Background The gap phenomenon in atrioventricular (AV) conduction is described as a block that occurs within a range of atrial coupling intervals. This block is assumed to occur between two adjacent parts of the conduction system having different refractory properties; thus, a gap would develop if the functional refractory period of the proximal unit was shorter than the effective refractory period of the distal unit. We describe a new electrophysiological mechanism based on dual pathways electrophysiology of the AV node.
Methods and Results In vitro experiments were performed on isolated superfused rabbit hearts. Standard electrophysiological pacing and recording techniques were used to generate conduction curves. The gap phenomenon was documented in 9 of 14 nodal preparations. With shortening of the atrial coupling interval, antegrade conduction block of the “fast” pathway wave front occurred while this impulse was still retrogradely interfering with slow pathway conduction. That is, the fast pathway wave front prevented propagation of the anterograde “slow” pathway wave front by collision or by creating a refractory barrier. This mechanism produced a gap and the block persisted until, at even shorter coupling intervals, the fast wave front penetration became insufficient and conduction was restored through the released slow pathway wave front. This mechanism was verified in AV nodal preparations with separated inputs, in which independent fast and slow wave fronts could be induced and programmed to collide.
Conclusions Our results established the functional interaction of fast and slow pathway wave fronts as an important electrophysiological mechanism underlying the AV conduction gap. This mechanism may be responsible for a variety of clinically observed conduction discontinuities.
The atrioventricular (AV) nodal conduction curve is the most widely accepted practical tool for evaluating the conduction properties of the AV node. The “normal” AV nodal conduction curve is monotonic and the delay A2H2 prolongs in an exponential or hyperbolic manner as the prematurity A1A2 shortened.1 2 3 4 The discontinuous AV nodal curve (or its counterpart, the H1H2 versus A1A2 curve) was first demonstrated in humans by Schuilenburg and Durrer5 and has been subsequently documented to be a frequent finding.6 7 8 9 10 11 This type of discontinuity, a large rise in AH interval for a small change in AA interval, has been attributed to the existence of dual pathways, that is, slow pathway and fast pathway, within the AV node.12
The phenomenon of conduction “gap” constitutes another form of discontinuity in the AV nodal conduction curve. When initially introduced by Moe et al,13 the term “gap” described a window of coupling intervals within which atrial premature beats did not conduct to the ventricle. Depending on the site of block, the gap has been described as atrial-nodal, nodal, nodal-His, or His-ventricular.14 15 The AV nodal gap referred to conduction block observed within the AV node for a range of atrial coupling intervals while successful conduction was preserved with both longer and shorter intervals.
The classic explanation for the gap phenomenon assumes that conduction fails at a site that forms the boundary between two areas of disparate refractory periods. Specifically, the proximal tissue is proposed to have a functional refractory period that is shorter then the effective refractory period of the distal tissue. Conduction is reestablished at coupling intervals shorter than the gap due to proximal conduction delay that allows recovery of the distal tissues. This explanation, although never directly proven, has been applied to explain the nodal gap phenomenon in both animal and human studies.14 15 16
The purpose of this study was to evaluate the gap phenomenon, considering the dual pathway nature of AV nodal conduction. We hypothesized that interference from wave fronts invading the AV node from different directions may produce a zone of conduction gap and tested this hypothesis in experiments on isolated rabbit atrial–AV nodal preparations. In addition, we applied the observations obtained in the experiments to a model of the AV node with separated atrial inputs to further elucidate this phenomenon. Preliminary report on these studies appeared in abstract form.17
Atrioventricular Nodal Preparation
The experiments were performed in vitro in 18 preparations (14 intact and 4 with separated AV nodal inputs, as explained below) obtained from the hearts of New Zealand White rabbits. Rabbits of either sex weighing 2 to 2.2 kg were anesthetized by sodium pentobarbital injection (50 mg/kg) into an ear vein. After a midsternal incision, the heart was removed and placed in oxygenated (95% O2, 5% CO2) modified Tyrode’s solution (described later) at room temperature. The ventricles and left atrial tissues were discarded. The final preparation (Fig 1A⇓) contained right atrial tissues consisting of the interatrial septum (IAS), the right atrial appendage with the musculae pectinatis (MP), the septal leaflets of the tricuspid valve (TrV), the central fibrous body, and the small portion of the ventricular septum just below the bundle of His.
The preparations were pinned down on a thin silicon disk with the endocardial surface oriented up and kept in a thermostat-controlled glass superfusion system. An initial period of equilibration preceded the placement of the recording and stimulating electrodes (see below). The AV nodal preparations survived and conducted impulses in a stable manner for at least 8 hours.
The solution was prepared with deionized, distilled water and had the following composition and concentrations (in mmol/L): NaCl (128.2), KCl (4.7), CaCl2 (1.3), MgCl2 (1.05), NaHCO3 (20), NaH2PO4 (1.19), and glucose (11.1). It was oxygenated with a mixture of 95% O2 and 5% CO2 and maintained at 35.5°C with a pH between 7.25 and 7.35.
To improve the stability of microelectrode recordings, 2,3-butanedione monoxime (BDM) was added to the basic solutions. Preliminary experiments with BDM in our laboratory as well as others18 did not show any deteriorating effects on conduction. Furthermore, we found that 10 mmol/L BDM not only fully eliminated mechanical motion of the preparations, thus allowing more stable recordings, but also improved stability of conduction. The latter effect was possibly related to favorable changes in metabolic demands.
Electrical Stimulations and Recordings
Bipolar stimulating and recording electrodes were custom made from 0.20-mm Teflon insulated platinum-iridium wire. The interelectrode distance was 0.5 mm. Electrical stimuli (2 ms, twice threshold) were applied at the crista terminalis (CrT) and/or the IAS input sites of the AV node (Fig 1⇑). Electrodes for recording surface bipolar electrograms were placed at these inputs and at the bundle of His. The atrial recording electrodes were located within 1 mm of the stimulating electrode and between the corresponding stimulating electrode and the AV node. This resulted in a stable stimulus-to-response latency of less than 2 ms. All electrodes were mounted on micromanipulators (WPI, M3301) for precise location at sites on the preparation surface.
Leads from the stimulating electrodes were connected to optically isolated stimulator units (WPI, A360). Stimulation was determined by an eight-channel, programmable stimulator (AMPI, Master-8). Leads from the recording electrodes were connected first to high-resistance, differential-input probes located close to the preparation and then to an eight-channel, programmable, computer-controlled signal conditioner (Axon Instruments, CyberAmp 380).
The microelectrodes were made with the use of a Flemming/Brown micropipette puller (Sutter Instruments Co, P-47). After filling with 3 mol/L KCl solution, they were connected through Ag-AgCl wire to the input probes of a two-channel microelectrode amplifier system (WPI, S-7100A). The input probes rested on hydraulically controlled micromanipulators (Narishige, MO150). Microelectrodes were impaled in cells from the N region (“proximal” cells) as well as in cells from the NH region (“distal” cells).
The experimental setup rested on an air-cushioned “floating” table (TMC, Micro-g). Amplified signals were displayed for monitoring on storage oscilloscopes (Tektronix, 5111A, and Nicolet, 310). In addition, they were digitally recorded on tape (Vetter Digital, 4000A) for subsequent analysis with the use of a ZEOS 486 computer and dadisp data analysis software.
The preparations were beating spontaneously at the start of the experiment, and the sinus cycle length was 365±25 ms. The basic paced cycle length was 300 ms in all preparations. After each 30th basic beat, labeled S1, two premature beats were introduced and the cycle then was repeated. The first of the two premature beats, called conditioning beat S2, was introduced with a predetermined coupling interval, S1S2. Several different S1S2 intervals, from 300 to 125 ms, were used to set baseline conditions before introduction of the subsequent test beat. The use of the conditioning prematurity S1S2 permitted the addition of a “refractory burden,” which is then exaggerated by the test stimulus. The latter, labeled S3, was introduced with progressively shorter coupling intervals, S2S3, in each stimulating cycle. The S2S3 was changed from 300 ms in steps of 10 ms at longer S2S3 and 5 ms or less at shorter S2S3 until the effective refractory period of AV node was reached. The latter was defined as the longest S2S3 prematurity (for any given S1S2 conditioning interval) resulting in block. The conduction times S3H3 were measured between the test stimulus and the fastest inscription of the corresponding bundle of His electrogram. The intervals were measured electronically to the nearest 1 ms. These stimulation protocols and measurements were applied either at the CrT or IAS input sites. Besides the S3H3 intervals (conduction times), the output intervals H2H3 between the corresponding bundle of His electrograms were also measured and plotted as a function of S2S3. The shortest observed H2H3 interval (for any given S1S2 conditioning interval) determined the functional refractory period of the AV node.
In four experiments, the atrial inputs at the IAS and CrT entry sites of AV node were separated. The preparations were cut from the base of the coronary sinus in two directions: upward across the CrT toward the MP and downward toward the central region of the AV node (Fig 1B⇑). This procedure was performed during basic pacing at the CrT site and was considered successful when it resulted in a marked prolongation of the S1H1 interval, indicating conduction through the slow pathway alone. In addition, the following conditions existed: (1) anterograde conduction could be initiated by electrical stimulation at either the CrT or IAS input sites, confirming the integrity of the N region; (2) delayed retrograde activation (via the N region) of each input site occurred when the opposite input site was paced; and (3) conduction time with basic pacing at the IAS site, ie, through the fast pathway, was similar to baseline before cutting.
Besides the above described pacing protocols at each input site, combined pacing of the CrT and IAS was used in these preparations. In this case test beat S3 was introduced at the two inputs simultaneously. Three conduction curves, one for each mode of stimulation (ie, CrT, IAS, and combined inputs), were generated.
All procedures used in this study were in strict accordance with the institutional guidelines for the care and use of experimental animals, and the protocols were examined and approved in advance.
Observations in AV Nodal Preparations With Intact Atrial Input Sites
Fig 2⇓ shows a set of typical conduction curves demonstrating the gap phenomenon. Both S3H3 (bottom) and H2H3 (top) curves generated with either CrT (filled symbols) or IAS (open symbols) pacing are illustrated. Note that conduction curves S3H3 crossed at S2S3=110 ms, with the IAS stimulating site yielding shorter delays for all coupling intervals to the right of the cross-point. For S2S3 coupling intervals between 140 and 120 ms there was conduction block, while for both longer and shorter coupling intervals conduction was preserved. The corresponding H2H3 curves showed that during shortening of S2S3, the AV nodal functional refractory period had been reached at S2S3=150 ms, ie, before the start of the gap. This suggested that the gap could not be explained exclusively by postulating a premature arrival of the anterograde wave front into the distal node.
Observations similar to those illustrated in Fig 2⇑ were made in 9 of the 14 preparations studied and are summarized in the Table⇓. In the remaining 5 cases, a gap could not be induced. As seen in the Table⇓, stimulation of either the CrT or the IAS input site could produce the gap phenomenon. Note that in some preparations, a gap could be demonstrated for a wide range of conditioning S1S2 intervals, while in others the gap developed only with shorter conditioning intervals. The width of the gap was also influenced by the choice of conditioning interval and tended to increase with shortening of S1S2.
Fig 3a⇓ further illustrates the development of the gap in the same preparation as in Fig 2⇑. Besides the surface electrograms, a microelectrode was impaled in the N region of the AV node. The timing of recorded action potentials (see below) identified the recorded site as being close to the fast pathway insertion into the compact node. Figs 3A through 3F illustrate the events developing with progressive shortening of S2S3 applied at the IAS input site. The following analysis will focus on the cellular action potentials AP3 in response to the test beat S3. Fig 3b⇓ represents 12 superimposed panels from the same experiment (including the panels shown in Fig 3a⇓) obtained by varying the coupling interval S2S3 from 85 to 140 ms.
With long coupling intervals S2S3≥135 ms (Fig 3a⇑, panel A), conduction proceeded through the fast pathway. The cell was depolarized after the IAS electrogram with a delay that did not change substantially with shortening of the coupling intervals, although a progressive decrease in the action potential amplitude (filled dot) and duration was noticeable. The gap started at S2S3=133 ms (not shown), and conduction block persisted with further shortening of prematurity (panel B), while the fast pathway wave front was still able to depolarize the impaled fiber. At S2S3≤120 ms (panel C) conduction was restored. Note that there was a substantial jump in the S3H3 delay (146 versus 106 ms), and some decrement in the AP3 amplitude (dot) as well as the appearance of a secondary delayed hump (vertical arrow).
These observations suggested the following working hypothesis. At S2S3≤133 ms, the effective refractory period of the fast pathway had been reached. The proximal driving force produced by it was insufficient to ensure activation of the distal NH region and block occurred proximal to the His bundle. This, however, did not preclude the fast pathway wave front from advancing into the N region, as indicated by the substantial amplitude of the recorded action potentials (panel B). This penetration resulted in a local N region refractoriness or, as demonstrated later, even retrograde conduction through the slow pathway, which then prevented conduction of the anterogradely proceeding slow pathway wave front. The net result was “annihilation” of the otherwise potent slow pathway wave front, thus initiating the gap phenomenon. Further shortening of the S2S3 coupling intervals, however, progressively reduced the fast pathway wave front penetration and eventually resulted in a slow pathway wave front breakthrough, ending the gap. It should be noted that the cellular impalement was not at the precise point where the two wave fronts met. Accordingly, one cannot judge from the presented data if there was a collision of two “active” wave fronts or instead, the earlier fast wave front created a refractory zone that was impenetrable by the slow pathway wave front until its final release at S1S2=120 ms.
One piece of evidence for the anterogradely propagating slow pathway wave front with S2S3≤120 ms is the secondary depolarization (vertical arrows) that follows the initial AP3 response (dots) as seen in panels C through F (see also Fig 3b⇑, upward arrow). Note the reverse relationship between the amplitudes of the primary and secondary depolarizations. With shortening of S2S3, the earlier depolarizations (Fig 3a⇑, dots, and Fig 3b⇑, downward arrow) were decremental and eventually disappeared, consistent with a “dying” fast pathway activation, while the secondary humps (Fig 3a⇑, vertical arrows, and Fig 3b⇑, upward arrow) grew from a local electrotonus to a full action potential corresponding to increasing penetration of the slow pathway wave front. As seen in Fig 3a⇑ (panel D), the slow pathway wave front arrived only briefly after the fast pathway wave front (the secondary depolarization started before the first one reached the diastolic membrane potential). One can conclude from this that the actual point of collision was proximal to the point of impalement toward the compact N region. In addition, these secondary responses were frequently associated with reentrant atrial echo beats (panels D and E). These events supported the above hypothesis by showing that the slow pathway wave front not only proceeded anterogradely toward the bundle of His but also retrogradely invaded the fast pathway area (site of recording) progressively deeper at the shortest S2S3 coupling intervals. Such behavior is in accordance with the basic principles of the dual pathway AV nodal electrophysiology.12
As previously mentioned, the microelectrode impalements in Fig 3a⇑ and 3b⇑ were evidently done at some distance from the area of collision of the fast and slow wave fronts. As a result, the “signatures” of the fast and slow waves were identified by the two distinct components of the action potential AP3 (Fig 3b⇑, downward and upward arrows, respectively). The records in Fig 4a⇓ were obtained from a different preparation in which a roving microelectrode was used in an attempt to impale a site closer to the site of the collision. Superimposed traces from the same experiment obtained with 13 different S2S3 coupling intervals in the range 105 to 190 ms are shown in Fig 4b⇓. Same traces are shown in Fig 4c⇓ but are superimposed so that the cellular coupling intervals can be evaluated. As expected, the desired impalement proved to be more distal than in Fig 3⇑, closer to the junction between the dual pathways. Here again, shortening of the S2S3 interval resulted in a gap, starting at S2S3=170 ms (Fig 4a⇓, panel A). Note the progressively decremental amplitude of action potential AP3 (dots), which was documented inside the gap, up to S2S3=140 ms (panels B and C). According to our working hypothesis, this represented the gradually “dying” fast pathway wave front, which up to this point successfully opposed the slow pathway wave front. At S2S3<140 ms, the microelectrode began to register the passage of the slow pathway wave front. Due to “withdrawal” of the competing fast pathway wave front, the action potential amplitude started to grow (panel D) and subsequently, at S2S3=120 ms (panel E), conduction was restored. At S2S3=105 ms (panel F) the slow pathway wave front initiated a reentry beat.
The dynamics of the changes in action potential in relation to test beat S3 can be better appreciated by the analysis of the records shown in Fig 4b⇑. They represent the cellular responses (in the same experiment as in Fig 4a⇑) to test beat S3 for 13 progressively shortened S2S3 coupling intervals. These records were superimposed by synchronizing all S3 beats. Note the development of a gap associated with the diminishing action potential amplitude (dots, downward arrow) and the reversal of this tendency (upward arrow) with the transition to resumed slow pathway conduction. It is worth mentioning that this biphasic course of the action potential amplitude was not related to the local cellular coupling interval in the distal AV node. In fact, the shortest H2H3 interval was achieved at S2S3=190 ms, ie, well before the start of the gap (see Fig 4A⇑). Moreover, as evident from Fig 4c⇑, the amplitude reversal was observed at almost the same diastolic cellular coupling interval, suggesting that the applicable mechanism was not based on insufficient distal nodal recovery. Instead, the gap more likely started as a result of insufficient depolarization produced by the fast wave front and persisted because this partial depolarization prevented progress of the more delayed slow pathway wave front. With sufficient weakening of the fast pathway wave front, the slow pathway impulse no longer encountered its impeding effect and was thus able to propagate through the N region and activate the His bundle. However, as with the observations in Fig 3⇑, the annihilated conduction during the gap might have resulted either from a head-on collision or from the impediment of the slow pathway wave front by the refractoriness wake of the fast pathway wave front.
Fig 5⇓, which is organized similar to Fig 3a⇑, illustrates superimposed traces obtained from a preparation in which the gap phenomenon could not be revealed. Note that in this case the presence of the two wave fronts is nevertheless obvious. It is reflected in the double humped action potential with the earlier declining component (downward arrow) associated with the fast pathway and with the later component with rising amplitude that represents the slow pathway wave front. In this experiment, however, the interaction of the two wave fronts did not result in a block. Although a variety of explanations could be proposed, all of them would share a common bottom line: The “dying” fast wave front could not prevent the anterograde progress of the slow pathway wave front. This resulted in uninterrupted switch from fast pathway to slow pathway conduction.
AV Nodal Preparations With Separated Atrial Input Sites
Whether the impediment to slow pathway conduction is due to refractoriness “tail” of the fast pathway impulse that penetrated the slow pathway domain or due to actual propagation of the fast pathway wave front retrogradely and collision with the slow pathway wave front cannot be defined with the above observations. To distinguish between these two scenarios, independent control of the atrial inputs into the slow and fast pathways was needed. Therefore, we carried out four experiments in which the atrial input sites into the AV node were separated. In this case (see Fig 1⇑, panel B), stimulation applied at the IAS produced anterograde engagement only of the fast pathway. In contrast, stimulation at the CrT engaged anterogradely only the slow pathway. This was easily verified, since anterograde conduction initiated at either input site was never associated with a prompt (within 20 to 40 ms) activation of the opposite input, as happens in the intact preparations (see the IAS-CrT intervals in Figs 3⇑ and 4⇑). In fact, activation of the nonstimulated input always occurred after a long delay (representing the propagation within the stimulated pathway through the compact N region and the nonstimulated pathway) and frequently even after the inscription of the His bundle electrogram (see below). Fig 6⇓ shows an example of data obtained from such an experiment. The conduction curve with IAS stimulation (open circles) indicated shorter conduction times and longer effective refractory period as compared with the conduction curve obtained with CrT stimulation (filled circles). This is consistent with commonly appreciated properties of the slow and fast pathways in human AV node. Note that no gap existed in either conduction curve. However, when both inputs were stimulated simultaneously, the conduction curve (triangles) indicated a gap. For S2S3>140 ms, the conduction corresponded to that seen with the fast pathway alone. No conduction was seen at S2S3<140 ms and S2S3>115 ms, although slow pathway conduction was intact when activated in isolation at those coupling intervals. Conduction resumed for S2S3<115 ms but occurred through the slow pathway.
The proposed electrophysiological mechanism is illustrated in the inset in Fig 6⇑. Although the fast pathway wave front could not reach the bundle of His for coupling intervals shorter than its effective refractory period (here 135 ms), it nevertheless was still able to invade the N region and subsequently retrogradely collide with and annihilate the otherwise potent slow pathway wave front.
The recordings shown in Fig 7⇓ further illustrate this mechanism. Each panel illustrates the responses to test beat S3 only. At longer coupling intervals, ie, S2S3=140 ms, each wave front could reach the His bundle anterogradely and reach the nonstimulated input retrogradely (panels A and B). The IAS-initiated wave front was the faster one, and this “advantage” determined its predominant role during the simultaneous stimulation (panel C) when the S3H3 remained as short as in panel A.
The situation was quite different at a coupling interval of 125 ms, that is, inside the gap. Stimulation of the IAS alone (panel D) was associated with block to His (dashed line). Notably, however, the CrT input site was activated retrogradely (arrow, circle) with a delay of 173 ms. For this to happen, the fast pathway wave front had to invade and successfully transverse the slow pathway retrogradely. Since the slow pathway wave front alone was also able to reach the N region and His bundle (panel E, S3H3=170 ms), the conditions existed for collision of the simultaneously induced fast pathway and slow pathway wave fronts (panel F, star). From the observations noted in panels D, E, and F, one can conclude that conduction block shown in panel F, when both sites were stimulated, must have occurred as a result of collision of the slow pathway and fast pathway wave fronts.
With a short S2S3 of 115 ms, stimulation of the fast pathway at the IAS input (panel G) resulted in a “bidirectional” block since neither His nor CrT electrograms were recorded. That is, the impulse not only failed to propagate to the His bundle but also failed to retrogradely transverse the slow pathway. Stimulation of the slow pathway at the CrT input site (panel H), however, produced anterograde conduction to His (S3H3=180 ms) and retrograde conduction to the IAS. When both inputs were stimulated (panel I), anterograde conduction to the His bundle was unaltered, ie, the slow pathway was predominant but retrograde conduction to the IAS was blocked, apparently because of the residual refractoriness left by S3 delivered at the IAS.
These sequences of events provide strong evidence that collision of wave fronts initiated in the AV node from the fast pathway and slow pathway input sites can form the basis of the observed gap phenomenon. Similar results were obtained in three of the four preparations with separated inputs. A gap could not be demonstrated in the fourth experiment (see “Discussion”).
The major finding of this study is the demonstration of a new mechanism to explain atrioventricular conduction gap, based on specific interaction of wave fronts in a dual input AV nodal structure. The electrophysiological mechanism underlying this phenomenon requires that the fast pathway wave front has longer effective refractory period than the slow pathway wave front, a condition that is well recognized in the human AV node. In addition, the phenomenon takes place as a result of a critical balance between the inability of the fast pathway wave front to proceed anterogradely toward the bundle of His and its ability to invade retrogradely the slow pathway domain. This retrograde invasion annihilates the anterogradely propagating slow pathway wave front through collision or interference by a refractory barrier. Further shortening of the atrial coupling interval results in more proximal block of the fast pathway wave front, eliminating its impeding ability and restoring conduction, which now occurs through the slow pathway. While experiments in the intact AV nodal preparations could not assess whether collision or interference was the mechanism, the preparations with separate IAS and CrT inputs were able to demonstrate that collision was the mechanism in those cases. It is likely, therefore, that collision could have been the mechanism of the gap phenomenon in at least some of the intact nodal preparations.
The mechanism proposed here differs from the classic explanation of AV nodal gap. The classic explanation is based on a model of the AV node that consists of serially connected proximal and distal units. The functional refractory period of the proximal unit is shorter than the effective refractory period of the distal unit. This postulate is partially based on the characteristic difference between the action potential duration in the proximal and distal nodal cells.29 Progressive shortening of the atrial coupling intervals produces a nonmonotonic change of the output intervals of the proximal unit. Thus, after its functional refractory period has been reached, further shortening of the atrial prematurity begins to produce longer proximal output intervals. These longer output intervals may restore activation of the distal unit. In such a scheme the concept of dual AV nodal pathways is not essential. However, experimental evidence demonstrating this mechanism is lacking.
There are similarities as well as important differences between the classic mechanism and the annihilation gap mechanism reported here. In both cases, initiation of the gap is triggered by anterograde block of an impulse in the distal AV node. However, in the dual pathway gap model proposed here, this condition is insufficient to explain the occurrence of AV block since the slow pathway wave front is still capable of conduction at those coupling intervals where the gap exists (see Fig 7⇑). Therefore, the added condition of the fast pathway wave front colliding or interfering with the slow pathway impulse must exist for complete AV nodal block to occur. Resolution of block at shorter coupling intervals of atrial premature beats also differs between the two models. In the classic model, resolution occurs due to proximal slowing of conduction, resulting in greater distal recovery and thus allowing conduction to resume. In our annihilation model, resolution of block involves the removal of an obstacle, the fast pathway wave front, to allow conduction of another wave front from the slow pathway to propagate through the N region of the node. This removal occurs due to decreasing penetration of the fast pathway impulse as the atrial coupling interval is shortened (see Fig 3⇑). With sufficient prematurity, the fast pathway wave is no longer able to penetrate the N region with the amplitude needed to interfere with propagation of the slow pathway wave front. Importantly, the gap does not end exclusively because of substantial conduction delay accumulated in the proximal node (see Fig 4c⇑). In fact, the refractoriness of the distal node (NH and H regions) does not appear to be involved in the development of the gap (see Fig 2⇑).
For the sake of simplicity in presentation of the experimental data, we assumed that the resumed conduction to the left of the “annihilation” gap (ie, with the shorter S2S3 coupling intervals) represented participation of the slow pathway wave front alone. Although this appeared to be the case in preparations with separated inputs (Fig 7⇑), in general it might be an oversimplification. Since the resumed conduction to the left of the gap takes place at very short atrial coupling intervals, the slow pathway wave front should itself be weakened and, although the fast pathway wavefront may not be able to cause annihilation, the two may still interact. One possible outcome of this interaction can be summation of weak wave fronts as previously reported.30 31 32 An alternative mechanism can be described as “electrotonic inhibition.” Liu and colleagues33 demonstrated in single AV nodal cells that a subthreshold response inhibits subsequent excitability secondary to partial inactivation of the transient calcium current. This inhibition could explain some characteristics of the type of curves shown in Fig 2⇑. When conduction resumed at short atrial coupling intervals on the left side of the gap, there was initial improvement of conduction when the coupling intervals were shortened. With further shortening of the coupling intervals, the conduction time again began to prolong. As the fast pathway wave front must block in order to allow the slow pathway wave front to propagate at this point, electrotonic inhibition due to increased refractoriness of the N region (resulting from partial depolarization by the “dying” fast pathway impulse) may be the cause of additional conduction delay seen with the first conducted beats on the left side of the gap. With further shortening of the coupling intervals, conduction improved somewhat due to removal of electrotonic inhibition. With this removal, further shortening of the coupling intervals only resulted in gradual decrement in the amplitude of the N cell response and subsequent conduction block. It is reasonable to assume that the three-dimensional structure of the AV node might provide conditions for a variety of space-time interactions as previously demonstrated in experimental studies.30 31 33 While our data can be explained by two-pathway model of the AV node, we do realize that other inputs into the AV node, such as left atrial inputs, may contribute to the observed phenomena. Although the data from this study cannot address all of the above issues, they nevertheless illustrate the basic mechanism underlying the phenomenon of gap in dual pathway AV nodal structure.
In 5 intact AV nodal preparations and in 1 with separated inputs, AV nodal gap was not observed. Based on data from experiments like the one shown in Fig 5⇑, we can postulate a reasonable explanation based on the demonstrated mechanism of the gap in the tissue preparations. If failure of the fast pathway wave front to conduct to the His bundle occurs in concert with its inability to conduct retrogradely into the slow pathway domain, and if the slow pathway wave front is sufficiently delayed, no gap would exist. These conditions would prevent collision of the fast pathway and slow pathway wave fronts. In addition, the delay of the slow pathway wave front arrival at the distal node will permit recovery of this region from the fast pathway wave front penetration. Alternatively, one can postulate that the decremental fast pathway wave front that would have blocked may arrive at the distal node at a similar time as the slow pathway wave front. With appropriate timing their “collision” would result in summation32 and enhanced conduction to the His bundle. In such a case, the shift from fast pathway to slow pathway conduction occurs “smoothly” and a gap cannot occur. Both of the above explanations are realistic possibilities consistent with AV nodal physiology described in this paper and previously.
The results shown in the Table⇑ indicate that shortening of the conditioning atrial interval (S1S2) tended to widen the gap zone predominantly by increasing the upper end of the gap while not significantly affecting the lower end. This implies that antegrade block of the fast pathway wave front can be readily influenced by the conditioning impulse S2, whereas retrograde interference of the slow pathway by this wave front is less vulnerable to conditioning. This difference suggests that moderate decrement in the amplitude of the fast pathway wave front may be adequate to cause failure in propagation to the bundle of His, thus causing an earlier start of the gap. However, this may still be associated with a sufficient electrotonic depolarization to interfere retrogradely with slow pathway conduction either by collision or by interference via a refractory tail.
While our data strongly support the occurrence of collision as a mechanism for conduction gap, identifying the precise site of this collision is more difficult. This is an important limitation that does not permit unequivocal determination of the type of interaction underlying the development of the gap, ie, collision-annihilation or creation of refractory wake. Future application of technologies such as optical imaging34 may demonstrate the site of actual wave front interaction by providing more detailed mapping of the dual pathways. Until then, one should consider a more general hypothesis for the discussed gap phenomenon. Namely, the anterogradely moving slow pathway wave may be stopped at a refractory zone left by the preceding fast pathway wave front in the N region rather than by the precise synchronous arrival, collision, and annihilation of both.
The results of this study do not disprove the mechanism of the classic AV nodal gap. Yet, while it is easy to conceptually understand the classic mechanism, its demonstration in live tissue may be more difficult. Certainly, conduction gaps occurring at the trough of a functional AV nodal conduction curve (HH versus AA) can be consistent with the classic mechanism. However, such gaps do not exclude annihilation as a mechanism. Our observations would suggest that the classic mechanism for conduction gap in its pure form is perhaps less common than the mechanism proposed in this report. Indeed, in the separated input model reported here, conduction gap could only be demonstrated when both inputs were stimulated. The conduction curves from stimulation of one input were always continuous. Besides the lack of experimental verification of the classic gap mechanism, there are physiological reasons why delayed conduction may not be able to reactivate the distal node in a single pathway model. While conduction delay can be readily achieved in one-pathway structure, it is also accompanied by amplitude decrement. This decrement may prevent distal activation of the NH region despite later arrival of the impulse in the N region and greater recovery of the NH region.
While the differentiation between the two mechanisms appears possible in controlled in vitro experiments (Figs 6⇑ and 7⇑) and even more so in mathematical models, it may be a difficult task in vivo and especially in patients due to the limited information about the conduction circuits. Nevertheless, the recognition of more than one electrophysiological mechanism underlying discontinuity in conduction in general and AV nodal conduction gaps in particular may have relevance to possible clinical observations. Consider cases of discontinuous conduction curves associated with AV nodal reentrant tachycardia where slow pathway ablation is used. The annihilation gap hypothesis would predict that the gap should no longer be manifest once slow or fast pathway conduction is eliminated. It should be noted that if the discontinuity is associated with a classic gap mechanism, it could remain even after successful slow or fast pathway ablation since the presence of only one pathway (provided not modified by the ablation) would be enough to produce a classic gap phenomenon.
Finally, the results of this study clearly demonstrate the necessity for careful evaluation of the mechanisms underlying the conduction in a system as complex as the AV node. It is obvious that the annihilation gap mechanism adds yet another feature, which may be manifested alone or in combination with others. For example, it has been previously reported that gaplike behavior can be observed as a result of an enhanced vagal tone.35 Although the specific effects of the vagus on the dual pathway AV nodal electrophysiology are only partially understood,36 37 38 it is reasonable to expect that the mechanism of the annihilation gap observed in this study would be modified in the presence of an enhanced autonomic tone.
The authors wish to express their deep appreciation to Dr José Jalife for sharing his observation on the reported results of these experiments, for his review and discussion of the concepts advanced in this study, and for his encouragement in the completion of this work. We also acknowledge the work done by Dr Wanzhen Zeng on computer modeling of the described phenomena that was helpful in clarifying our working hypothesis. Thanks also to Dr Igor Efimov for his helpful comments. We are also grateful to Rosemarie Capone for her expert secretarial assistance.
- Received October 24, 1994.
- Revision received April 24, 1995.
- Accepted June 8, 1995.
- Copyright © 1995 by American Heart Association
Billette J, Janse MJ, van Capelle FJL, Anderson RH, Touboul P, Durrer D. Cycle-length dependent properties of AV nodal activation in rabbit hearts. Am J Physiol. 1976;231:1129-1139.
Teague S, Collins S, Wu D, Denes P, Rosen K, Arzbaecher R. A quantitative description of normal AV nodal conduction curve in man. J Appl Physiol. 1976;40:74-78.
Meijler FL, Janse MJ. Morphology and electrophysiology of the mammalian atrioventricular node. Physiological Reviews. 1988;68:608-647.
Schuilenburg RM, Durrer D. Atrial echo beats in the human heart elicited by induced atrial premature beats. Circulation. 1968;37:680-693.
Denes P, Wu D, Dhingra RC, Chuquimia R, Rosen KM. Demonstration of dual A-V nodal pathways in patients with paroxysmal supraventricular tachycardia. Circulation. 1973;48:549-555.
Denes P, Wu D, Dhingra RC, Amatylion F, Wyndham C, Rosen KM. Dual AV nodal pathways: a common electrophysiological response. Br Heart J. 1975;37:1069-1076.
Rosen KM, Denes P, Wu D, Dhingra RC. Electrophysiological diagnosis and manifestation of dual A-V nodal pathways. In: Wellens HJJ, Lie KI, Janse MJ, eds. The Conduction System of the Heart. Leiden, Netherlands: HE Stenfert Kroese BV; 1976:453-466.
Sung RJ, Waxman HL, Saksena S, Zulfikar G. Sequence of retrograde atrial activation in patients with dual atrioventricular nodal pathways. Circulation. 1981;64:1059-1067.
Kay GN, Epstein AE, Dailey SM, Plumb VJ. Selective radiofrequency ablation of the slow pathway for the treatment of atrioventricular nodal reentrant tachycardia: evidence for involvement of perinodal myocardium within the reentrant circuit. Circulation. 1992;85:1675-1688.
Mendez C, Moe GK. Demonstration of a dual A-V nodal conduction system in the isolated rabbit heart. Circ Res. 1966;19:378-393.
Moe GK, Mendez C, Han J. Aberrant A-V impulse propagation in the dog heart: a study of functional bundle branch block. Circ Res. 1965;16:261-286.
Wu D, Denes P, Dhingra R, Rosen K. Nature of gap phenomenon in man. Circ Res. 1974;34:682-692.
Damato AN, Akhtar M, Ruskin J, Caracta A, Lau SH. Gap phenomenon: antegrade and retrograde. In: Wellens HJJ, Lie KI, Janse MJ, eds. The Conduction System of the Heart. Leiden, Netherlands: HE Stenfert Kroese BV; 1976:504-528.
Wit AL, Damato AN, Weis MB, Steiner C. Phenomenon of the gap in atrioventricular conduction in the human heart. Circ Res. 1970;27:679-689.
Tchou P, Mazgalev T, Keim S, Mowrey K, Cheng YN, Mehdirad A. A new mechanism for atrioventricular nodal gap phenomenon: interference between wavefronts from dual atrioventricular nodal pathways. Circulation. 1993;88(suppl I):I-203. Abstract.
de Bakker JMT, McGuire MA, Vermeulen JT, Janse MJ, van Nemel NM. Origin of slow potentials in the atrioventricular junctional area. Circulation. 1993;88(suppl I):I-202. Abstract.
Irisawa H, Nakayama T, Noma A. Membrane currents of single pacemaker cells from rabbit S-A and A-V nodes. In: Noble D, Powell T, eds. Electrophysiology of Single Cardiac Cells. London, England: Academic Press; 1987:167-186.
Hoshino K, Anumonwo J, Delmar M, Jalife J. Wenckebach periodicity in single atrioventricular nodal cells from the rabbit heart. Circulation. 1990;82:2201-2216.
Irisawa H, Noma A. Pacemaker mechanism of rabbit sinoatrial node cells. In: Bouman LN, Jongsma HJ, eds. Cardiac Rate and Rhythm. The Hague, Netherlands: Nijhoff; 1982:35-61.
Irisawa H, Brown HF, Giles W. Cardiac pacemaking in the sinoatrial node. Physiol Rev. 1993;73:197-227.
Noble D, Noble SJ. A model of sino-atrial node activity based on a modification of the DiFrancesco-Noble (1984) equations. Proc R Soc Lond. 1984;B222:295-304.
Billette J. Atrioventricular nodal activation during premature stimulation of the atrium. Am J Physiol. 1987;252:H163-H177.
Zipes DP, Mendez C, Moe GK. Evidence for summation and voltage dependency in rabbit atrioventricular nodal fibers. Circ Res. 1973;32:170-177.
Janse MJ. Influence of the direction of the atrial wavefront on A-V nodal transmission in isolated heart of rabbits. Circ Res. 1969;25:439-449.
Mazgalev T, Dreifus LS, Iinuma H, Michelson EL. Effects of the site and timing of atrioventricular nodal input on atrioventricular conduction in the isolated perfused rabbit heart. Circulation. 1984;70:748-759.
Liu Y, Zeng W, Delmar M, Jalife J. Ionic mechanism of electrotonic inhibition and concealed conduction in rabbit atrioventricular nodal myocytes. Circulation. 1993;88:1634-1646.
Mazgalev T, Dreifus LS, Michelson EL. A new mechanism for atrioventricular nodal gap: vagal modulation of conduction. Circulation. 1989;79:417-430.
Keim S, Mazgalev T, Tchou P. Physiologic effects of subthreshold burst stimulation on the human AV node. PACE Pacing Clin Electrophysiol. 1993;16:158. Abstract.