Novel concepts and approaches in ablation of atrial fibrillation: the role of non-pulmonary vein triggers

Abstract

Ablation of non-pulmonary vein (PV) triggers is an important step to improve outcomes in atrial fibrillation ablation. Non-pulmonary vein triggers typically originates from predictable sites (such as the left atrial posterior wall, superior vena cava, coronary sinus, interatrial septum, and crest terminalis), and these areas can be ablated either empirically or after observing significant ectopy (with or without drug challenge). In this review, we will focus on ablation of non-PV triggers, summarizing the existing evidence and our current approach for their mapping and ablation.

Introduction

Ablation of atrial fibrillation (AF) has emerged as an alternative to drugs for rhythm control of this commonly encountered arrhythmia. Pulmonary vein isolation has been the mainstay of every AF procedure since the recognition of importance of the pulmonary veins (PVs) for initiation of AF in the late 1990s.¹ During the course of the 2000s, PVI has moved antrally pulmonary vein antral isolation (PVAI), both to reduce the risk of complications (namely, PV stenosis) and increase the success rate of the procedure, as the antrum shares the same embryological origin as the PVs.² While PVAI is a highly effective procedure in paroxysmal AF patients, in whom spontaneous PV firing is frequently the only trigger for AF paroxysms, by itself it is not enough in patients with non-paroxysmal AF.³ As we move forward, performing ablation in patients with persistent (PrAF) and long-standing persistent atrial fibrillation (LSPAF), other targets should be sought and, in the recent year, research efforts have focused on finding the best ablation approach in these populations. In general, when AF persists, electrophysiological and structural changes of the atrial myocardium predispose to the development of triggers originating from sites other than the PVs (the so-called non-PV triggers), and make the atria the perfect substrate to sustain AF.⁴ Throughout the years, many ablation strategies for substrate modification have been proposed, none of which has been proven to be superior to triggers ablation alone.^5–9 This can be explained on a mechanistic point of view: while substrate determines the persistence of AF, there is no AF without a trigger initiating it. This concept is corroborated by the lack of correlation between procedural AF termination and atrial tachycardia (AT)/AF recurrence in non-paroxysmal AF patients: despite effective substrate modification, without targeting non-PV sites that de facto trigger AF, outcomes do not improve.^10,11 In this review, we will focus on ablation of non-PV triggers, summarizing the existing evidence and our current approach for their mapping and ablation.

Substrate-based ablation

Over the last two decades, many ablation strategies for substrate modification have been proposed: from linear lesions to compartmentalize the left atrium, to ablation of complex fractionated electrograms (CFAE) or autonomic ganglionated plexi, and scar homogenization.^12–15 Despite their widespread use, none of these approaches has been found to be superior to triggers ablation in patients with both paroxysmal and non-paroxysmal AF.^5–7,9 In the recent years, substrate-based ablation strategies have focused on mapping of localized organized re-entrant activity thought to drive AF, such as rotors, high frequency areas, and subsequent iterations.^16–21 While this approach is gaining traction, there are some important limitations that needs to be addressed.²² Most importantly, AF drivers are identified by means of inadequate mapping, either because it’s non-simultaneous or has low-resolution, requiring significant post-processing by proprietary software. Indeed, stable rotor-like re-entries or other forms of focal AF drivers have not been consistently identified in studies using simultaneous high-density mapping or alternative computational methods.^23–28 Moreover, as with other forms of substrate modification, ablation of AF drivers has only been evaluated in few small studies with a relatively short follow-up, showing contrasting results and an overall lack of benefit when compared to PV isolation or triggers ablation.⁸ Therefore, before incorporating these approaches into standard clinical practice, further experimental and clinical validation are needed.

Ablation of non-pulmonary vein triggers

Non-pulmonary vein triggers are ectopic beats triggering AF, originating in areas other than the PVs. They usually cluster in specific regions such as the left atrial posterior wall (PW), other thoracic veins [superior vena cava (SVC), coronary sinus (CS), vein of Marshall)], crista terminalis (CR), interatrial septum (IAS), and the left atrial appendage (LAA). These structures have myocardial cells that retain the ability to automatically depolarize or serve a substrate for micro-re-entry due to their rapid conduction, thus serving as independent triggers for AF.^29,30 The prevalence of non-PV triggers is variable among different studies, as it depends on the protocol used to induce them and the definition adopted for ‘significant’ non-PV triggers. A low yield of non-PV triggers is common in those studies using low-dose (<10 mcg/min) or stepwise increasing doses of isoproterenol; moreover, some others consider significant non-PV triggers only those that reproducibly initiate AF.³¹ However, when performing AF ablation under deep sedation or general anaesthesia, a high-dose isoproterenol infusion is necessary to induce non-PV triggers and even repetitive isolated premature atrial contractions (PACs) and non-sustained focal atrial tachycardias, in addition to beats triggering AF/atrial flutter (AFL), appear to be clinically relevant.³² As such, non-PV triggers can be observed in up to 60% of patients with AF,³³ being more prevalent in patients with non-paroxysmal AF, female gender, obesity, sleep apnoea, older age, low left ventricular ejection fraction, severe left atrial scarring, hypertrophic cardiomyopathy, and mechanical mitral valve (Table 1); in these cohorts, non-PV triggers should be targeted at the time of the first procedure.^9,34–42 Moreover, non-PV triggers are responsible for the late recurrences observed in patients with paroxysmal AF and persistent PV isolation; in this cohort, they should be sought and targeted at the time of the repeat procedure.⁴³

Ablation of non-PV triggers can be empirical or performed after induction using high-dose isoproterenol (an infusion of 20–30µg/min for 10–15 min, with concomitant adequate pressure support). For the latter approach, it is important that any antiarrhythmic drug is stopped at least 5 half-lives before the procedure, to minimize the chance of non-inducibility. In our laboratory, mapping non-PV triggers is guided by multiple catheters positioned along both the right and left atrium (Figure 1): a 10-pole circular mapping catheter (CMC) in the left superior PV recording the far-field LAA activity (to avoid mechanical ectopies), the ablation catheter in the right superior PV that records the far-field IAS, and a 20-pole linear catheter with electrodes spanning from the SVC, right atrium/CR to the CS. With this simple catheter setup, when focal ectopic atrial activity is observed, their activation sequence is compared to that of sinus rhythm, allowing to quickly identify their area of origin:

• beats originating from the right atrium: earliest activation in the proximal duo-decapolar catheter; the specific activation sequence varies depending on the site of origin, resembling that of sinus rhythm for ectopies originating from the SVC (Figure 1, green).

• beats originating from the IAS area: both the CS and proximal duo-decapolar catheter are early, usually preceded by the far field atrial activity recorded with the ablation catheter (Figure 1, magenta).

• beats originating from the CS: earliest activation in the distal duo-decapolar catheter (Figure 1, blue).

• beats originating from the LAA: far-field activity in the CMC is early, preceding that recorded in the distal CS (Figure 1, yellow).

For significant non-PV triggers (repetitive isolated PACs, focal atrial tachycardias or beats triggering AF/AFL), a more detailed activation mapping can be performed in the area of origin.

In our laboratory, every patient, including the paroxysmal AF cohort, undergoes empirical PVAI, left atrial PW isolation, and SVC isolation. Moreover, every patient is challenged with high-dose isoproterenol to elicit other triggers, that are targeted if deemed significant. In patients with an history indicative of a higher prevalence of non-PV triggers (see Table 1), we also perform empirical CS isolation and extend left atrial LA ablation along the septum (anterior to the right PV antra) and the inferior wall of the left atrium down to the CS, to empirically address IAS and additional left atrial PW triggers. Empirical LAA isolation at the time of the first procedure is usually reserved for LSPAF patients, who commonly display severe scarring in the PV/left atrial PW (thus less likely to trigger AF).

In general, non-PV triggers are targeted with focal ablation, exception being the triggers originating from the SVC, LAA, or CS, in which cases complete isolation of these structures is the ablation strategy of choice. In the next sections, we will review the main non-PV triggers in more detail (see Table 2 for a selection of studies published on catheter ablation of non-PV triggers).

Left atrial posterior wall

The left atrial PW is that area of the left atrium located between the right and left PVs. From an embryologic, anatomic, and electrophysiological standpoint, it should be considered an extension of the PVs.⁴⁴ Over the years, many surgical and transcatheter ablation studies have shown that effective PW isolation improve outcomes both in paroxysmal and non-paroxysmal AF patients.^45–52 Therefore, empirical isolation of the left atrial PW should be performed in all patients undergoing AF ablation.

After completing PVAI, left atrial PW isolation can be achieved either by linear ablation (box lesion) or electrogram-based ablation. With linear ablation, a roof and floor line are created to connect the superior and inferior PVs, respectively. With transcatheter radiofrequency (RF) ablation, it is not easy to create a single continuous transmural lesion, and even a single gap in the ablation line results in reconnection of the whole PW, affecting the long term success rates.^53–55 In electrogram-based ablation, all potentials identified moving the CMC along the PW are sequentially targeted (Figure 2A): reconnection of the whole PW is unlikely, and usually limited to areas close to the oesophagus, where RF energy delivery is limited. The endpoint is to achieve electrical isolation, as documented by the absence of any electrical activity in the PW (Figure 2C and D). For patients with non-paroxysmal AF, PW isolation is usually extended more anteriorly (anterior wall and mid-septal areas) and inferiorly (floor of the left atrium down to the endocardial aspect of the CS) (Figure 2B).

Of note, the PW is in close relationship with the oesophagus: to reduce the risk of oesophageal damage and atrio-oesophageal fistula, it is important to know its location during the procedure (with intracardiac echocardiography, contrast oesophagography, or using an oesophageal probe), use real-time temperature monitoring, and titrate power and/or contact force while moving the catheter quickly along the entire PW. It is also important to remember that the oesophagus is wider than the temperature probe, therefore sometimes no or minimal temperature change is recorded, despite ablating over the oesophagus—thus, if no additional real-time imaging of the oesophagus is available, it is a good rule to move quickly whenever ablating in the PW regardless of the location of the oesophageal probe.

Superior vena cava

A common site of non-PV triggers is the SVC. Empirical SVC isolation is a reasonable addition to PVAI in patients with paroxysmal AF and should be performed in every patient demonstrating firing originating from the SVC.^56–62 As for the PVs, SVC isolation is preferred to focal ablation of triggers arising from the muscular sleeves, which is time consuming, not as effective, and carries the risk of SVC stenosis with secondary SVC syndrome.

To perform SVC isolation, the CMC is positioned at the junction between the right atrium and SVC, with potentials resembling that of the PVs, with a sharp vein potential superimposed to a blunt far-field atrial electrogram (Figure 3C). Intracardiac echocardiography can help define the SVC-right atrial junction by visualizing the CMC between the right superior PV and the right pulmonary artery (Figure 3A). SVC isolation can be achieved with a segmental approach starting on the septal aspect (Figure 3B). When targeting the lateral aspect, it is important remember its relationships with the phrenic nerve and the sinus node. Before ablation, phrenic nerve mapping with high-output bipolar pacing is performed; if the patient is under general anaesthesia, paralytic agents should be avoided or pacing performed after an adequate wash-out, confirming the presence muscle contraction with a nerve stimulator. Given the potential risk of phrenic nerve injury, in up to 10% of patients complete isolation is not possible, and an alternative approach—targeting the right atrial posterior wall, can be employed.⁶³ Another possible complication is damage of the sinus node, which lies laterally below the SVC: if acceleration of the sinus rate is observed during ablation (a sign of impending sinus node injury), RF energy should be discontinued. As for the PVs, endpoint is complete isolation with documented entrance conduction block (Figure 3D).

Special mention merits to the subset of patients with left persistent SVC (LPSVC). Its isolation should be considered as first line treatment in all patients, given its arrhythmogenicity and close relation with other important left atrial structures (left PVs and LAA).^64,65 Isolation of the LPSVC is guided by the CMC that can be easily advanced through a dilated CS reaching the level of the left superior PV/LAA ridge (Figure 4). The CMC is continuously pulled back while performing electrogram-guided ablation: LPSVC isolation is usually performed along with CS isolation (see below), with the endpoint of obtaining dissociation or abolition of sharp potentials along the two structures. In these patients, the CS is dilated and the CMC might not be in good contact with the entire vessel wall for any given position: for this reason, it is important to manipulate the catheter around the wall to confirm the absence of sharp vein potentials. Moreover, careful left phrenic nerve mapping as well a monitoring the oesophageal temperature should always be performed, as the dilated CS and LPSVC are in close relationship with both structures.

Coronary sinus

Non-pulmonary vein triggers originating from the CS are common in patients with PrAF and LSPAF.^66–69 Coronary sinus isolation is preferred to focal ablation, as it eliminates all its potential triggers and has been shown to be associated to better outcomes.^70,71 Coronary sinus isolation is challenging, given the presence of myocardial sleeves that surround it connecting the CS to the left atrium.⁷² Therefore, to achieve complete CS isolation it is important to target the vessel both endocardially and epicardially.

Endocardial ablation is performed in the anterior aspect of the infero-lateral left atrium, with the ablation catheter positioned at the level of the mitral valve annulus, parallel to the one positioned in the CS. Ablation is performed with the endpoint of elimination of local sharp electrograms spanning from the lateral left atrium to the septum, adjacent to the CS os (this usually requires looping the catheter up to 180°; Figure 5A). Epicardial ablation, inside the CS, is started distally and the ablation catheter is continuously dragged back to the CS os (Figure 5B), making sure the catheter tip is freely moving, and not wedged in a small branch to avoid steam pops (any sudden impedance rise from the baseline should prompt RF discontinuation). The procedural endpoint is complete isolation, as demonstrated by dissociation or abolition of any potentials along the CS body (Figure 5C and D). Of note, an important precaution should be taken when ablating proximally in the CS, given the vicinity of the AV node: beat-to-beat PR monitoring is crucial, with immediate RF energy discontinuation in case of PR prolongation. The CS is also in close relationship with the oesophagus, thus real-time oesophageal temperature monitoring should be performed to avoid the possible formation of cardio-oesophageal fistulas.⁷³

Left atrial appendage

Non-pulmonary vein triggers can originate from the LAA, with their prevalence increasing in patients with non-paroxysmal AF.^74,75 Indeed, empirical LAA isolation can be considered as a first-line approach for patients with longstanding persistent AF.^76,78 Again, given the high recurrence rates in patients undergoing focal ablation, LAA isolation is to be preferred.⁷⁹

Left atrial appendage isolation can be achieved with a technique similar to that of PVI, although it is more challenging given the wide variability of the LAA ostial anatomy. The CMC is positioned at the level of the LAA ostium (as easily assessed by intracardiac echocardiography, see Figure 6A) and ablation is performed targeting the earliest LAA activation site, aiming for delay until complete isolation with entrance block is documented (Figure 6B–D). This is preferably done during sinus rhythm, when it’s easier to define the left atrium-LAA breakthrough. Moreover, a contact force sensing ablation catheter is helpful to ensure adequate catheter tip-tissue contact. It is also important not to advance the CMC deep into the LAA to avoid ablation inside the LAA: the wall is thin between the pectinate muscles and the risk of perforation and left phrenic nerve injury are heightened.

With persistent LAA isolation, its mechanical function is impaired despite maintenance of sinus rhythm; therefore, patients should continue long-term oral anticoagulation or be considered for LAA occlusion, unless the mechanical function of both the LAA (good contractility with pulsed-wave Doppler emptying velocities ≥0.4 m/s) and left atrium (consistent A waves at the transmitral pulsed-wave Doppler flow) are preserved as demonstrated with transoesophageal and transthoracic echocardiogram follow-up (usually performed 6 months after ablation).

Other sites

Other common locations of non-PV triggers are the CR, IAS, ligament of Marshall (LoM), mitral and tricuspid valve annuli, but foci can originate anywhere in the atria, especially in patients with atriopathy.^57,80 These triggers are usually targeted with focal ablation after activation mapping identifies the site of origin, the endpoint being termination/non-inducibility of the AT or disappearance of the repetitive PACs. Exception is for triggers originating from LoM, where isolation along the whole ligament is the endpoint of choice. The LoM is an epicardial vestigial fold that marks the location of the embryological left SVC and contains the vein of Marshall, nerves and muscular fibres.⁸¹ It runs from the mid-distal CS through the posterolateral left atrium, up to the epicardial aspect of the ridge between the left superior PV and LAA. It is possible to target the LoM area endocardially by ablation in the mid-lateral left atrium (between the CS and left inferior PV) up to the ridge.⁸² Alternatively, direct ethanol injection in the vein of Marshall has shown to be an effective way to achieve ablation along the whole LoM without significant complications.⁸³

Limitations

As with PVAI, the main limitation of non-PV triggers ablation lies on its durability. A successful procedure depends on the creation of lasting transmural ablation lesions leaving no gap of viable tissue, with efficacy varying by technology. Using contact force feedback and high power (we aim at 7–15 g, using 35–45 W, depending on the location), RF energy delivery is very effective, resulting in less recurrences of the original arrhythmia and a lower incidence of de novo arrhythmias caused by the pro-arrhythmic effect of an incomplete lesion.⁸⁴ Among other energy sources, cryoablation with balloon-based catheters has been used to isolate the posterior wall and the LAA, with promising results^77,85; with the current technology, the main limitation is the inability to target focal sites or small structures, such as the CS. Though surgical ablation has been shown to create more consistently transmural durable lesions, it is an invasive procedure, with limited intraoperative mapping techniques, and—if performed epicardially only—can’t address the majority of sites responsible for non-PV triggers (CS, LAA, IAS).

Another challenge when performing non-PV trigger ablation, is the inability to perform accurate activation mapping when infrequent beats that originating from uncommon sites, that are not efficiently identified by fixed multicatheter mapping as described above. In these cases, point-by-point mapping is not efficient to regionalize the area of interest and non-invasive single-beat electro-anatomical mapping may potentially overcome this limitation, identifying triggers critical to a patient’s AF, even prior to their entry into the EP laboratory; however, further studies are necessary to evaluate the spatial resolution of the available technology when used to place focal triggers.

Conclusion

Ablation of non-PV triggers is an important step to improve outcomes in AF ablation. Non-pulmonary vein triggers typically originates from predictable sites (such as the left atrial PW, SVC, CS, IAS, CR) and these areas can be ablated either empirically or after observing significant ectopy (with or without drug challenge). Mapping of non-PV triggers can be performed quickly with the aid of multielectrode catheters positioned in key areas of the right and left atrium. They can be targeted with focal ablation, exception being the triggers originating from the SVC, LAA, or CS, for which complete isolation is the ablation strategy of choice.

Conflict of interest: none declared.

References