A unified theory for the circuit of atrioventricular nodal re-entrant tachycardia

Abstract

Atrioventricular nodal re-entrant tachycardia (AVNRT) is the most common regular tachycardia in the human, but its exact circuit remains elusive. In this article, recent evidence about the electrophysiological characteristics of AVNRT and new data on the anatomy of the atrioventricular node, are discussed. Based on this information, a novel, unified theory for the nature of the circuit of the tachycardia is presented.

Introduction

Atrioventricular nodal re-entrant tachycardia (AVNRT) represents the most common regular tachycardia in the human.¹ It is amazing indeed, that 65 years after having established a relation of this arrhythmia with dual atrioventricular (AV) nodal conduction, the circuit is still elusive.² The time-honoured concept of dual AV nodal pathways as the substrate for AVNRT provided explanations for many aspects of its electrophysiological behaviour, but several observations were incompatible with this model, and these two pathways had not been demonstrated histologically.³ In 1998, Becker et al.,^4–6 established the presence of the inferior extensions of the AV node in the human heart, as initially described by Tawara in 1906, and speculated that these extensions and the inputs they facilitate may be involved in slow pathway conduction. The role of the inferior extensions was subsequently supported by animal and human electrophysiology and ablation studies,^7–9 and models based on this concept have been proposed for all forms of the arrhythmia.^7,10–12

The identification and characterization of the inferior extensions mainly addressed the issue of the location of the slow pathway. The substrate of the so-called fast pathway still remained obscure since discrete tracts that constitute insulated pathways had not been histologically demonstrated in the human. Animal studies have demonstrated histologic and electrophysiologic evidence of multiple atrial inputs to the AV node.^13–16 In the human, there had also been electrophysiologic evidence for atrio-nodal and atrio-hisian connections, but their role in the AVNRT circuit was not obvious.^17–19 Furthermore, histologic proof of their existence was lacking. Recently, Anderson et al.²⁰ identified ubiquitous connections to the compact node through the working myocardium of the atrial septum. Connections were composed of ordinary myocardium, and transitional cells as forming the ‘bridge’ between the septum and the body of the node were identified in only a minority of the hearts examined. Usually, these myocardial continuities, and especially the ‘last’ one before the node becomes insulated as the bundle of His were provided by the left-sided, or deep, layer of the septum, but it could also take their origin from the superficial, or rightward, side of the septum. Their superior location and composition of working myocardium made these inputs a suitable substrate of the so-called fast pathway.

This new information, especially in the context of recent observations about the electrophysiological nature of typical and atypical forms of AVNRT, compels us to consider a new perspective about the circuit of this fascinating arrhythmia.

Inferior nodal extensions and atrial inputs

The evidence supporting the notion of the inferior extensions as the substrate of the slow pathway in the human has been derived from post-ablation post-mortem studies. Histopathologic examination of the septum following successful ablation at the presumed anatomical area of the slow pathway has demonstrated interruption of a long right inferior atrial extension, and also distinct endocardial thickening on the crest of the underlying left ventricular septal surface, without affecting the node.^21,22 Inoue and Becker⁴ in their seminal publication considered the inferior extensions as showing similar cellular and architectural characteristics to those of the compact AV node, but we know now that two embryonic rings have been described.²³ The so-called primary ring is composed of ‘specialized’ myocardium that encircles the primary interventricular foramen. It expands to encircle the inlet of the right ventricle and the outlet of the left ventricle. The second ring is the AV canal, which is able to produce AV delay prior to the stage that it is possible to recognize ‘specialized’ myocardium. The rings intersect at the sites of formation of the regular AV node and the so-called retro-aortic node, for which we have still to find a function. The rightward inferior extension is part of the primary ring, and retains its ‘specialized’ histological characteristics. The leftward inferior extension is part of the AV canal myocardium. As such, it could be considered to be slowly conducting, but would not be histologically specialised.²³ In this perspective, both extensions can be perceived, therefore, as a plausible substrate of the slow pathway. The right extension, as well as the superior input that is composed of ordinary myocardium,²⁰ could also serve as the substrate of a fast pathway. Still, the issue is anything but settled. An intraoperative ice mapping study localized the fast pathway at several points, including the anatomic position of the right inferior extension (points T3 and T5 in that report).²⁴ Staining and genotyping of connexins (Cx), i.e. the gap junctional proteins that are particularly expressed in the AV junction, has also allowed a three-dimensional reconstruction of the area and identification of specific structures with greater variability in the space constant of tissue, and poor gap junction connectivity due to differential expression of connexin isoforms that result into substantially different conduction properties.²⁵ A connexin genotyping study in four human hearts has identified the right inferior extension as an area of high Cx43 expression and, consequently, faster conduction than the node and the left extension where Cx43 expression was low.²⁶ However, the location of the presumed left inferior extension in this study was rather closer to the node itself rather than the left extension as described in pathology studies.^4,20 It is probable, therefore, that detected differences merely represented the different conduction properties of the node and its extensions. To complicate the issue more, left septal ablation eliminates typical and atypical AVNRT, not only when a right-sided approach has failed, but also as a de novo procedure.^11,27,28 Furthermore, during atypical AVNRT, and especially the slow-slow form, the prolonged AH and HA intervals that are usually recorded suggest the existence of more than one slow pathway. It seems that these conflicting observations merely reflect the remarkable variability of atrial inputs to the node, inferior and superior, as demonstrated by Anderson et al.²⁰ In the seminal publication of Inoue and Becker⁴ as well, the extensions of the AV node in the human heart could be rightward, leftward or both. Considerable variability also exists in the arrangement of the superficial atrial muscle fibres in the area of the triangle of Koch,²⁹ and during typical AVNRT, atrial activation within this area displays anatomic variability and spatially heterogeneous activation.³⁰

Superior atrial inputs

The superior atrial inputs, and especially the “last” one as described by Anderson et al.,²⁰ may participate in the circuit of typical AVNRT as the fast pathway since they are composed of working, relatively fast conducting atrial myocardium. The described location variability explains the retrograde atrial conduction patterns during slow-fast AVNRT, that may display variable earliest activation either in the right or the left aspect of the interatrial septum.^2,30–33 Theoretically, they could be also serve as fast pathway in the fast-slow, atypical form. However, in patients with co-existent typical and atypical AVNRT, we have shown that this is unlikely, since fast-slow is not an inverse slow-fast using the same circuit in the opposite direction.³⁴ Thus, at least in these patients an additional fast pathway is involved. Furthermore, the AH interval during fast-slow AVNRT approximates 95–100 ms, while in the slow-fast variety the HA interval is 45–60 ms.^10,34 Still, AH intervals as small as 30 ms have been seen in fast-slow AVNRT.³⁴ Thus, in patients with fast-slow AVNRT, any fast conducting inputs may be involved. Theoretically is also possible that, depending on their connexin expression in different patients, superior inputs they could even have a role for slow conduction in the circuit in the occasional patient(s). Electrophysiological demonstration of multiple pathways,^34–36 or discontinuities in the AV node conduction curves,⁷ demonstration of different fast pathway conduction during typical AVNRT and ventricular pacing,³² differentiation between the antegrade and the retrograde slow pathway in patients with atypical AVNRT,³⁷ and the observed difference between antegrade and retrograde fast pathways in patients with coexistent typical and atypical AVNRT,³⁴ clearly argue against the involvement of ‘fixed’ pathways. They rather suggest a dynamic environment with several possibilities among different patients.

Atrioventricular nodal re-entrant tachycardia forms, types, and confusion

The distinction between ‘fast-slow’ and ‘slow-slow’ forms is of no practical significance, since certain cases of atypical AVNRT cannot be classified according to described criteria.^2,38 Even separating typical and atypical AVNRT as different arrhythmia entities is not in keeping with all observed patterns of the tachycardia. We know now that both typical and atypical AVNRT can be induced by either atrial or ventricular pacing,¹¹ atypical AVNRT is ablated exactly at the same site as the typical^11,27 and may also be induced with antegrade conduction jumps,^11,34 whereas typical AVNRT can be induced in the presence of continuous antegrade conduction curves.³⁹ Simultaneous conversion of a typical arrhythmia to atypical, and vice versa, can also be seen in the electrophysiology lab, and occasionally atypical AVNRT may display HA intervals close to 70 ms.^11,34 All forms of AVNRT, typical or atypical may display anterior, posterior and middle or even left atrial retrograde activation patterns.^{2,30–33,40–44} Thus, the distinction between typical and atypical AVNRT is only based on relatively different conducting properties of involved pathways.

If atrial inputs are a normal phenomenon why not all people develop atrioventricular nodal re-entrant tachycardia?

The observed variations of the location, relative length and, probably, connexin expression of both inferior and superior atrial inputs to the node are fundamental for our conception and understanding of the AVNRT circuits. The morphology of the AV node apart from being subjected to random anatomic variation,²⁰ also depends on age.⁵ These facts might explain why AVNRT, as opposed to AVNRT, is not an arrhythmia mainly of the young, or, perhaps, why a higher incidence of atypical, as opposed to typical, AVNRT has been documented in athletes.⁴⁵ It seems, therefore, that it is not the mere presence of such extensions capable of facilitating re-entry. The location, size, orientation and connexin expression of these structures in the vicinity of the triangle of Koch has to satisfy particular requirements in order to allow the electrophysiologic conditions for re-entry to occur. The considerable variation even in the apparently ‘normal’ heart in subjects without arrhythmia, therefore, suggests a clearly probabilistic phenomenon. Thus, although these structures are universal findings in the normal human heart, not all humans develop AVNRT.

A probabilistic conjecture

A conceptual circuit that satisfies discussed observations is not compatible with ‘fixed’ slow and fast pathways in certain anatomic locations. The circuit appears to be just confined within variable superior and inferior atrial inputs. Depending on the relative conduction properties, any input can contribute to the fast or slow component of the tachycardia circuit. Different anatomic locations, orientation, length and, perhaps, connexin expression between all atrial inputs allow them to behave as the relatively slower or faster conduction limb than the others. Since the circuit involves part of the node itself, none of the inputs alone comprises the so-called slow pathway. A conduction jump simply reflects refractoriness of a working input and taking over by a relatively slower one that delays conduction in a way that the circuit starts operating by also involving the node as part of the slow pathway.

The right inferior extension because of its size and structure as part of the primary ring is the most probable, but not necessarily the only one, to participate in the circuit as part of the slow pathway, and represents the easier target to ablate. By ablating in the anatomical area of the inferior extensions we do not necessarily abolish slow pathway conduction, since abolition of dual AV nodal conduction is not a prerequisite for a successful procedure.²⁷ We may therefore deduce that we simply modify one of the inferior atrial inputs, that is involved in the circuit. With the usual right septal approach this could be either the right or even an accessible left extension (through a deep right septal lesion). The lesions interrupt or critically delay conduction in an extension (or extensions) that may well serve as the slow or fast component of the circuit, depending on the form of AVNRT. Nevertheless, a successful procedure certainly does not abolish superior inputs, and consequently normal AV nodal conduction.

If the described atrial inputs, inferior and superior represent ‘dead ends’ how does the circuit close? We can speculate that this happens by capturing and excitation of the very close atrial or transitional tissue that is capable of fast enough conduction to support the tachycardia. Spread of activation through the atrial myocardium to the node in the absence of identifiable sino-atrial pathways also happens during sinus nodal rhythm.⁴⁶ The potential co-existence of AVNRT with atrial fibrillation is a phenomenon similar to that of a ‘fluttering’ right and a ‘fibrillating’ left atrium, or vice versa, that can be recorded simultaneously in the electrophysiology lab. Thus, none of the inputs is a sole fast pathway, since part of the very close atrial tissue also participates in fast conduction over the circuit. Another speculative scenario might involve the rings of conduction tissue that have been described in animals in an elegant immunohistochemical study.⁴⁷ These structures take their origin from the inferior extensions of the atrioventricular node, passing rightward and leftward to encircle the orifices of the tricuspid and mitral valves, and reuniting to form an extensive retroaortic node. However, their electrophysiologic function remains unknown. If they could participate as the slow conducting limb with the 'last' superior extension as the fast conducting one, then the AVNRT circuit might be a well-defined entity. This is a speculative, albeit provoking, hypothesis that deserves further study.

Atrioventricular nodal re-entrant tachycardia therefore may be defined as a re-entrant tachycardia involving the AV node and its extensions, as proven by exclusion of an operating septal accessory pathway with decremental properties.⁴⁸ The nature of the resultant re-entrant circuit depends on the properties of superior and inferior atrial inputs. Figure 1 provides a hypothetical probabilistic model based on the presented unified theory about the AVNRT circuit. All involved inputs and transitional cells are subjected to great anatomical variability. The so-called slow pathway contains a component from the AV node, whereas the fast pathway contains a component of atrial tissue. The right inferior input has a higher Cx43 expression than the node itself, whereas no characterization of connexin isoforms exists for the true left inferior and the superior ‘last’ inputs. Since they are composed of atrial myocardium we could hypothesize that their conduction properties are faster than those of the right inferior input which in most cases may represent the main contributor to the slow pathway. However, this assumption should be considered within the context of the described anatomic variability of all involved structures. Simultaneous right and left circuits or double-loop re-entry entailing both sides cannot be ruled out. A figure-of-8 re-entry with continuous crossing over of antegrade activation through an inferior input to the contralateral superior input via the node might perhaps explain RR alternans that are sometimes observed in AVNRT.⁷

In-depth characterization of the conduction properties of all atrial inputs in hearts with proven AVNRT by means of detailed histochemistry and connexin genotyping of all known 4 connexins, and at strictly specified anatomic sites, are necessary in order to allow Oedipus to solve the myth of the Sphinx.

Acknowledgements

I am indebted to Professor Robert H. Anderson for his insights in reviewing the manuscript, and his incredible assistance and guidance in my comprehension of the complex embryology of the AV nodal structures. Without his constructive support, this manuscript would have not been possible.

Conflict of interest: none declared.

References