PVI represents the cornerstone of AF ablation worldwide with a class I (level of evidence A) recommendation from current guidelines [2]. As per a survey conducted in 2015, PVI was adopted as the primary procedural strategy for persistent AF patients in over 60% of the 30 European centers participating [13]. It is worth noting that the current focus on PVI signifies a shift towards a simpler approach following the extensive exploration of various additional ablation strategies over the past two decades. Described initially by Haïssaguerre et al. [14] early studies of PVI only in persistent AF patients yielded largely underwhelming results with higher recurrence rates compared to those with paroxysmal AF [15, 16, 17]. The inferior outcomes observed in the persistent AF cohort may be attributed to a more advanced atrial substrate, distinct from trigger-driven AF commonly observed in the paroxysmal pattern of disease, that responds favorably to PVI. In order to target this arrhythmia substrate, additional ablation techniques were proposed as additional strategies beyond PVI.

An adjunct strategy of linear mitral isthmus and left atrial roof ablation to achieve bidirectional conduction block, in addition to PVI, was first reported in 2004 [9, 18]. The rationale behind this approach, was derived from knowledge gained from the surgical maze procedure and concept of atrial compartmentalization, by which lines of electrical block would be theorized to reduce the overall electrical conducting area of the atria and thus limit its capacity to sustain AF [19]. This strategy gained significant traction in the face of several studies demonstrating higher success rates with linear ablation compared to PVI alone, an effect that was most notable in persistent AF patients [20, 21, 22]. These studies emphasized the difficulty in achieving acute mitral block, however, with often low final rates of bidirectional block in this region. The rate of achieving roofline block was notably greater in comparison. Although adjunct linear ablation may reduce the incidence of recurrent atrial tachycardia (AT), in the presence of incomplete block the risk of AT appears to increase [23]. In general, while some studies indicate a potential incremental benefit of linear ablation in patients with persistent AF, evidence is conflicting, and disappointing long-term outcomes have been observed in other studies [5, 21]. The randomized STAR AF II trial (Approaches to Catheter Ablation for Persistent Atrial Fibrillation) conducted in 2015, involving 589 patients across multiple centers, stands as the most rigorous assessment to date of an adjunctive ablation strategy in patients with persistent AF [5]. At 18 months no benefit was observed beyond PVI in those undergoing mitral and roof lines with only 46% free from recurrent AF. It is worth noting that block across both mitral and roof lines was achieved in only 75% of patients in this study, however and overall, the efficacy and long-term success of linear ablation as an adjunctive strategy for persistent AF patients remain uncertain.

In 2007, Kumagai et al. [24] introduced the concept of box isolation as a linear, end-point driven ablation technique targeting the posterior left atrial wall. The objective of box isolation was to isolate focal and re-entrant drivers responsible for persistent AF, which were frequently identified in this region [25]. A mechanistic study utilizing phase mapping in patients with persistent AF provided support for the rationale behind the box isolation technique and revealed a decrease in left atrial critical mass and AF drivers after isolation of the posterior wall was achieved [26]. Similar to the mitral isthmus, achieving isolation of the posterior box can be challenging, however, and while this approach was not evaluated in STAR AF II, several randomized studies have shown no additional benefit to this strategy beyond PVI only [27, 28, 29]. Recently the multicentre, randomized CAPLA trial (Effect of Catheter Ablation Using Pulmonary Vein Isolation With vs Without Posterior Left Atrial Wall Isolation on Atrial Arrhythmia Recurrence in Patients With Persistent Atrial Fibrillation) addressed this question in a larger population of 338 persistent AF patients [6]. Once again, no difference in arrhythmia free survival at one year was seen in patients undergoing PVI alone vs PVI plus posterior box isolation with modest success rates seen in both groups (53.6 vs 52.4%, p = 0.98). Lack of benefit may relate to the iatrogenic creation of pro-arrhythmic substrate due to incompletely ablated tissue and/or failure to achieve complete isolation. Concerns regarding ablation at the posterior wall relate to its close proximity to the oesophagus, with oesophageal injury and the rare but often fatal atrio-oesophageal reported in the setting of extensive posterior wall radiofrequency (RF) ablation [5, 30, 31]. Indeed, intra-oesophageal temperature rise resulting in operator reluctance to persist with ablation is a common reason for failure to achieve complete isolation. This in turn is associated with posterior wall reconnections and treatment failure [32]. As such it carries only a IIB level of evidence C recommendation [33].

CFAE are characterized by the presence of complex fractionated potentials, continuous electrical activity, or short mean cycle lengths [34]. These electrograms are often observed in areas of slow conduction, conduction block, as well as at anchor points for re-entry circuits, wavefront collision, and ganglionated plexi [10, 35, 36]. In 2004, success rates from ablation of CFAE sites (mapped during AF), without PVI, were reported by Nademanee et al. [10] in both paroxysmal and persistent AF patients. While promising results were seen initially, a subsequent study failed to replicate these findings, with only 33% of persistent patients maintaining sinus rhythm at a follow-up of over 1 year [37]. Moreover, when CFAE ablation was performed in addition to PVI, several randomized trials reported mixed results in the persistent AF population [38, 39, 40]. The inconsistency in defining CFAEs and endpoints among different studies poses a significant challenge when comparing their results. Furthermore, mechanistic studies have suggested that a substantial portion of CFAEs might not actively contribute to the perpetuation of atrial fibrillation AF [41], leading to potential concerns regarding excessive ablation of noncritical regions. Consequently, this approach can prolong procedure times without a clearly defined objective. Adjunct CFAE ablation was evaluated in STAR AF II, again without a significant demonstrable benefit beyond PVI alone and only 49% undergoing this strategy free of recurrent arrhythmia on 18-month follow up [5].

In 2005, Haissaguerre et al. [42] described a ‘stepwise’ approach for persistent AF patients, combining PVI, CFAE ablation, and linear mitral and roof ablation. The procedural endpoint of this approach was AF termination. Several studies reported high rates of sinus rhythm maintenance after multiple procedures and over longer-term follow-up of up to 5 years [43, 44]. Although the STAR AF II trial did not evaluate this specific approach, proponents of the method argued that extensive substrate ablation over multiple procedures could achieve high long-term success rates in restoring sinus rhythm.

The development of various advanced mapping techniques for characterization of mechanisms of persistent AF coincided with the assessment of the aforementioned ablation strategies. Phase mapping is a mathematical technique utilized to characterize spatial and temporal patterns of electrical activity in cardiac tissue. Its application to ablation of AF involves identifying and mapping periodic rotations, known as “rotors”, which play a crucial role in sustaining AF [45]. An invasive endocardial mapping approach, “focal impulse and rotor modulation” (FIRM), employs a specialized catheter with 64 electrodes arranged in a basket configuration (FIRMap by Topera, Palo Alto, CA, USA; Constellation by Boston Scientific, Marlborough, MA, USA). This technique enables the creation of endocardial maps illustrating AF propagation, facilitating real-time identification and ablation of stable foci exhibiting certain spatial and temporal characteristics [46]. The FIRM mapping approach was evaluated in the CONFIRM (Conventional Ablation for Atrial Fibrillation With or Without Focal Impulse and Rotor Modulation) trial in predominantly persistent AF patients undergoing conventional ablation vs conventional plus FIRM guided ablation [11]. The elimination of stable rotors or focal impulses through ablation led to the termination or deceleration of AF in 86% of participants. Notably, the single-procedure success rates were significantly higher in individuals who underwent adjunct FIRM-guided ablation compared to those who did not (82.4% vs 44.9%, p 0.001) [47, 48]. Nevertheless, replicating these findings proved challenging [47, 48], and the randomized REAFFIRM trial (Prospective randomized comparisonof rotor ablation vs conventional ablation for treatment of persistent atrial fibrillation), which involved 375 patients with persistent AF, failed to demonstrate any outcome benefit after one year [49].

Electrocardiographic imaging (ECGi) mapping is a non-invasive form of phase mapping employing a multi-electrode body surface vest in conjunction with thoracic imaging. By analyzing unipolar electrograms, this method generates activation maps that aid in the detection of re-entrant circuits and focal activities. In an early clinical ECGi mapping study, conducted in 2010, the presence of multiple wavelets was the most prevalent activation pattern in patients with AF with targeted, ECGi-guided ablation resulting in termination of AF [50]. Following that, the ECVUE ECGi mapping system (developed by CardioInsight, Cleveland, OH, USA) was employed to study 103 patients with persistent AF. The findings identified unstable re-entry circuits, rather than stable focal sources observed through FIRM mapping, as the primary mechanism for AF perpetuation in the patients studied [51, 52]. Subsequently the AFACART study (Multicentre evaluation of non-invasive biatrial mapping for persistent atrial fibrillation ablation), conducted in 2017 across multiple centers, investigated the effectiveness of ECGi-guided ablation in patients with persistent AF. The primary endpoint of the study was the termination of AF, with 77% of patients free from recurrence at the one-year mark [53]. It is important to note, however, that there are no randomized trials evaluating this technique in AF ablation.

The complexity involved in the transformation of electrograms obtained using phase mapping with limited ability for real-time intraprocedural analysis of raw signals are clear drawbacks to the phase-mapping approach. Additionally, limitations associated with electrode density and mapping resolution, as well as conflicting findings regarding the stability of drivers when comparing FIRM and ECGi mapping, have raised concerns regarding the reproducibility and validity of the technique. In recent years, additional mapping systems have emerged to enable the identification of AF drivers without relying on phase mapping transformation. Two notable systems are the Cartofinder contact mapping system (developed by Biosense Webster) and the AcQMap non-contact ultrasound-based mapping system (developed by Acutus Medical). Studies have shown that targeted elimination of focal or rotational activity identified using these platforms has been associated with high rates of AF termination and midterm success rates of up to 70% in patients with persistent AF [54, 55]. However, similar to ECGi mapping, there are currently no randomized trials available to assess the efficacy of these techniques. Further research is needed to evaluate and validate their utility in AF mapping and ablation procedures.

Collectively, when considering the findings of the studies mentioned above, and in alignment with the STAR AF II trial, there is no substantial evidence to support adjunct ablation strategies beyond PVI or ablation guided by mechanistic mapping techniques.

One possible explanation for the lower success rates observed in previous studies could be the failure to create durable transmural lesions during the initial procedure. This is reflected in the significant reduction in rates of pulmonary vein (PV) reconnection seen at redo procedures, after the advent of contact force sensing catheters [58, 59]. In recent years, increased emphasis has been placed on better understanding of the biophysics of ablation to maximise procedural efficacy through safe creation of transmural ablation scar. The Ablation Index (Carto3, Biosense Webster) is a metric which allows for real time assessment of lesion formation using the key parameters of contact force, time, and power combined in a weighted formula. Preclinical studies have demonstrated its ability to predict lesion depth [60] and in humans, the Ablation Index has shown promise in identifying sites of pulmonary vein reconnection during repeat procedure [61].

The “CLOSE protocol” workflow, which focuses on creating continuous and optimized RF lesions with specific target Ablation Index values, has shown promising outcomes in patients with paroxysmal AF with high rates of freedom from recurrence after a single procedure [62, 63]. In the PRAISE-AF study (Use of Ablation Index-Guided Ablation Results in High Rates of Durable Pulmonary Vein Isolation and Freedom From Arrhythmia in Persistent Atrial Fibrillation Patients), the CLOSE protocol was evaluated in a small group of patients with persistent AF. Following a protocol-mandated repeat procedure, with pulmonary vein re-isolation in 22% of patients demonstrating reconnections, 95% of patients were free from recurrent AF at one year [64].

These findings may suggest that optimized Ablation Index-guided PVI alone can achieve high clinical success rates in patients with persistent AF.

Very high-power, short-duration ablation offers the potential advantage of enhanced procedural safety and efficacy, with pre-clinical studies suggesting predominantly resistive tissue heating, translating into a high rate of lesion contiguity and transmurality [65, 66, 67, 68]. In humans, the randomized POWER PLUS study (Very High-Power Ablation for Contiguous Pulmonary Vein Isolation) of PVI using 90 W/4 second vs conventional ablation with the QDOT Micro catheter (Biosense webster, Diamond Bar, CA, USA) in mixed population of paroxysmal and persistent AF patients confirmed a five-minute reduction in procedure time with a similar safety and midterm efficacy profile [69]. A trend towards lower first pass isolation was noted, with gaps frequently identified at the thick-walled anterior carina in the 90 W group. This is in line with a recent animal study reporting significantly smaller lesion sizes with this modality, reflecting a lower overall energy deposit to the tissue [70]. Further studies will shed light on the efficacy of very high-power ablation beyond the PVs, however the possibility of smaller lesions is a consideration when ablating at thicker regions such as the mitral isthmus.

An alternative form of energy delivery, cryoablation, has emerged over the last decade as a major competing technology with RF for PVI only. This ‘single-shot’ technique results in pulmonary vein isolation via delivery of a refrigerant to a custom-made balloon placed at the ostia of the pulmonary veins [71]. In paroxysmal AF patients it was found to be non-inferior in safety and efficacy to RF ablation in a large randomized controlled trial [72]. Several non-randomised studies have subsequently supported the feasibility of this approach in persistent AF patients, with modest results, comparable to those obtained with RF ablation [73, 74]. Results from the randomized, multicentre Fire and Ice II trial, comparing cryoballoon to RF ablation in persistent AF patients, are awaited [75].

While increased durability of PVI has translated into improved success rates in paroxysmal AF patients, in persistent AF patients with advanced substrate remodelling, it may not be a sufficient strategy to prevent recurrences.

Atrial fibrosis is a prominent feature of atrial structural remodeling, characterized by an excessive accumulation of extracellular matrix, as observed in histological studies. It increases with advancing age, can be present in cases of lone AF as well as AF associated with underlying cardiac conditions, and furthermore has been correlated with the severity and persistence of AF [76, 77, 78, 79]. Two methods exist for identifying and quantifying this important disease marker clinically: atrial late gadolinium enhancement cardiac magnetic resonance (LGE-CMR) imaging and invasive voltage mapping. The extent and severity of high signal intensity regions on LGE-CMR, or low voltage areas on invasive mapping (as surrogate markers for atrial fibrosis) has been shown to be a powerful independent predictor of recurrence post ablation [80, 81, 82]. Increased availability of LGE-CMR imaging and ongoing advances in catheter technology have allowed for the routine assessment of underlying fibrotic substrate in persistent AF patients before or during ablation, with the targeting of these areas for ablation a focus of recent trials.

Several studies have investigated the efficacy of voltage-guided ablation strategies, which involve ablating or isolating regions of low voltage in the atria. Rolf et al. [83] reported comparable success rates between patients who underwent voltage-guided ablation of low voltage areas and those without low voltage areas undergoing standard ablation. A subsequent small retrospective study demonstrated improved outcomes with additional voltage-guided ablation, specifically the posterior left atrial wall [12]. In the randomized, mulitcentre STABLE-SR study (Low-Voltage Myocardium-Guided Ablation Trial of Persistent Atrial Fibrillation) of persistent AF ablation examining the stepwise approach vs PVI plus homogenization of low voltage areas, comparable success rates were seen (74% vs 71%, p = 0.325) with lower procedure and fluoroscopy times in the group undergoing voltage guided ablation [84]. Several factors should be considered with voltage-guided ablation. Mapping resolution, relating directly to catheter electrode size and spacing, may influence the appearance of voltage maps, as can activation direction and local conduction velocity, which have been associated with changes in voltage amplitude in atrial and ventricular myocardium [85, 86]. Atrial rhythm may also affect recorded voltage which has been demonstrated to be lower during AF than sinus rhythm [87]. Furthermore, not all low voltage areas may be mechanistically relevant for atrial arrhythmogenesis. In the absence of robust histological validation for pre-ablation voltage mapping, optimum thresholds for defining relevant low voltage zones remain somewhat unclear. Most centres routinely apply a value of 0.5 mV to define low voltage, based on animal studies of post myocardial infarction scar in the ventricle [88, 89]. In a recent study Rillo et al. [90] identified an optimum range of 0.3–0.6 mV for identifying low voltage zones and surmised that only ‘non-compact’ zones are of interest for ablation. The comprehensive MASH II trial (Characterization of Atrial Substrate to Predict the Success of Pulmonary Vein Isolation: The Prospective, Multicenter MASH-AF II (Multipolar Atrial Substrate High Density Mapping in Atrial Fibrillation) Study) in 262 AF patients sought to further define low voltage zones and their association with outcome post ablation. It found that the impact of low voltage area on outcome differs between paroxysmal and persistent AF patients, suggesting that not all low voltage zones have the same prognostic implications [91].

Regarding non-invasive CMR assessment of atrial substrate, the last two years has witnessed the publication of two important randomised trials examining adjunct fibrosis guided ablation using CMR or voltage mapping in persistent AF. The multicentre DECAAF II trial (Effect of MRI-Guided Fibrosis Ablation vs Conventional Catheter Ablation on Atrial Arrhythmia Recurrence in Patients With Persistent Atrial Fibrillation) randomised 843 persistent AF patients undergoing 1st time ablation to PVI vs PVI plus LGE-CMR guided atrial fibrosis ablation [8]. Ablation strategy in the latter arm included encircling or covering all areas of fibrosis identified on atrial LGE-CMR. After a follow up period of 12–18 months there was no difference in arrhythmia recurrence between groups (46 vs 43%, p = 0.63). Furthermore, more safety events occurred in patients undergoing additional fibrosis guided ablation (p = 0.001), particularly relating to a higher rate of ischaemic strokes in this group. Subsequently the multicentre ERASE trial (Low-Voltage Myocardium-Guided Ablation Trial of Persistent Atrial Fibrillation) randomised 324 persistent AF patients to PVI only vs PVI plus low voltage ablation based on the findings of invasive mapping [7] (Fig. 1). In contrast to the DECAAF study, the primary endpoint of arrhythmia recurrence of $>$30 seconds occurred in significantly more patients undergoing PVI only vs PVI plus voltage guided ablation (50 vs 35%, p = 0.006). It is worth noting that in this study, only those with low voltage areas underwent additional substrate ablation and that all those with ‘healthy’ atria underwent PVI only, regardless of randomisation group. As such only 54 out of the 324 patients randomised underwent voltage guided ablation. Furthermore, similar to DECAAF II a trend towards more complications were seen in the group randomised to additional substrate ablation. As for voltage mapping, lack of robust histological validation remains an inherent limitation. Several methods exist for thresholding atrial LGE images and the optimal threshold for identifying clinically relevant fibrotic regions to be targeted for ablation is still subject to debate. Inherent difficulties with imaging the very thin-walled atria and the requirement for bespoke software for image processing has limited the uptake of the technique to a relatively small number of expert centres, and differences in imaging protocols and thresholding techniques has resulted in difficulty reproducing results between centres.

Fig. 1.

Low voltage myocardium guided substrate ablation as employed in the ERASE trial. (A) PA view of left atrium with low voltage areas identified in the posterior wall. (B) Ablation strategy in those randomized to PVI only. (C) Box isolation of left atrial posterior wall in addition to PVI as an example of strategy in those randomized to PVI plus substrate modification. Reproduced from Huo et al. ‘Low-Voltage Myocardium-Guided Ablation Trial of Persistent Atrial Fibrillation’. NEJM Evidence. 2022 Oct 25;1(11):EVIDoa2200141. Reprinted with permission from Massachusetts Medical Society. PVI, pulmonary vein isolation; PA, postero-anterior.

Interestingly, using both invasive and non-invasive measures of fibrosis characterisation, high success rates with PVI only, of up to 84%, were seen in sub-analyses of the STABLE SR and DECAAF II trials in patients without evidence of atrial fibrosis [8, 84]. This again underscores the point that in persistent AF patients with early-stage disease, PVI alone is likely to be a sufficient strategy.

The traditional 7-day AF classification does not capture the true extent of underlying arrhythmia substrate in the persistent AF cohort. Heterogeneity in disease subtype may be better characterised by more specific markers of atrial remodelling including atrial size and presence of fibrosis, duration of continuous AF episodes, rhythm at time of procedure and number of and response to prior cardioversions. Significant variation exists, however, regarding the consistent reporting of such parameters in studies and therefore subtype of AF is likely to differ between trials rendering direct comparison of results difficult. The refinement of AF definition coupled with the reporting of a wide range of such disease related parameters will be important for future trial design to enhance interpretation of results and to identify or correct for selection bias.

When evaluating trial outcomes, it is crucial to take into account the techniques utilized for post-procedural rhythm monitoring. Greater sensitivity in detection of arrhythmia recurrence is seen with increasing intensity of rhythm monitoring [111] however there is significant variation among studies in terms of the duration and approach to follow-up monitoring. Consistent between most studies to date, is the use of the endpoint of $>$30 seconds of atrial tachyarrhythmia to define arrhythmia recurrence, but this parameter may underestimate the clinically relevant benefit derived by the patient from ablation. As such, AF burden, defined as the percentage of time in AF, and most accurately measured with implantable cardiac monitors (ICM), has been increasingly employed as an outcome measure. This parameter is associated with symptomatology and quality of life [112, 113] as well as hard clinical outcomes including heart failure, stroke and mortality [57, 114, 115, 116] and may represent a more meaningful marker of success post ablation [117]. In the CAMERA-MRI study (Catheter Ablation Versus Medical Rate Control in Atrial Fibrillation and Systolic Dysfunction), ICM (implanted at time of procedure)-determined atrial tachyarrhythmia burden at 6 months post ablation was 1.5% in persistent AF patients with high rate of prior cardioversion, despite an overall freedom from recurrence of only 56% [118]. Similarly in the CASA-AF trial (Catheter ablation vs. thoracoscopic surgical ablation in long-standing persistent atrial fibrillation), 77% of long-standing persistent AF patients undergoing ICM implantation at the time of ablation, demonstrated a reduction in AF burden $>$75% despite an only 28% freedom from recurrence [119]. The CLOSEMAZE trial (Impact of Catheter Ablation on Arrhythmia Burden in Patients With Shock-Resistant Persistent Atrial Fibrillation) assessed the impact of catheter ablation on AF burden through ICMs implanted 2 months pre-procedure and noted a median atrial tachyarrhythmia burden reduction of 100%, despite an overall modest arrhythmia free survival of 55%, adding weight to the argument that conventional arrhythmia-free survival analysis does not capture the true impact of catheter ablation in this patient cohort [120] (article in press, JACC EP 2023, Fig. 3).

Fig. 3.

Results from the Closemaze trial in persistent AF patients undergoing patient-tailored catheter ablation. (Left panel) patient population and monitoring regime. (Right panel) Plot demonstrating ATA burden on implantable cardiac monitoring before and after ablation, with a significant reduction in burden post-ablation. ICM, implantable cardiac monitor; AF, atrial fibrillation; ATA, atrial tachyarrhythmia; PVI, pulmonary vein isolation.

The diverse range of atrial remodeling observed in patients with persistent AF indicates that a personalized or “patient-tailored” approach to ablation may be more suitable than a generic “one size fits all” approach for this patient population. This approach is employed to some degree in the ERASE and DECAAF II trials whereby ablation strategy is dependent on the location and extent of low voltage or fibrotic regions on MRI. The heterogeneity of ablation strategies employed in these trials can be a limitation when interpreting results, however. In our centre, the choice of ablation strategy at index procedure in persistent AF patients is influenced by AF burden and the presence of parameters suggestive of non-PV triggered AF. Patients without significant atrial remodelling and with self-terminating AF are classified as having ‘pseudo’ persistent AF and undergo PVI only without adjunct ablation [3]. Those who do not fulfill the above criteria are classified as having truly persistent AF and undergo additional substrate ablation as an adjunct to PVI with linear ablation at the roof and mitral isthmus as well as Vein of Marshall ethanolisation. Ongoing randomized trials of patient tailored ablation and improved patient selection for adjunct strategies will be needed to optimise results in those with more advanced disease.