Since the first description of catheter ablation for focally initiated atrial fibrillation and focal atrial fibrillation in 1995, the popularity of this catheter intervention has grown practically exponentially and the ablation strategy underlying the intervention has evolved considerably. From ‘point’ ablation within the pulmonary venous vasculature,1 ablation has progressed through circular mapping guided pulmonary vein (PV) isolation,2 anatomical PV encirclement,3 extra-PV foci ablation,4 double ostial PV isolation,5 adjuvant left atrial linear ablation,6,7 ablation of complex fractionated potentials,8 isolation of the superior vena cava,9,10 ablation within the coronary sinus,11 ablation targeting the ligament of Marshall12 and, recently, even isolation of the left atrial appendage.13
The multitude of different choices of ablation strategies emphasises the underlying acknowledgement that the optimal strategy is unknown, that the optimal choice may vary according to individual or group variations and even that the methods underlying analysis of outcomes after catheter ablation may have significant limitations.
Classification of Atrial Fibrillation Subtypes
It may well be that we have oversimplified atrial fibrillation by blurring boundaries between varying and different pathophysiologies. The younger individual (≥50 years of age) without any structural heart disease and short-lasting, frequent paroxysms of symptomatic atrial fibrillation is clearly very different from the 70-year-old with mitral valve disease, dilated left atrium and chronic long-standing persistent atrial fibrillation. The latter patient has clear structural and anatomical differences from the former, many of which are echocardiographically evident and others that are predominantly ultrastructural and microscopic. Nevertheless, the current dogma is to search for the universally applicable single solution to the electrocardiogram (ECG) syndrome of atrial fibrillation. The older patient with structural heart disease such as mitral valve incompetence often has extensive collagenous replacement of the atrial myocardium: adding ablation-induced extensive scars can only be of limited help to restore electrophysiologic sinus rhythm but will certainly only worsen the mechanical function of this patient’s left atrium.
The currently accepted nomenclature and classification of atrial fibrillation is based on clinical simplicity. Paroxysmal is distinguished from persistent atrial fibrillation by the duration of untreated episodes lasting beyond seven days in case of persistent fibrillation.14 However, if atrial fibrillation terminates after antiarrhythmic treatment (only presumably due to it) before seven days or is electrically cardioverted before seven days, then in the presence of uncertainty as to how long the untreated episode would have lasted, current definitions classify the fibrillation as persistent. The practical implications for catheter ablation for atrial fibrillation are that many patients currently classified with persistent atrial fibrillation may actually have paroxysmal atrial fibrillation and perhaps lesser degrees of substrate remodelling. The outcomes of catheter ablation of these patients categorised with persistent atrial fibrillation are inevitably more similar to those of paroxysmal atrial fibrillation. Along the same lines, permanent atrial fibrillation is defined by the acceptance of continuing atrial fibrillation or the inability to cardiovert the fibrillation.14 Thresholds for accepting continuing atrial fibrillation vary within the medical community and the inability to cardiovert fibrillation may also vary depending upon the availability of biphasic defibrillation, the placement of defibrillatory paddles, the amount of current reaching the heart (obesity, pressure on the paddles, etc.) and the use or otherwise of concurrent anti-arrhythmic drug treatment with effects on the defibrillation threshold. In addition, it is not unlikely that immediate re-initiation of atrial fibrillation after cardioversion is misinterpreted as ineffective cardioversion. All the foregoing factors have considerably weakened the correlation of the type of atrial fibrillation with the underlying EP substrate and the optimum strategy of catheter ablation.
Non-uniform and Imperfect Follow-up
The accurate long-term assessment of rhythm outcome after catheter ablation is critical for assessing ablation efficacy and choosing strategies of ablation. Lack of symptom correlation, asymptomatic arrhythmias, the inability to measure total arrhythmia burden and the frequent lack of distinction between flutter or organised tachycardias and atrial fibrillation are all important issues that render the assessment of the outcome of individual strategies of ablation difficult.
Ablation Strategies and Resulting Tissue Changes
A prerequisite for a precise analysis of the optimal strategy for catheter ablation of atrial fibrillation is the knowledge of the extent of the electrophysiological changes and of tissue necrosis produced as a result of myocardial ablation. Radiofrequency (RF) current delivery produces coagulative tissue necrosis as a result of tissue heating. A thin rim of resistive tissue heating at the interface with the electrode provides the source for conductive heating of surrounding tissue. The resulting temperature isochrones exceeding about 50–55°C determine the extent of irreversible tissue necrosis15 but cannot currently be imaged or delineated. Borderzone tissue, thought to exhibit reversible electrophysiological dysfunction, can completely recover after a period of days to weeks, but cannot currently be recognised during ablation. The extent of irreversible coagulative tissue necrosis is thought to be smaller than the acute lesion as demarcated by extracellular potential changes. The rhythm outcome at six, 12 and 24 months after catheter ablation is most likely related to the extent and distribution of irreversible ablation-produced scar rather than the acute lesion16 and our current inability to discern and delineate scar tissue resulting from catheter ablation makes analysis of any ablation strategy based on lesion delivery imprecise. The recent development of delayed enhancement magnetic resonance imaging (MRI) of scar in the atria is a step in the right direction with the provisio that pre-existing disease-related scar needs to be distinguished from ablation-produced scar for an accurate analysis of ablation strategy and outcomes.17
Identifying the Poor Outcome Candidate
Despite the above constraints, the last 10 years of clinical experience have provided a few indications about the ideal candidate and the optimum technique. Candidate patients for catheter ablation are currently selected based on the underlying presence of refractory, symptomatic atrial fibrillation. Identifying the poor outcome candidate is probably as important as identifying the best outcome candidate. Such patients include those with a very poor likelihood of maintaining stable sinus rhythm and those who are unlikely to benefit from electrical sinus rhythm maintenance probably because of poor mechanical function. It is unclear whether these represent the same group of patients.
During the initial experience with catheter ablation of atrial fibrillation, parameters associated with poor outcomes in the surgical literature (maze and its variants) have been empirically used to identify (and exclude) poorer-outcome patients. Large left atrial size, older age, mitral valve disease and duration of atrial fibrillation have all been identified as predictors of poor rhythm outcome after maze surgery or its variants.18 With the accrual of an increasing volume of experience with catheter ablation, the relevance of these parameters has been generally confirmed, although usually in single-centre, retrospective studies. Left atrial size appears to correlate with rhythm outcomes and severe left atrial dilatation is considered a poor prognostic marker for stable restoration of sinus rhythm.19 Left atrial volume measurements, whether by 3D echo, 3D mapping, angiography, computed tomography (CT) or MRI, may be more reliable in identifying severe dilatation and a poor prognosis for restoring sinus rhythm compared with 2D parameters.20
Rheumatic mitral valve disease results in atrial pressure overload and severe dilatation and is associated anecdotally with poor outcomes after catheter ablation,21 but structured and rigorous evidence is still lacking. Similarly, advanced age and long-duration atrial fibrillation are associated with atrial dilatation and poor rhythm outcomes but also require further validation.
Other parameters that have been proposed include extensive areas of low-voltage electrograms or no endocardial signal in the left atrium,22 which correspond to pre-existing scar. Recent reports suggest that extensive baseline delayed enhancement after gadolinium injection during MRI may be predictive of a poor rhythm outcome after catheter ablation.17 However, the tissue pathology resulting in delayed enhancement is unclear and may or may not correlate with an electrophysiologically relevant substrate.
Tailoring the Ablation Strategy and Technique
A ‘sartorial’ analogy applied to catheter ablation of atrial fibrillation makes intuitive sense because of obvious individual differences in patient characteristics. An individualised strategy may avoid unnecessary ablation resulting in shorter procedure duration, lesser effects on atrial mechanical and endocrine/paracrine function, perhaps less atrial stasis and fewer atrial tachycardias during follow-up. The recommendations that follow are based on personal experience and a review of the available evidence. It is clearly desirable to derive inferences from randomised, prospective, multicentre studies but, for the most part, these are either under way or not available.
Accordingly, young patients (<35 years of age) without structural heart disease who manifest with so-called ‘focal’ atrial fibrillation, essentially a pulmonary-vein-originating tachycardia,1 should be treated with isolation of the individual arrhythmogenic pulmonary vein. In the event of interconnections with the ipsilateral pulmonary vein, additional isolation of that vein may be necessary to ensure stable maintenance of sinus rhythm.5 Young patients should undergo this strategy of minimum ablation in the absence of evidence of multi-PV involvement.
Patients with paroxysmal atrial fibrillation may also have an underlying paroxysmal supraventricular tachycardia (PSVT) that secondarily degenerates or triggers atrial fibrillation. In such patients, it appears reasonable to ablate the PSVT and evaluate the clinical response. Typically, younger patients respond well with clinical elimination of atrial fibrillation, whereas older patients may require additional ablation or isolation of the pulmonary veins.
Patients with paroxysmal atrial fibrillation and short-lasting episodes (usually <12 hours) appear to benefit from pulmonary vein isolation alone. Circumferential mapping guided segmental PV isolation or encircling PV ablation with electrophysiologically proven isolation are probably equally effective provided a sufficiently proximal isolation is performed. Ostial segmental PV isolation relies upon point-by-point ablation, a single PV ostium at a time. More extensive and proximal PV isolation may be achieved by the use of a variable loop circular mapping catheter (allowing more proximal positioning towards the left atrium) and/or a second level of more proximal lesions in continuity with the mitral isolating lesion. There is evidence to indicate that more proximal ablation provides better short- and medium-term rhythm outcomes.23 An imaging technique (pre-acquired CT or MRI images) when combined with circular mapping at the PV ostium can allow more-proximal ablation isolation as described for the double ostial PV isolation technique.24 This technique requires an extensive circumferentially complete line and theoretically is prone to a higher likelihood of gaps, although published single-centre results are good.
It is unclear whether adjuvant ablation strategies, such as linear left atrial ablation or complex fractionated atrial electrogram (CFAE) ablation, significantly improve the rhythm outcome in patients with paroxysmal or short-lasting persistent atrial fibrillation.
Patients with documented typical atrial flutter in addition to paroxysmal atrial fibrillation should also undergo cavotricuspid isthmus ablation. In our laboratory, if sustained typical atrial flutter is induced or observed during the procedure, cavotricuspid isthmus ablation is performed despite the absence of prior clinical documentation of flutter. We found no difference in typical flutter rates during long-term follow-up in two groups of patients undergoing atrial fibrillation ablation: one group with documented or observed typical atrial flutter who underwent additional cavotricuspid isthmus ablation and the other group without flutter who did not.25
A minority of patients with atrial fibrillation also have atypical left atrial flutter and should undergo ablation for both the fibrillation and the flutter. If the flutter is PV-related an isolation of the pulmonary veins may suffice but if the flutter is dependent on a perimitral circuit or a circuit around the ipsilateral pulmonary veins, then additional linear lesions with the achievement of conduction block are required. In the event that the flutter cannot be induced in the electrophysiology lab, determining the exact circuit based on the ECG is not reliable enough to propose a specific ablation strategy. Although a minimaze-like linear lesion scheme (a line joining the left and right PVs and another from the left PVs down to the mitral annulus) has been proposed, the present-day linear lesion-making technology is not effective enough to exclude slow-conducting gaps and may in fact generate atrial reentrant tachycardias.26
Patients with atrial fibrillation and a persistent left superior vena cava draining into the coronary sinus should undergo not only PV isolation but also isolation of the left superior vena cava (SVC).27
Those with documented evidence of (right-sided) SVC ectopy or tachycardia originating from the SVC should undergo SVC isolation although the prospect of phrenic nerve injury may limit this strategy.28 Ablation within the coronary sinus is often performed as part of the strategy of ablating complex fractionated atrial electrograms.8,11 The ligament of Marshall is not directly targeted for ablation because of the risks involved in its cannulation12 and arrhythmias suspected of originating from this structure may be targeted endocardially from the posterior left atrium in the vicinity of the left PV ostia or from within the coronary sinus depending upon the specific exit point.
For the large subgroup of patients with long-standing persistent AF, the rhythm outcome success rates are unsatisfactorily low after PV isolation alone and additional ablation in the atria is necessary. The creation of left atrial linear lesions6,7 and the ablation of complex fractionated electrograms11 have been shown to improve sinus rhythm maintenance rates when combined with PV isolation. However, additional left atrial ablation, whether in the form of linear lesions or targeting complex fractionated electrograms, results in an increased incidence of post-ablation atrial tachycardias, typically multiple forms and circuits.29 The commonest linear lesions deployed in the left atrium include one across the roof and posterior wall of the left atrium extending from the left superior to the right superior PV ostia and another from the left inferior pulmonary vein ostium to the postero-lateral mitral annulus. Other lesions have been created from the right superior PV ostium down to the anterior mitral annulus30 as well as from the left atrial appendage–left superior PV ostium junction down to the anterior mitral annulus. Point-by-point lesions are strung together to try to create a confluent linear lesion and often no attempt is made to verify or achieve conduction block across these lesions.
Complex fractionated electrograms are widely distributed, their definition is open to broad variation, the underlying activation mechanisms support both a bystander as well as an active maintaining role, and the temporal evolution of the electrogram morphologies is unknown. The development of reliable multielectrode simultaneous activation mapping during atrial fibrillation could allow targeting complex electrograms maintaining atrial fibrillation and avoid ablating bystander electrograms, which could contribute to generating pro-arrhythmic reentrant tachycardias. The extensive atrial ablation required by both these adjuvant ablation strategies may compromise mechanical function and perhaps result in pro-coagulative consequences. Performing additional atrial ablation in stages once the transiently arrhythmic ‘blanking period’ (probably related to post-ablation inflammation) has run out may reduce the number of patients requiring additional atrial ablation and/or re-ablation for atrial tachycardias. Other strategies include adding prophylactic linear lesions or attempting to ablate and eliminate all inducible atrial tachycardias during the index ablation, but these strategies do not account for the pro-arrhythmic consequences of the additional ablations performed to eliminate inducible tachyardias and their evolution during the weeks that follow.
The recently proposed isolation of the left atrial appendage as an adjuvant ablation strategy will require further evaluation.13 The left atrial appendage functions as a key contractile element of the left atrium, is thought to be important to maintaining left atrial compliance and is a major source of atrial natriuretic peptide secretion. Its isolation may therefore have significant haemodynamic and pro-coagulative consequences. The atrial fibrillation suppressing effect of left atrium appendage isolation will require careful analysis in a study designed to distinguish the effect of individual debulking lesions and to define the consequences of isolating this structure.
In animal studies, atrial fibrillation requires strong vagal stimulation or vagomimetic stimulation to be sustained. Stimulation and inhibition of epicardial ganglionic plexi in experimental studies has a significant influence on the arrhythmogenicity of the PVs and the inducibility of atrial fibrillation.31 The current inability to selectively target GPs by ablation in humans has tempered interest in its clinical use. Nevertheless, there is likely to be a significant component of ganglionic plexus ablation occurring with current strategies of endocardial ablation and its exact contribution is not known.
An individualised strategy linked to an imaging modality capable of delineating the reversible and/or irreversible lesions in realtime may allow the optimisation of lesions during the index procedure so as to ensure the efficacy of eliminating atrial fibrillation as well as blocking potential isthmuses capable of sustaining reentrant atrial tachycardias. MRI may well be a strong candidate for such a modality although significant challenges related to imaging thin-walled atria as well as working in a high signal noise environment with an extremely strong magnetic field remain.
All catheter ablation techniques should try to maximise the probability of a stable, permanent lesion set. For PV isolation, this means careful verification of conduction block at intervals out to 60 minutes (or more). Realtime evaluation of catheter tissue contact has recently become clinically available and has the potential to reduce conduction recurrence.32
Patients with atrial fibrillation requiring rhythm control and undergoing concomitant cardiac surgery have the option of undergoing intra-operative ablation, typically in line with the catheter ablation strategies outlined above. A ‘cut and sow’ technique ensures lesion completeness but when using intra-operative RF energy or cryoenergy, testing for lesion completeness is essential. The surgical approach is hampered by not being able to verify electrophysiological completeness and the stability of the lesions with the same sophistication and perserverance as in the electrophysiology lab because of the difficult operating room environment and the imperatives of coming off bypass.
New Ablation Devices and the Choice of Ablation Energy
Radiofrequency energy is the current modality of choice. It is simple to use, does not require complicated equipment and is well understood from a bio-physics point of view. An extensive body of experience has also accumulated over the last two or three decades with the use of this energy modality. Probably the most important weaknesses of RF energy are its dependence on tissue contact and the tendency for borderzone recovery of electrophysiological and cellular function.
Among current alternative candidates, cryoenergy and laser are probably the most interesting. Cryoenergy has been touted as allowing reversible cryo-mapping at locations where tissue necrosis could result in complications or negative consequences. Additionally, cryoenergy delivery produces an ice ball or a layer of ice that adheres the catheter/energy delivering surface to the target tissue and avoids the problems of intermittent or fluctuating contact. However, the effective cooling of target tissue may be difficult to achieve uniformly in the circulating body temperature blood pool. Cryolesions for SVTs appear to be associated with a high rate of clinical recurrence and this may extend to PV isolation as well. The time cycle of effective cryoablation with present-day technology is long and therefore a disadvantage. Furthermore, cryoablation lesions delivered to the atrial endocardium have in fact produced oesophageal lesions.33 The introduction of cryoballoon isolation of the PVs will have a significant impact on the ablation strategy only if the procedure can be simplified and shortened without compromising efficacy and patient security. The recent demonstration of the feasibility of PV isolation with a visually guided circumferential spot ablating laser balloon has re-kindled interest in an energy that was thought to be difficult to control.34
Circular multielectrode RF ablation devices make intuitive sense for the purpose of isolating the pulmonary veins, and the recent introduction of such a catheter combining bipolar and unipolar RF current delivery is an attempt in this direction.35 However, currently there is no ‘one-shot’ ablation device and the old-fashioned technique of point ablation still offers many advantages.
Synthesis
Being able to select out the worst outcome candidates for catheter ablation should allow electrophysiologists to propose rate control strategies to this segment of the atrial fibrillation population with confidence. The patients with short-lasting, frequent bouts of atrial fibrillation with or without typical flutter are best treated with some form of electrophysiologically verified stable PV isolation and if necessary the creation of additional cavotricuspid isthmus block. For patients with persistent atrial fibrillation, electrophysiological PV isolation is certainly the cornerstone of the ablation strategy. It is unclear whether left atrial linear lesions and/or complex fractionated atrial electrogram ablation is necessary to achieve stable sinus rhythm. Avoiding unnecessary ablation in order to preserve contractile atrial myocardium and avoid atrial pro-arrhythmia should dictate ablation strategies. Recent progress in MRI may allow individualised strategies based on realtime recognition of underlying scars and diseased tissue as well as reversible and irreversible tissue necrosis produced by catheter ablation.