low-stroke risk had no risk factor. The choice of 55 and 65 years for the two age groups was based on the mean age of patients undergoing ablation and the age at which the prevalence of AF begins to increase significantly.
A success rate of 80% for the LACA procedure (with a 30% rate of second procedure during the first year) and an annual relapse rate back to AF of 2% were assumed for the base analysis. Antithrombotic or anticoagulant treatment was assumed in all treatment arms, warfarin in patients with a moderate stroke risk, and warfarin or aspirin in those with a low-stroke risk. Patients achieving SR were assumed to continue warfarin for 6 months before switching to aspirin. Costs were calculated in 2004 US dollars, and costs and life expectancy were discounted at 3% per year. Baseline annual stroke risks were considered to be 3.0% and 1.4%, respectively, in patients at moderate- and low-risk without anticoagulation, 2.3% and 1.1%, respectively, with aspirin therapy, and 1.3% and 0.7% with warfarin.
The base-case analysis indicated the greatest number of quality-adjusted life years (QALYs) with LACA (11.06) compared with rate-control (10.81) and rhythm-control (10.75) in patients aged 65 years at moderate risk of stroke, but represented a higher total cost ($52 369) than the pharmacological rate- and rhythm-control ($39 391 and $43 358, respectively). The incremental cost-effectiveness (ICER) of LACA compared with rate-control (i.e. the difference in cost divided by the difference in QALYs achieved) was $51 800 per QALY in this population and $28 700 per QALY in patients aged 55 years with a moderate risk of stroke, but $98 900 per QALY in patients with a low-risk of stroke. Traditionally, the upper limit of ICER for a therapy to be considered cost-effective relative to a reference therapy is $50 000. On this basis, the authors considered that LACA may be cost-effective in patients with moderate stroke risk, particularly younger patients with a longer life expectancy, provided that this procedure is not associated with long-term adverse effects or a high late-recurrence rate. In contrast, it is unlikely to be cost-effective in patients with a low risk of stroke.
Limitations of this analysis include the assumption that SR attained by catheter ablation prevents stroke and other embolic complications, for which there is little evidence to date, the assumption of an 80% success rate for LACA, which may be expected in younger patients with paroxysmal AF and no structural heart disease, but is probably an overestimate for older patients with persistent and chronic AF, and finally the current paucity of data concerning long-term outcomes of catheter ablation.
In conclusion, the results of these three studies suggest that catheter ablation may be a cost-effective strategy, particularly in the relatively young patients with paroxysmal AF currently forming the majority of those undergoing this procedure. However, as information on actual costs and long-term outcomes are currently limited, these analyses necessarily involve reliance on estimates, assumptions, and retrospective data and their results need to be confirmed by long-term prospective randomized trials. The initial cost of catheter ablation for AF is high, and its cost-effectiveness relative to alternative therapies in different patient populations will depend greatly on the success rates achieved and the need for repeat procedures, as well as on the long-term impact of SR restoration on morbidity and mortality.
For which patient subgroups can catheter ablation be considered a reasonable first-line option?
Catheter ablation for AF as a first-line treatment option seems to be most justifiable in two main patient subsets:
Patients with clearly focally triggered AF with constant staccato firing, in whom a limited ablation procedure can be envisaged;
Patients with contraindications to, or a limited choice of antiarrhythmic drugs, for example, those with HF, coronary disease, etc.
Verma and Natale86 suggested that patients with symptomatic paroxysmal or persistent AF and mild to moderate structural heart disease might be considered for first-line catheter ablation and that this might particularly benefit younger patients with lone AF for whom antiarrrhythmic drug therapy and anticoagulation over a very long period presents particularly high risks and lifestyle costs. Padanilam and Prystowsky87 considered that patients with highly symptomatic AF who refuse antiarrhythmic therapy or for whom the only choice of antiarrhythmic agent is long-term amiodarone, those at high risk for stroke who refuse or cannot tolerate long-term warfarin therapy and young patients with paroxysmal AF and sinus node dysfunction who might otherwise require insertion of a pacemaker might be appropriate candidates for such a treatment strategy. Our proposal of an electrophysiological criterion for envisaging catheter ablation as a first-line treatment option takes into account procedural duration and complexity, as well as reservations concerning the possible long-term effects of extensive left atrial lesions, which are as yet largely unknown.
In all cases, treatable causes of AF, such as obesity, sport, high alcohol consumption, disordered sleep breathing, etc., should be sought and targeted before envisaging an ablation procedure. Techniques of ablation are evolving rapidly and this may also be a factor in deciding whether or not ablation should be proposed as first-line therapy or delayed until after other treatments have been tried.
Catheter ablation as a first-line treatment of AF should only be envisaged by centres with considerable experience in performing AF ablation in which the benefit–risk ratio of this approach is likely to be optimal.
Catheter ablation of atrial fibrillation in patients with congestive heart failure
Background
Treatment of AF in patients also suffering from CHF is currently a major focus of attention. These two common cardiac disorders are inextricably linked and are frequently encountered together, patients with one of these disorders having a substantially greater risk of developing the other. Both conditions show an exponential increase in incidence and prevalence with advancing age and are associated with higher mortality, more frequent hospitalizations, decreased quality of life, and substantial healthcare costs.88 In the Framingham Heart Study, 26% of the 921 subjects diagnosed with AF had a previous or concurrent diagnosis of CHF and 16% later developed CHF. Similarly, of the 931 subjects diagnosed with CHF, 24% had a previous or concurrent diagnosis of AF and 17% developed AF subsequently.89 AF preceded CHF with approximately the same frequency that CHF preceded AF, and the cumulative incidence of first CHF in subjects with AF and vice versa increased steadily over time. Clinical trial data indicate that the prevalence of AF in the context of CHF is related to the severity of the latter pathology, ranging from 10% to 20% in patients with mild to moderate CHF up to 50% in patients with more advanced disease90 (Figure 3).
Prevalence of atrial fibrillation in clinical studies of congestive heart failure. (Adapted from Savelieva and Camm.90)
The adverse haemodynamic consequences of AF, including chronically elevated cardiac filling pressures, irregular ventricular intervals, and lack of effective atrial contraction and atrioventricular synchrony, can impair ventricular function and contribute to symptoms of CHF. Similarly, elevated cardiac filling pressures, dysregulation of intracellular calcium, and autonomic and neuroendocrine dysfunction associated with CHF can markedly increase the risk of development of AF.91 In particular, long-term up-regulation of the renin-angiotensin-aldosterone system, resulting in salt and water retention by the kidneys, vasoconstriction and structural remodelling, including fibrosis of the atrial and ventricular myocardium, leads to a self-perpetuating cycle of progression of both the disorders92 (Figure 4).
Pathophysiology of atrial fibrillation (AF) and congestive heart failure (CHF). Upregulation of the renin-angiotensin-aldosterone system (RAAS) leads to salt and water retention, increasing left atrial pressure and stretch, as well as atrial and ventricular myocardial fibrosis and remodelling. This produces a detrimental positive-feedback loop, worsening CHF and predisposing to AF, thereby further activating the RAAS system. (Adapted from Heist and Ruskin.92)
AF is more likely to precipitate CHF in patients with underlying LV dysfunction or cardiac hypertrophy leading to reduced compliance. However, symptoms of HF may also be triggered by rapid and irregular ventricular activity even in patients with little structural heart disease. The risk of tachycardia-induced cardiomyopathy is greatest in patients with persistent AF, particularly those with sustained ventricular rates greater than 120 b.p.m. LV dysfunction caused by AF in the absence of significant underlying heart disease may completely reverse after the restoration of SR or control of ventricular rate.90
An electrophysiological and electroanatomic mapping study, comparing 21 patients with symptomatic CHF (LVEF 26 ± 6%) and age-matched controls with normal LVEF, showed clear atrial remodelling owing to CHF. This was characterized by increased atrial effective refractory period, increased atrial conduction time along the lateral right atrium and the coronary sinus, prolongation of P-wave duration and corrected sinus node recovery times, and a larger number of double potentials along the crista terminalis, as well as slowing of regional conduction and a greater number of areas with fractionated electrograms. These abnormalities may contribute to the increased propensity for AF in patients suffering from CHF.93
Most large studies have indicated a worse prognosis of AF and CHF combined than that of either disorder alone, although there is some controversy as to whether the presence of AF adversely affects prognosis in the context of CHF.88,92 A retrospective analysis of data on 6517 patients with left ventricular (LV) systolic dysfunction (LVEF ≤35%) included in the SOLVD (Studies Of Left Ventricular Dysfunction) prevention and treatment trials, followed-up for an average of almost 3 years, showed a significantly higher all-cause mortality in the 419 patients with AF at baseline than in those with SR at baseline (34 vs. 23%; P < 0.001), independent of age, LVEF, NYHA functional class, and medication use. Multivariate analysis showed that the relative risk (RR) of death for patients with AF when compared to those with SR was 1.34 (95%CI 1.12–1.62; P = 0.002). AF was also significantly associated with the composite endpoint of death or hospitalization for CHF (RR 1.21; 95%CI 1.03–1.42; P = 0.02).94
Analysis of data from the DIG (Digitalis Investigators Group) trial, including 7788 patients with CHF followed-up for a mean of 37 months, also demonstrated increased mortality among the 866 patients presenting supraventricular tachyarrhythmias, including AF, during the study period (43 vs. 32%). After adjustment for other risk factors, the development of supraventricular tachyarrhythmias independently predicted a higher risk of all-cause mortality (RR = 2.45; 95%CI 2.19–2.74; P = 0.0001), stroke (RR 2.35; 95%CI 1.68–3.29; P = 0.0001), and hospitalization for worsening HF (RR 3.00; 95%CI 2.71–3.33; P = 0.0001).95
In a study of 390 consecutive patients with NYHA functional class III–IV HF and a mean LVEF of 0.19, of whom 75 (19%) had AF (paroxysmal in 26 patients, chronic in 49), this disorder was found to be an independent predictor of both sudden death and all-cause mortality.96 Actuarial 1-year overall survival was 52% in patients with AF vs. 71% in those with SR (P = 0.0013), sudden death-free survival being 69 vs. 82% (P = 0.0013). Subgroup analysis revealed AF to be predictive of both all-cause and sudden death in patients with a pulmonary capillary wedge pressure <16 mmHg on therapy, but not in those with elevated pulmonary capillary wedge pressure. AF was also shown to be an independent risk factor for mortality (RR 1.46; 95%CI 1.22–1.75; P = 0.0001) in a study of 651 elderly patients with CHF after prior myocardial infarction, of whom 296 had a normal LVEF (≥50%) and 355 had an abnormal LVEF. In the group of patients with normal LVEF, higher mortality rates among patients with AF were seen at both 6-month follow-up (11 vs. 2% among patients with SR; P = 0.0005) and at long-term follow-up (93% at a mean follow-up of 27 ± 18 months vs. 74% at a mean follow-up of 37 ± 21 months; P = 0.0001).97
In contrast, the results of the Vasodilator in Heart Failure Trial (V-HeFT), which followed-up 1427 patients with predominantly NYHA functional class II–III HF for a mean of 2.5 years, did not reveal any significant increase in sudden death or all-cause mortality in 206 patients with AF compared to those with SR.98 Crijns et al. followed-up 409 patients with functional class III–IV HF for a mean of 3.4 years, and showed a numerically higher all-cause mortality among patients with AF vs. those with SR (60 vs. 47%; P = 0.04). However, the difference was no longer statistically significant after adjustment for age, LVEF, NYHA functional class, renal function, and blood pressure.99 Several other smaller studies also failed to demonstrate an independent prognostic significance of AF with regard to mortality in patients with HF,100–102 but these studies were not powered to detect small or moderate effects of AF on mortality.88
Restoration of SR in patients presenting both AF and CHF is particularly challenging in view of the demonstrated association between the use of class I antiarrhythmic agents and increased mortality in this population, particularly in patients presenting with ischaemic heart disease.29 Furthermore, these agents appear to increase clinical CHF in patients with impaired LV systolic function, possibly by further worsening LV function and haemodynamics, and are also relatively ineffective in suppressing AF in this population.103–105 The class III antiarrhythmic agents amiodarone24 and dofetilide19 have been shown to be effective in suppressing AF in patients with CHF. However, long-term use of amiodarone is limited by the risk of multiple organ toxicity and dofetilide can lead to marked QTc prolongation and potentially to life-threatening ventricular arrhythmias, including torsade de pointes, particularly in the elderly, in women, and in the setting of LV dysfunction.92 Azimilide has also demonstrated effectiveness in maintaining SR in patients with AF and LV dysfunction with a neutral effect on mortality.106
Catheter ablation may prove to be a valuable treatment option in this context and encouraging outcome data have been obtained in small studies.
Clinical trial data on catheter ablation of atrial fibrillation associated with congestive heart failure
Comparison of ablation outcomes in patients with congestive heart failure and matched controls
Hsu et al.107 compared the outcome of 58 consecutive patients with CHF and a LVEF of <45% who were undergoing catheter ablation for the treatment of AF resistant to at least two antiarrhythmic drugs (mean age 56 ± 10 years, 88% male), to that of a control group of patients without CHF, matched for age, gender, and classification of AF, who were similarly undergoing ablation. In both groups, 74% of the patients presented permanent AF, 17% persistent AF, and 9% paroxysmal AF. At baseline, ventricular rate was adequately controlled in 41% of the patients with persistent or permanent AF.
The ablation procedure was based on electrical isolation of the PVs44 in all patients, with additional left atrial linear ablation between the two superior PVs and/or extending from a PV to the mitral annulus53 in over 90% of the patients. Outcome was evaluated in terms of LV function and dimensions, symptom score, exercise capacity, and quality of life; all these parameters were measured at baseline and at 1, 3, 6, and 12 months after the ablation procedure. Recurrence of AF was assessed by 48 h ambulatory monitoring 1, 3, 6, and 12 months after the last ablation procedure. Anticoagulation therapy was stopped if SR had been maintained for 3–6 months, unless otherwise indicated.
At baseline, all the patients with CHF had symptoms in NYHA class II or higher, despite treatment with angiotensin-converting enzyme-inhibitors or angiotensin II-receptor blockers in 72% of patients, beta-blockers in 97%, and digoxin in 29%; 16% of patients had experienced at least one episode of class IV symptoms within the previous 6 months. The great majority of patients (91%) had persistent or permanent AF, 66% had a LVEF below 40%, and 45% had concurrent structural heart disease. Three patients were referred for catheter ablation to ameliorate symptoms while waiting for a heart transplant.
Following the initial ablation procedure, early recurrence of AF or atrial flutter led to a second ablation procedure in 50% of the patients with CHF and 47% of the control group. During a mean follow-up of 12 ± 7 months (range 3–34 months) after the final ablation procedure, 78% of patients with CHF, and 84% of those in the control group were in SR (P = 0.34), 69 and 71%, respectively, maintaining SR in the absence of antiarrhythmic drug therapy. In the CHF group, NYHA class improved from a mean of 2.3 ± 0.5 before ablation to 1.4 ± 0.5 at 1 month, remaining close to this level at 12 months (P < 0.001). No changes were seen in the control group. The condition of two of the patients awaiting transplantation improved sufficiently for them to be taken off the transplant list (symptoms improved by one NYHA class and LVEF increased by 8 and 12%, respectively). The third patient on the transplant list experienced a recurrence of AF 1 month after ablation and died 3 months later from worsening CHF.
Catheter ablation of AF was associated with marked improvement in LV function and dimensions in patients with CHF, the greatest improvement in LV function being seen within the first 3 months (Figure 5). LVEF increased by a mean of 21 ± 13% and LV shortening by 11.7% (P < 0.001 for both comparisons), a marked increase in LVEF (defined as an increase of 20% or more or to a value of 55% or more) being seen in 72% of the patients. Subgroup analysis revealed significant (P < 0.001) increases in LVEF even in patients with adequate ventricular rate control prior to ablation (mean increase 17 ± 15%), as well as in those with poor rate control (mean increase 23 ± 10%) Marked improvement in LVEF was seen in 86% of patients with poor rate control pre-ablation compared with 54% of those with adequate rate control (P = 0.02), the latter group notably comprising a higher percentage of patients with concurrent structural heart disease (75 vs. 21%). The greatest improvement in LVEF (24 ± 8%) was observed in patients with poor rate control pre-ablation and no structural heart disease. LV function also improved significantly in four patients whose permanent AF had been converted to paroxysmal AF by ablation.
Improvement in left ventricular (LV) ejection fraction (A) and LV fractional shortening (B) after ablation in patients with congestive heart failure. Number of patients at each time point—0: 58; 1 month: 55; 3 months: 48; 6 months: 40; 12 months: 34. (Adapted from Hsu et al.107)
Symptom Check List (SCL) and quality of life (SF-36) scores improved significantly in both the groups. SF-36 summary scores for the physical and mental components increased by a mean of 24 ± 21 points (P < 0.001) and 21 ± 19 points (P < 0.001), respectively, in the CHF group and by 18 ± 17 points (P = 0.003) and 14 ± 19 points (P = 0.004), respectively, in the control group. Statistically significant (P ≤ 0.001) increases in mean exercise time and mean exercise capacity compared with baseline were noted in both the groups.
The incidence of peri-procedural complications was low in both the groups. Pericardial tamponade requiring percutaneous drainage occurred in one patient in each group and one patient with CHF experienced a stroke during the procedure. Repeat ablations were performed in a similar percentage of patients in the CHF and control groups, and the total overall duration of the ablation procedure and the durations of radiofrequency ablation and fluoroscopy did not differ between the two groups.
Comparison of ablation outcomes in consecutive patients with and without left ventricular dysfunction
Chen et al.108 analysed the outcomes of PV isolation in a cohort of 377 consecutive patients (mean age 55 years, 80% male) of whom 94 (25%) had impaired LV function (LVEF <40%) and 283 had normal LV function. Selection criteria for ablation in this centre were symptomatic AF, refractoriness to antiarrhythmic drug therapy, and no indication for open-heart surgery. No patient was considered unsuitable for ablation on the grounds of low LVEF. Among patients with LV systolic dysfunction, this was predominantly caused by ischaemic heart disease (78%), valvular heart disease (16%), idiopathic cardiomyopathy (4%), and hypertensive heart disease (2%). Mean LVEF in this group was 36 ± 8%, 30% of the patients were in NYHA functional class II, 68% in class III, and 2% in class IV; 32% had an implantable cardioverter-defibrillator and 10% had a pacemaker.
PV isolation (with an endpoint of abolition of all PV potentials) was not accompanied by linear lesions from the PV to the mitral annulus in any patient. However, 10% of patients with normal LV function and 7% with impaired LV function underwent cavotricuspid isthmus ablation during the PV isolation procedure for concomitant atrial flutter and a further 7% and 4% of patients, respectively, had undergone cavotricuspid isthmus ablation in a previous procedure. PV isolation was guided by angiography in the first 53 of the 337 consecutive patients included in this analysis and by intracardiac echography in the remaining patients. Recurrence of AF was assessed by loop event-recorder 1 month after ablation, 24 h Holter monitoring at 3, 6, and 12 months, and 12-lead electrocardiography at 2, 3, 6, and 12 months. Additional monitoring with a loop event-recorder or Holter monitor was considered if symptoms suggestive of AF occurred. Recurrence of AF was also confirmed by interrogation of implanted devices.
After a mean follow-up of 14 ± 5 months, the incidence of AF recurrence was 27% in patients with impaired LV function compared with 13% in those with normal LV function (P = 0.03). The remaining patients in the two groups remained free of AF without antiarrhythmic drug therapy. Among the 25 patients with LV dysfunction experiencing recurrence of AF, this was controlled by previously ineffective antiarrhythmic drugs in three patients and eliminated by a second PV isolation procedure in 21 patients. Another patient presented fewer and shorter AF episodes after ablation. AF was consequently eliminated in 96% of patients with impaired LV function, after a second procedure in 22% of the patients. Overall, mean LVEF increased non-significantly from 36% before ablation to 41% after this in the group of patients with LV dysfunction, 60% of these patients showing an increase in LVEF (mean increase 7.2 ± 3%), the remaining patients experiencing no change in LVEF or in a few cases a minimal decline. Among the patients completing the SF-36 quality-of-life questionnaire before and 6 months after ablation (46 and 53%, respectively, of those with impaired and normal LVEF), general health, energy, physical functioning and emotional well-being, and social functioning scores were significantly increased after the procedure, to a similar extent in the two groups.
The incidence of complications was low and similar in the two groups: periprocedural stroke in two (2%) of the patients with impaired LVEF and in three (1%) of the patients with normal LVEF, severe PV stenosis in one (1%), and five (2%) of the patients, respectively, and pulmonary oedema leading to premature termination of the first procedure in one patient with impaired LVEF.
Tondo et al.109 compared the outcome of 105 consecutive patients with (mean age 57 ± 10 years, 70% men) and without (mean age 56 ± >8 years, 89% men) LV dysfunction (LVEF <40%) and a definite diagnosis of CHF undergoing CPVA with the additional linear lesions at the mitral isthmus and the tricuspid isthmus. Among patients with LV dysfunction (38% of the cohort), mean LVEF was 33 ± 2%, 25% of the patients had mitral valve disease, 25% ischaemic/hypertensive disease, 5% hypertrophic cardiomyopathy, and 45% had idiopathic dilated cardiomyopathy. All the patients in this group exhibited dyspnoea on exertion, markedly limiting their daily activities (mean NYHA functional class 2.8 ± 0.1) and 75% of the patients presented with persistent AF. All the patients had optimal treatment for CHF, but several drug trials had failed to maintain SR or provide effective ventricular rate control. Arrhythmia recurrence was assessed by continuous telemetric ECG monitoring for the first 72 h, then by transtelephonic monitoring three times a week and whenever symptoms were perceived, as well as by 12-lead ECGs and 24 h Holter monitoring at 1, 3, 6, and 12 months.
During the first 48 h post-ablation, eight (20%) patients with LV dysfunction experienced AF or atypical atrial flutter, of whom seven required external cardioversion to restore SR, compared with 15% of the patients with normal LV function (difference non-significant). Repeat ablation was required for 13 patients in the group with LV dysfunction (10 for AF, three for atrial flutter) and 14 patients in the group with normal LV function (seven for AF, seven for atrial flutter).
During a mean follow-up of 14 ± 2 months post-ablation, 87% of patients with LV dysfunction were in stable SR (50% in the absence of antiarrhythmic drug therapy, 37% with previously ineffective drug therapy). Mean LVEF was significantly improved in these patients, from 33 ± 2% at baseline to 47 ± 2% (P < 0.01), mean fractional shortening increasing from 19 ± 4% to 30 ± 3% (P < 0.01). Mean exercise time was also significantly increased (P < 0.001). Among patients with normal LV function, 92% were in stable SR (69% in the absence of antiarrhythmic drug therapy). All the patients with stable SR, irrespective of baseline LV function, reported marked improvement in general health, physical functioning and emotional wellbeing (SF-36 scores), improvement from baseline being statistically significant (P < 0.05).
Two patients in the group with LV dysfunction experienced procedural complications (AV fistula) and four patients in the group with normal LV function (one case of tamponade and three cases of AV fistula). No stroke event occurred in either group.
In which patients with congestive heart failure is ablation likely to be most beneficial?
The results of the studies described above indicate that catheter ablation of AF in patients with HF can be particularly beneficial, restoration of SR leading to improvement in LV function and quality-of-life even in patients with adequate rate control prior to ablation.107 Since a reduced LVEF is an important predictor of mortality,110 the significant improvement in LV function after ablation could lead to enhanced survival.
In the study reported by Hsu et al.,107 the presence of concurrent structural heart disease in patients with impaired LV function did not significantly affect the outcome of ablation with regard to restoration and maintenance of SR, but a marked increase in LVEF (by 20% or more or to a value of 55% or more) was achieved in a significantly higher percentage of patients without concurrent structural heart disease (88 vs. 54%; P = 0.007), consistent with historical observations on the benefits of restoring SR in patients with tachycardia-mediated cardiomyopathy. The greatest improvement in LVEF (24 ± 8%) was seen in the subgroup of patients with poor rate-control pre-ablation and with no co-existing heart disease. In this group, 92% of the patients showed marked improvement in LVEF after ablation of AF. In the group of patients with impaired LV function studied by Chen et al.,108 no correlation was detected between improvement in LVEF and the type of structural heart disease, age, type, or duration of AF, or left atrial size.
Patients without severe congestive features of HF, in whom long-term (e.g. 2-year) AF precedes the onset of CHF, i.e. those in whom tachycardia-induced cardiomyopathy is a major factor in the development of CHF, may be considered most likely to gain benefit from AF ablation, even if rate control is successful but irregularity of heart rhythm persists. Conversely, in patients with class III–IV HF who subsequently develop AF, this may contribute little to the manifestations of HF, and ablation is less likely to result in improved LV function. Improvement in HF symptoms by amiodarone would increase confidence in the likely benefit of AF ablation in patients with CHF, a poor response to amiodarone possibly predicting a less successful outcome.
PV isolation is technically more challenging in patients with impaired LV function, because of the elevated filling pressures and larger PV ostia, resulting in the necessity for more extensive ablation. In the study reported by Chen et al.,108 the mean ostial diameter of the four PVs in patients with LV dysfunction was 2.2 vs. 1.4 cm in patients with normal LV function (P < 0.05). Patients with LV dysfunction may also have hypertrophy of the atrial muscle which could result in thickening of the areas targeted for ablation.108 Nevertheless, studies comparing peri-procedural complications in patients with and without impaired LV function,107–109 did not detect any marked differences in the rate and type of complications occurring. In the study reported by Hsu et al.,107 the duration of the ablation procedure as a whole, the total duration of radiofrequency application, and the duration of fluoroscopy did not differ significantly between the groups of patients with and without LV dysfunction.
In conclusion, the limited data currently available indicate the safety of ablation to treat AF in patients with CHF and its efficacy with regard to both restoration of SR and improvement of HF symptoms. However, the processes underlying CHF and the co-existing morbidities vary from one patient to another and the factors predictive of a successful outcome of ablation in terms of AF recurrence, LV function, and quality-of-life have not yet been fully defined. Reported studies assessing the benefits and risks of ablation in patients with CHF have been confined to relatively young patients (mean age below 60 years) and have included few women (12–30% of the total study population). Randomized prospective trials on larger patient populations and with longer follow-up durations are required to confirm the benefits and risks of AF ablation in patients presenting both AF and CHF and to define the patient subsets for whom this procedure is most strongly indicated and is more appropriate than atrioventricular junction ablation and pacemaker implantation.
Anticoagulation after catheter ablation for atrial fibrillation
One of the key questions is when and in whom anticoagulation can be stopped after catheter ablation for AF. Although there is a fairly broad consensus about stopping anticoagulation in low-risk patients and not stopping it in high-risk patients, there are many patients who fit into a grey zone where the optimal strategy is unclear. Which patients make up the subsets in which anticoagulation can be safely stopped and what are their clinical features? How can we be sure that a patient is really in SR when most patients undergoing ablation do not have implantable devices? What is the appropriate timing of warfarin discontinuation and what sort of long-term follow-up protocol is optimal for ensuring that a decision made at 3 months or 6 months is in fact safe and effective for that patient 1, 2, or 5 years after the procedure? As yet there are no clear answers to these questions. If a more convenient and safe anticoagulant therapy, with equivalent efficacy to warfarin were available, this might not be an issue, but for many patients, especially young patients, long-term warfarin therapy is perceived as burdensome.
By restoring SR in the majority of patients, LACA may reduce the risk of thrombo-embolic events. However, patients with a history of AF and co-morbidities may still be at risk of these events even when in SR.20,111 When AF appears to be completely suppressed or cured, based on the absence of clinical symptoms, and the presence of SR is documented by repeated intermittent monitoring, anticoagulation may be stopped to avoid exposing the patient unnecessarily to the risks and burden of warfarin therapy. Yet, available evidence suggests that continued episodes of asymptomatic AF are no less likely to be associated with a risk of stroke than symptomatic episodes,14 and asymptomatic AF is difficult to monitor continuously in the long-term other than with implantable devices. Furthermore, up to 30% reductions in left atrial transport have been reported following ablation in the left atria,112 and this may predispose patients to thrombo-embolic events even if SR is restored and pre-ablation risk of thrombo-embolic events was low.113
Analysis of the relationships between SR, treatment and survival in the large AFFIRM trial20 indicated that use of warfarin was as strongly associated with a decreased risk of death (HR = 0.50) as SR (HR = 0.53) and in the absence of long-term monitoring of asymptomatic AF recurrences, it may not be advisable to discontinue warfarin treatment even after apparent restoration of stable SR. Measurement of symptomatic and asymptomatic recurrence of AF in 19 consecutive patients with highly symptomatic drug-refractory AF undergoing catheter ablation using a mobile cardiac telemetry (MCOT) system revealed that although 70% of patients were free of symptomatic AF recurrences at 6 months post-ablation, the success rate fell to 50% when asymptomatic recurrences were included in the outcome.114 On the basis of these findings, the authors recommended that discontinuation of warfarin treatment after AF ablation should only be envisaged in patients with a low risk of stroke in whom aspirin is considered to provide adequate protection based on the current guidelines.
A study specifically analysing the risk of thrombo-embolic events after radiofrequency ablation of AF was conducted in 755 consecutive patients with paroxysmal (65%) or chronic (35%) AF, of whom 411 (56%) presented one or more risk factors for stroke.113 The mean age of patients in this latter group was 58 ± 10 years and that of patients with no risk factor for stroke was 51 ± 9 (P < 0.01). Male patients predominated in both groups (74 and 79%, respectively). Risk factors included CHF, hypertension, age over 65 years, diabetes mellitus, and prior stroke or TIA. A total of 929 ablation procedures were performed, 71% comprising CPVA28,52 and 29% a tailored approach.64 All the patients were anticoagulated with warfarin for at least 3 months after ablation. Rhythm status was assessed during follow-up by event monitor recordings, serial electrocardiograms, and 24 h Holter monitor recordings. The blanking period was 2 months.
A thrombo-embolic event occurred in seven patients (0.9%) within 2 weeks after ablation, corresponding to the period of highest risk before attainment of a therapeutic international normalized ratio (INR) after the switch back to warfarin after heparin treatment; six of these patients presented a sub-therapeutic INR (below 2.0). A further two patients (0.2%) experienced a late thrombo-embolic event, 6 and 10 months, respectively, post-ablation, both these patients being in SR but one presenting severely impaired left atrial transport function. A total of 522 patients (69%) remained in SR after left atrial ablation, of whom 256 had no risk factor for thrombo-embolic events and 266 had at least one risk factor. Kaplan–Meier analysis indicated freedom from recurrent AF and atrial flutter in the absence of antiarrhythmic drug therapy in 77% of patients with paroxysmal AF and in 66% of those with chronic AF at 12 months after the most recent ablation procedure, the corresponding percentages at 24 months being 73% and 62%.
Recurrent AF was more frequent in patients with chronic AF at baseline (32%) than in those with paroxysmal AF (23%; P = 0.01), whereas recurrent atrial flutter occurred with a similar frequency in the two groups. Warfarin was discontinued in 79% of patients in SR with no risk factor (at a median of 4 months post-ablation) and in 68% of those with at least one risk factor (at a median of 5 months post-ablation; P = 0.003 compared to patients with no risk factor). None of these patients experienced a thrombo-embolic event during 25 ± 8 months (range 10–40 months) of follow-up. Among the 233 patients with recurrent AF or atrial flutter after ablation, 94% continued warfarin anticoagulation. No thrombo-embolic events occurred during follow-up in 6% of patients with recurrent AF who were not anticoagulated. A cerebral haemorrhage occurred in two warfarin-treated patients, 1 and 3 months after ablation, in one patient in the context of an INR of 3.5 and in the other, following head trauma. Among the variables analysed [age over 65 years, gender, paroxysmal vs. chronic AF, hypertension, diabetes mellitus, CHF, and prior stroke or TIA], only prior stroke or TIA (OR 3.63; 95%CI 1.47–8.98; P = 0.005) and age over 65 years (OR 1.82; 95%CI 1.06–3.11; P = 0.03) were independently associated with continuation of warfarin therapy after a successful left atrial ablation procedure.
Bearing in mind the limitations that 31% of patients with a successful outcome, but with baseline risk factors for stroke, were maintained on long-term warfarin therapy and that follow-up was not continued beyond 24 months in most patients, the results of this study provide some evidence justifying discontinuation of warfarin therapy 3–6 months post-ablation in patients with an apparently successful outcome and without baseline risk factors for stroke or with risk factors other than prior stroke/TIA or age over 65 years.
A recently published long-term multi-centre follow-up study, investigating the incidence of cerebrovascular accidents in 2436 consecutive patients who had undergone successful PV isolation with discontinuation of oral anticoagulation 3–6 months post-ablation, indicated the safety of this practice even in patients at high risk.115 A minimum follow-up of 6 months after discontinuation of anticoagulant therapy was required prior to enrolment in this study. AF was paroxysmal in 1575 patients (65%), persistent in 396 (16%), and permanent in 460 (19%). The average duration of AF prior to ablation was 45 ± 34 months and 884 patients (36%) presented with structural heart disease (mean LVEF 55 ± 8%, mean left atrial diameter 43 ± 8 mm). The CHADS2 score for thrombo-embolic events was 0 in 1508 patients (62%), 1 in 646 patients (27%), and 2 in 282 patients (11%). During a mean follow-up of 31 ± 17 months, only one patient (0.04%) experienced a cerebrovascular accident, 1 year after PV isolation. This event occurred in a patient with a CHADS2 score of 0 while in SR. No patient with a CHADS2 score of 1 or 2 experienced a thrombo-embolic event. Subsequent repeat monitoring revealed asymptomatic paroxysmal AF 1 week later. These results, obtained in a very large patient cohort, suggest that discontinuation of oral anticoagulation after successful PV isolation is safe even in patients with a high risk of thrombo-embolic events.
Conclusion
Catheter ablation is increasingly widely used to treat AF and constantly evolving techniques and protocols have led to improved success rates and lower risks of complications in a broad range of patients. The results of comparative trials of catheter ablation vs. antiarrhythmic drug therapy, or a combination of these treatment strategies vs. antiarrhythmic therapy alone, consistently show higher rates of AF suppression with catheter ablation. Most of these comparative trials principally enrolled patients with paroxysmal AF, but superior outcomes with catheter ablation have also been shown in patients with persistent or chronic AF. Catheter ablation can be particularly beneficial in patients presenting HF secondary to AF in whom restoration of SR has been demonstrated to result in improved LV function and quality-of-life even in the context of adequate rate control prior to ablation.
The demonstrated success of catheter ablation in restoring SR encourages consideration of this approach as a first-line therapy in selected patients, particularly symptomatic patients with clearly focally triggered AF in whom a limited ablation procedure is likely to be successful and those with contraindications to antiarrhythmic drugs or a limited choice of these because of concomitant morbidities. Numerous unanswered questions nevertheless remain, including the long-term impact of catheter ablation on the natural history of AF and on major outcomes such as stroke, HF, and survival. Large-scale clinical trials comparing catheter ablation and pharmacological strategies as first-line approaches to the management of AF are badly needed, particularly in higher risk patient subsets, such as those studied in the AFFIRM trial.
Conflict of interest: Dr Jeremy Ruskin declares the following conflicts of interest relevant to AF ablation: Consultant to Astellas, Biosense Webster, Cardiome, CV Therapeutics and sanofi-aventis; Data Safety Monitoring Board member: CryoCath and CardioFocus; Scientific Steering Committee member: Pfizer.
Dr Etienne Aliot declares no conflict of interest relevant to AF ablation.
Funding
Funding for editorial support was provided by sanofi-aventis.
