Spatial characterization of the tachycardia circuit of atrioventricular nodal re-entrant tachycardia

Abstract

Aims

The exact circuit of atrioventricular nodal re-entrant tachycardia (AVNRT) remains elusive. To assess the location and dimensions of the AVNRT circuit.

Methods and results

Both typical and atypical AVNRT were induced at electrophysiology study of 14 patients. We calculated the activation time of the fast and slow pathways, and consequently, the length of the slow pathway, by assuming an average conduction velocity of 0.04 mm/ms in the nodal area. The distance between the compact atrioventricular node and the slow pathway ablating electrode was measured on three-dimensionally reconstructed fluoroscopic images obtained in diastole and systole. We also measured the length of the histologically discrete right inferior nodal extension in 31 human hearts. The length of the slow pathway was calculated to be 10.8 ± 1.3 mm (range 8.2–12.8 mm). The distance from the node to the ablating electrode was measured in five patients 17.0 ± 1.6 mm (range 14.9–19.2 mm) and was consistently longer than the estimated length of the slow pathway (P < 0.001). The length of the right nodal inferior extension in histologic specimens was 8.1 ± 2.3 mm (range 5.3–13.7 mm). There were no statistically significant differences between these values and the calculated slow pathway lengths.

Conclusion

Successful ablation affects the tachycardia circuit without necessarily abolishing slow conduction, probably by interrupting the circuit at the septal isthmus.

Proposed model of the AVNRT circuit and site of successful ablation (RF lesions indicated by red arrow). The left inferior extension is derived from the atrioventricular canal myocardium and is low in connexin C43 expression, thus being capable of only slow conduction. The right inferior extension could either be slowly or rapidly conducting, since it incorporates both the primary ring and the atrioventricular canal myocardium, and is an area of higher C43 expression. CS, coronary sinus; FO, foramen ovale; LI, left inferior extension; RI, right inferior extension; S, superior ‘last’ input; TV, tricuspid valve. Figure modified from Katritsis.⁶

Open in new tabDownload slide

Introduction

The circuit of atrioventricular nodal re-entrant tachycardia (AVNRT) remains unknown, both in terms of its exact anatomic location and its dimensions. Evidence derived from animal and human electrophysiology studies supports the notion that the inferior extensions of the node are the substrates of the slow pathway.^1,2 These speculations are supported by post-mortem studies of hearts subsequent to ablation.^3,4 A recent histologic study has now suggested that the connections to the compact node through the working myocardium of the atrial septum could serve as part of the fast pathway.^5,6 No data exists, however, regarding the size of these extensions, nor electrophysiological estimates of the length of the involved pathways. In consequence, the exact location and dimensions of the tachycardia circuit remain unknown.

We have previously reported on a method for calculating the activation time over the slow pathway in patients with co-existent types of typical and atypical AVNRT.⁷ Animal and human studies, furthermore, have provided data about the conduction velocity in the area of the atrioventricular (AV) node and its inferior extensions, and in working atrial myocardium.^8–11 On this basis, we hypothesized that, in patients with co-existent types of typical and atypical AVNRT, it should be theoretically feasible to calculate the approximate length of the slow pathway.

Our study, therefore, aimed at calculating the activation times of the slow and fast pathways, and subsequently deriving the estimated length of the slow pathway in patients with co-existent types of typical and atypical AVNRT. We then compared the length of the slow pathway to the distance between the compact AV node and the ablating electrode as derived from three-dimensional reconstruction of fluoroscopy images. For comparative purposes, we also measured the length of the histologically discrete right inferior nodal extension in 31 human hearts. To the best of our knowledge, this is the first human study to provide insights into the anatomical relationship of ablation lesions with the substrate of the slow pathway as defined by electrophysiologic measurements and histological studies.

Methods

Patients

Data were collated from adult patients with AVNRT undergoing catheter ablation at five centres, Beth Israel Deaconess Medical Center, Boston, MA, USA (2009–2013), Rhode Island Hospital, Providence, RI, USA (2007–2011); Athens Euroclinic and Hygeia Hospital, Greece (2007–2019); the Johns Hopkins Hospital, Baltimore, MD, USA (2011–2014); and the University of Michigan Health System, Ann Arbor, MI, USA (2009–2014). Included patients had to have both typical and atypical AVNRT induced during the same procedure, along with the availability of the intracardiac tracings from that procedure. All patients were studied in the post-absorptive state, under conscious sedation, and after all antiarrhythmic agents had been discontinued for more than five half-lives. No patient had received amiodarone for the preceding 3 months. The study received approval from each centre’s institutional review boards.

Electrophysiology study and catheter ablation

AVNRT was diagnosed by standard criteria,¹² and subsequent abolition of the tachycardia was accomplished by electrogram-guided anatomic ablation of the slow-pathway. Typical (slow-fast) AVNRT was defined by an atrial-His/His-atrial ratio (AH/HA) >1, and HA interval ≤70 ms. Atypical AVNRT was defined by delayed retrograde atrial activation with HA >70 ms. Differentiation between various atypical subtypes was accomplished according to the AH/HA relationships.¹²

Details of our ablation procedure have been published elsewhere.⁷ In brief, the ablation catheter was positioned at the inferior (posterior) part of the tricuspid annulus until an A/V ratio of <1 was recorded, and the atrial electrogram was delayed relatively to the atrial electrogram recorded at the His bundle. Radiofrequency current was delivered when multicomponent signals or separate, low-amplitude potentials were obtained. Care was taken to keep the ablation catheter below the ostium of the coronary sinus as visualized in the right anterior oblique projection. Ablation was not performed at the mid or superior paraseptal areas. Endpoints of the procedure were non-inducibility of the arrhythmia, usually after ablation-induced junctional rhythm, and despite isoproterenol challenge.

Calculation of slow pathway activation time

Details of our methodology for measurements of intervals during tachycardia have been described elsewhere.⁷ In order to ensure that the same limb was used for slow anterograde and retrograde conduction, that is in both typical and atypical tachycardia, only patients displaying atypical AVNRT with a clear HA>AH pattern and AH <200 ms were considered for comparative purposes (Figure 1). Assuming that conduction velocity over the slow pathway is essentially identical in both the anterograde and retrograde direction (i.e. Sa = Sr), as is the case in the AV nodal area,⁸ the following relationship is true (Figure 2):

HA (atypical) + AH (typical) = (Sr + A-H) + (Sa + H-A) = Sr + Sa = 2S.

As a corollary of this relationship, activation time of the slow pathway can be calculated by subtracting S from the tachycardia cycle length.

Calculation of slow pathway length

Studies have provided evidence on the conduction velocity in the area of the AV node and its inferior extension, calculating it to between 0.069 and 0.162 m/s in perfused canine and rabbit hearts.^8–10 In the human heart, mathematical modelling of the conduction velocity in these areas has provided a value of 0.04 m/s.¹¹ Assuming an average value of 0.04 mm/ms for the living human heart, therefore, the approximate length of the slow pathway can be calculated using a simple equation.

Measurement of compact AV node to ablating electrode distance

A recent anatomical study has provided evidence about the position and size of the AV node and the atrioventricular conduction axis in the human heart.¹³ Based on these findings, and assuming that the His potential represents the penetrating bundle, the AV node lies 2–3 mm behind and below it. Three-dimensional measurements of distances were accomplished with the use of a reconstruction algorithm developed by our group.¹⁴ The algorithm is based on the concept of epipolar geometry. It requires acquisition of angiographic images of the structure under investigation in two projections that diverge by more than 30°. Based on the acquired 2D angiographic images, and the information embedded in their DICOM header (c-arm angulation, source to image receptor distance, etc.), this algorithm allows accurate three-dimensional reconstruction of cardiovascular structures. The method has been successfully validated against computed tomography images and virtual phantoms and has been used for the reconstruction of coronary arteries and study of the anatomic characteristics of atherosclerotic lesions.¹⁴ In the current study, we used the algorithm to assess the angiographic images acquired during ablation, reconstructing the centrelines of the ablation and diagnostic catheters. This allowed accurate measurement of the proximity of the ablation electrode to the AV node as derived from the His catheter (Figure 3). Measurements were accomplished in both diastole and systole, and an average value was registered.

Histological study

For the histological study, we analysed 31 hearts from an adult population, albeit that the medical history of the donors was unknown for us. Because of questions posed to their close family, however, we have no reason to suppose that they were suffering from arrhythmic cardiac abnormalities. The donors had an age 56.7 ± 10.0 years, with a range from 43 to 68 years. The hearts weighted 350.0 ± 34.5 g, with a range from 279 to 420 g. Of the donors, 18 were of male gender. In all hearts, the full thickness of the triangle of Koch, along with the adjacent aortic root, was removed as a solitary block. The block was then serially sectioned at 12 microns thickness, cutting at right angles to the hinge of the septal leaflet of the tricuspid valve. The sections were then stained using Masson’s trichrome technique. For the purposes of this investigation, we selected 31 of these data sets in which the block removed had been of sufficient size also to include the atrial myocardium making up the cavotricuspid isthmus. When analysing the sections from these data sets, we calculated the distance in millimetres of the lower rightward extension of the atrioventricular node within the tricuspid vestibule, as judged on the basis of its histological characteristics, taking the superior extent of the node as the site of penetration of the atrioventricular conduction axis into the insulating tissues of the atrioventricular junctions (Figure 4). The protocol for this study was reviewed and approved by Ethical Committee of University of Extremadura, reference number: n°28/2018. All work was carried out within the provisions of the 1995 Declaration of Helsinki.

Statistical analysis

Continuous, normally distributed variables are presented as mean ± standard deviation and categorical data are expressed as frequencies (percentages). Data normality was analysed using the Shapiro–Wilk test. Mean values between groups were compared using the Student’s t-test. All reported P-values were based on two-sided tests and were compared with a significance level of 5%. Statistical calculations were performed with IBM SPSS Statistics version 22 (IBM Corporation, Armonk, NY, USA).

Results

Patients

A total of 1347 patients with AVNRT were studied at Beth Israel Deaconess Medical Center, Rhode Island Hospital, Boston, MA, USA (n = 188); Athens Euroclinic, Greece (n = 335); the Johns Hopkins Hospital, Baltimore, MD, USA (n = 271); and the University of Michigan Health System, Ann Arbor, MI, USA (n = 553). Using the criteria mentioned above, 21 patients had both typical and atypical AVNRT during the electrophysiology study. Among these 21 patients, 14 patients (66%) displayed atypical AVNRT with characteristics compatible with AH<HA and AH <200 ms. The mean age of these patients was 48.1 ± 11.9 years (range 32–75), and 6 patients (43%) were female. In all patients, both tachycardias were abolished following ablation of the anatomical slow pathway.

Activation time and length of slow pathways

Conduction intervals during tachycardias are shown in Table 1. The calculated activation time of the slow and fast pathways were 268.8 ± 32.4 ms and 101.9 ± 23.5 ms, respectively. The calculated length of the slow pathway was 10.7 ± 1.3 mm (range 8.2–12.8 mm) (Table 2).

Compact AV node to ablating electrode distance

Three-dimensional reconstructed images were derived from five patients during diastole and systole in two projections. The average distance between the compact AV node and the ablating electrode was 17.2 ± 1.6 mm for the studied patients (Table 2). This distance was significantly longer compared to the estimated length of the slow pathway (P < 0.001). For comparative purposes, this distance was also calculated in 17 patients subjected to slow pathway ablation for typical AVNRT. Derived measurements (17.8 ± 0.9 mm) were not different than those derived from the 5 previous patients (P = NS).

Histological study

The distance of the right inferior extension of the atrioventricular node within the tricuspid vestibule, as judged histologically, and as measured from the site of penetration of the conduction axis in 31 hearts, was 8.2 ± 2.4 mm, with a range from 5.2 to 13.6 mm. In 8 hearts (26%), this distance measured more than 9 mm. We found no statistically significant differences between these values and the measured slow pathway in 14 patients who were subjected to electrophysiology study.

Discussion

Our results indicate that the AVNRT circuit, as identified by our findings, occupies an area of several squared centimetres adjacent to the AV node. Such an arrangement is compatible with the feasibility of resetting and entrainment of AVNRT with extrastimuli, and pacing close to the circuit from both the atrium and the ventricle. It is also in keeping with the observed wide variation of retrograde atrial activity during typical and atypical tachycardia.¹² The length of the slow pathway, as calculated using our electrophysiological results, is also close to the measured length of the right nodal inferior extension as judged on its histological recognition in specimens obtained from human hearts. This observation provides further evidence for the role of the inferior extensions as the anatomical substrate of the slow pathway. In addition, it indicates that the usual successful ablation site at the septal isthmus may not affect the inferior extension itself. This is compatible with previous data showing that abolition of dual AV nodal conduction is not a prerequisite for a successful ablation procedure.¹⁵ As we have previously shown, the credible sign for delivery of radiofrequency energy at the correct site is the evoked junctional rhythm, indicating that a structure in anatomical continuity with the AV node was thermally stimulated.¹⁵

Histopathologic examination of the areas following successful ablation at the presumed anatomical site of the slow pathway has demonstrated distinct endocardial thickening on the crest of the underlying left ventricular septal surface, without affecting the node.^3,4 Furthermore, an ice-mapping study during surgery has also demonstrated that the perinodal atrial tissue is a requisite part of the tachycardia circuit.¹⁶ We infer, therefore, that by successful ablation we interrupt, at some point along its length, the AVNRT circuit, which seems to contain the septal isthmus between the coronary sinus ostium and the tricuspid valve. In this way, we create a blocking line in the septal isthmus between the mouth of the coronary sinus and the tricuspid valve. This does not permit the continuation of the arrhythmia, but does not necessarily affect the inferior extension itself (Figure 5). An alternative explanation might be that the inferior extensions of the node participate in the circuit,^6,17 but it is the proximal end of the extension at the junction with the node that is mainly responsible for slow conduction. Both inferior extensions may continue in the form of the rings of myocardium that have been described in animals in an immunohistochemical study.¹⁸ Of note, within our conceptual model on which our calculations were based, the slow pathway represents not only conduction along the inferior extension, but also involves the node itself.⁶

We have identified structures that may represent the anatomical substrate of the slow and fast pathways,^1–6 but the described atrial inputs, inferior and superior (the so-called ‘last’ one⁵), represent ‘dead ends’ and not the entire circuit. We can speculate that the complete circuit requires capture and excitation of the very close atrial or transitional tissue that is capable of conduction fast enough to support the tachycardia. An alternative explanation might be involvement of the so-called ‘ring tissues’, and the ‘retroaortic node’, in the circuit.¹⁸ As yet, no data are available regarding the electrophysiologic function and conduction properties of these structures.

Another study has provided similar data about the length of the slow pathway as derived from fluoroscopy images.¹⁹ In this report, however, the site of ablation was arbitrarily accepted as lying within the vicinity of the slow pathway, and conduction times were calculated by assuming that the AH interval represents slow pathway conduction in AVNRT, an assumption that is not correct.^7,20

Clinical implications

Our study has important clinical implications. First, attempts at entrainment should be close to the described circuit. Second, right-sided ablation should be aimed at the septal isthmus between the mouth of the coronary sinus and the hinge of the septal leaflet of the tricuspid valve. Higher, mid-paraseptal lesions in difficult cases should not be performed. Third, successful ablation interrupts the tachycardia circuit, not the slow pathway itself. Prolonged and extensive procedures for the abolition of antegrade dual conduction that may inadvertently result in nodal damage are not necessary.

Study limitations

Our calculations of the conduction velocity of the nodal and atrial tissues are derived from mathematical modelling of the human heart conduction properties, based on samples from human cadaveric hearts. Thus, our results should be considered as providing a rough estimate of the size of the pathways in the circuit. Our sample is limited to 14 patients. A larger number might provide a different comparative estimate than the one derived in our study. We could not retrieve fluoroscopic images with DICOM data in order to provide estimates of the actual distance of ablation lesions in all studied patients. Finally, our results apply to patients capable of displaying both typical and atypical AVNRT. Whether this is applicable to patients with typical only, slow-fast tachycardia, cannot be deduced from our study.

Conclusion

In conclusion, our study provides a novel method for estimating the length of slow and fast pathways, and, consequently, the size of the tachycardia circuit in typical AVNRT. Our results correlated with measured lengths of the right nodal inferior extension in histology specimens from human hearts, thus further supporting the role of the inferior extensions as the anatomical substrate of the slow pathway. Comparison of these values with the distance of the ablating electrode from the compact AV node indicated that a successful procedure interrupts the circuit without necessarily abolishing slow pathway conduction.

Conflict of interest: none declared.

Data availability

Data are available on request from the corresponding author.

References