Ventricular Echoes EVIDENCE FOR DISSOCIATION OF CONDUCTION AND REENTRY WITHIN THE AV NODE. By Robert J. Mignone and Andrew G. Wallace, M.D.

Transcription

1 Ventricular Echoes EVIDENCE FOR DISSOCIATION OF CONDUCTION AND REENTRY WITHIN THE AV NODE By Robert J. Mignone and Andrew G. Wallace, M.D. ABSTRACT These experiments were designed to examine further the concept of a dual pathway for AV transmission in the heart, and to determine whether or not the atrium is a necessary link in the reentry pathway used for a ventricular echo. Studies were performed on anesthetized and awake dogs with recording electrodes implanted on the atrium, bundle of His, and right ventricle. When paired stimuli (VJVJ) were applied to the ventricle with an appropriate delay, the second stimulated beat was followed by a nonstimulated ventricular echo. The echo response was of supraventricular origin. During paired stimulation with echoes, stimulation of the atrium prior to arrival of the retrograde response to V 2 produced a propagated ventricular beat in advance of the expected echo. Stimulation of the atrium immediately after an echo was associated with apparent supernormal AV conduction. These observations support the conclusion that the AV node can be dissociated functionally into at least two pathways. It was possible to render the atrium refractory without abolishing the echo. This finding indicates that the atrium per se does not participate in the reentry path used for ventricular echo. ADDITIONAL KEY WORDS paired stimulation reciprocal rhythm extrasystole dog arrhythmias Echo responses are an unusual type of coupled rhythm in which an impulse initiated in one chamber of the heart propagates to another and then returns to its chamber of origin. Despite their infrequent spontaneous occurrence, echoes have been the subject of considerable experimental study because they have important implications in our understanding of the normal function of the AV node. Moe, Preston, and Burlington 1 first described a technique for producing ventricular echoes. When the ventricle was paced at a regular frequency and a premature ventricular stimulus was applied at a short interval after the basic response, nonstimulated beats of From the Department of Medicine and the Department of Surgery, Division of Thoracic Surgery, Duke University Medical Center, Durham, North Carolina. This work was supported in part by a grant from the Walker P. Inman Fund, in part by grant HE A1, from the U. S. Public Health Service, and in part by research grant HE Accepted for publication April 25, supraventricular origin coupled to the premature stimulus were observed. The premature impulse was presumed to encounter one group of cells in the AV node which were still refractory from the preceding beat and another group which had recovered. The impulse utilized the latter path to propagate slowly to the atrium, and then entered that portion of the node which was not used for retrograde transmission and returned to the ventricle. According to this hypothesis a dual pathway was created within the AV node when the premature impulse arrived early enough to dissociate groups of cells with differing refractory periods. The dual pathway then provided the necessary conditions for a reentry circuit. Several other authors 2 " 4 have noted that the premature impulse must be conducted to the atrium for a ventricular echo to occur, and they have concluded that the atrium is an essential link in the circuit used for an echo. 8 ' 4 Recently, however, Wallace and Daggett 5 showed that atrial echoes can occur in the 638 OrcmUtitm Rtsurcb, Vol. XIX, Stpumbtr 1966

2 VENTRICULAR ECHOES 639 presence of complete AV block. Since it is apparent that the ventricle is, therefore, not essential to the occurrence of an atrial echo, we believed it would be of interest to examine further the role of the atrium in ventricular echoes. These studies were designed in an attempt to clarify the role of the atrium in ventricular echo responses and to provide additional data concerning the dual pathway hypothesis for AV nodal conduction. Methods Experiments were performed on 29 dogs. Anesthesia was induced with thiamylal, 10 to 20 mg per kg, and ventilation was maintained with a Harvard respirator. A right thoracotomy was performed, and bipolar surface recording electrodes were implanted on the epicardial surface of the right atrium and right ventricle. A similar electrode was implanted over the bundle of His during temporary inflow occlusion. Pacing electrodes were sutured to the right atrium and right ventricle, and the sinoatrial node was crushed to slow the intrinsic heart rate. Three dogs, prepared in the above manner, were allowed to recover and were studied while awake and unmedicated several weeks later. The pacing electrodes were connected through isolation transformers to a series of Tektronix pulse generators. The first two generators were programmed to produce continuous "paired pulse stimulation" 6 by pacing either the atrium or the ventricle at a basic cycle length (A^j) or (V,V,) and introducing a second stimulus (A 2 ) or (V;,) at any desired interval after each basic response (i.e. A]A 2, A^o, AjA^, etc. or VjV.>, V,^.,, VjVo, etc.) A third generator supplied a test pulse which could be delivered to either atrium (A s ) or ventricle (V 8 ) at any desired interval after each sixth basic response. All impulses were 5 msec in duration and approximately twice threshold. Signals from each recording electrode were amplified with Tektronix 122 preamplifiers and displayed on a four-channel Tektronix oscilloscope. Frequencies below 80 cycle per sec and above 1000 cycle per sec were filtered to sharpen the desired signals. A Lead II ECG also was recorded. Responses were recorded on a Consolidated oscillograph at paper speeds of 4 or 8 inches per sec. To define accurately the sequence of atrial excitation in the region near the AV node, 3 dogs were placed on complete cardiopulmonary by-pass and maintained at 38 C. The right atrium was opened and a bipolar probe electrode was used to record activity at different locations from Circulation Rtsircb, Vol. XIX, Stpttmktr 1966 the endocardial surface of the atrial floor and septum. The sequence of atrial excitation during both antegrade and retrograde propagation was determined by measuring the interval between the pacing stimulus and local activity at each point. Results 1. THE RESPONSE OF THE AV NODE TO PREMATURE EXCITATION Paired stimulation of the ventricle at varying ViV 2 intervals produced a pattern of retrograde conduction as shown in Figure 1 (panel A). Retrograde conduction across the AV node was estimated from the interval between ventricular responses (Vi or V 2 ) and the corresponding activity recorded from the atrial septum (ASi or AS 2 ). At long V]V 2 intervals, V 2 AS 2 equaled V1AS1. As the ViV 2 interval was decreased, V 2 AS 2 remained essentially constant until a critical interval was reached, after which further reduction of ViV 2 resulted in a progressive lengthening of retrograde conduction time. At very short V X V 2 intervals retrograde conduction became fixed at a prolonged interval. A similar pattern was observed for antegrade conduction during paired stimulation of the atrium (Fig. 1, panel B). In most experiments the effective refractory period of the ventricle limited further shortening of the ViV 2 interval and was encountered before complete retrograde block at the AV node. During paired stimulation of the atrium, however, block of A 2 at the level of the node always limited further reduction of the AiA 2 interval. 2. COUPLE VENTRICULAR PACE WITH ECHOES Echo responses were observed in all dogs during paired stimulation of the ventricle. An example of bipolar electrograms recorded from 1 dog is shown in Figure 2. The electrograms from top to bottom were recorded from right atrium (RA), bundle of His (His), right ventricle (RV) and Lead II of the ECG. The His electrode recorded activity from the subjacent atrial septum (AS), bundle of His (H) and underlying ventricular septum. Panel A shows a control recording of a spontaneous atrial beat. In panel B, paired stimuli were applied to the right ventricle at a ViV a inter-

3 640 MIGNONE, WALLACE "loo 140 ISO FIGURE 1 Panel A shows retrograde conduction times across AV node; panel B shows antegrade conduction times across AV node. In panel A, V,V t = the interval between paired stimuli applied to the ventricle; VjAS, = retrograde conduction time following V x ; VjAS = retrograde conduction time following V,. In panel B, A t A a = the interval between paired stimuli applied to the atrium; ASj^H, = the antegrade conduction time between atrial septum and His bundle following A t ; AS 2 Hj = the antegrade conduction time between atrial septum and His bundle following A 2. Basic cycle length (V t V t ) in panel A = 1000 msec. Basic cycle length (A,A t ) in panel B = 1000 msec. RA- His FIGURE 2 He I r- 200 msec A ventricular eclw. Panel A shows a s^iontaneoiis alrial beat; panel B shows paired stimuli to ventricle followed by an echo. Leads are recorded from the right atrium (RA), His bundle (His), right ventricle (RV) and a Lead II ECG (L-II). A = atrium, AS = atrial septum, H = His bundle, V = ventricle. The arrows show the stimulus artifact. See text for further description. val of 258 msec. V x propagated retrograde to the atrial septum with a V^St interval of 112 msec. V 2 also propagated to the atrium, but retrograde conduction time across the AV node was prolonged (V2AS2 = 162 msec). A nonstimulated echo beat (V E ) occurred 262 msec after V L>. The activation sequence during the echo proceeded from His bundle CircuUlwn KtstMrcb, Vol. XIX, Stpitmbtr 1966

4 VENTRICULAR ECHOES 641 (H E ) to ventricle (V E ) and the QRS complex was identical to that of a normal supraventricular beat. Ventricular echoes were observed when the basic cycle length was sufficiently long to allow inclusion of both the premature beat and the echo response, and when the VjVo interval was sufficiently short to produce long V2AS2 intervals. Table I summarizes measurements from 13 experiments in which the V1V0 interval was altered in a stepwise manner, and the effects on retrograde conduction were observed. In all animals V1AS1 varied by less than 5 msec regardless of the ViV«interval. In any given dog there was a range of ViV L. intervals at which echoes were observed. This range for the group of animals was from 20 to 140 msec. The interval between V L. and the echo remained constant in 9 dogs and increased by 20 to 60 msec in 4 dogs as the VtVo interval was decreased within the echo range. Regardless of the interval between Vu and the echo (V E ), V 2 AS 2 increased, and AS L.V E decreased as the interval between paired stimuli was decreased. 3. THE ROLE OF THE ATRIUM IN VENTRICULAR ECHOES When paired stimuli were applied to the ventricle at appropriate ViV 2 intervals the V 2 echo interval was constant in 9 dogs. In these animals, a VjVo interval was selected at which the echo recurred at a stable time TABLE 1 Interval Measurements when VjV s was Adjusted within the Echo Range Eipt. No. V,Vt msec A. Observations from dogs in B. Observations from dogs in V.E mtec which V t E was constant as which V,E varied as V / V J after V 2 and then a third stimulus (A3) was applied to the atrium starting at ViA 8 intervals which equaled ViVj. The V1A3 interval was then adjusted by approximately 10 msec increments to approach the expected AS 2. At short ViA 8 intervals, As failed to conduct to the ventricle and blocked both retrograde conduction V 2 and the echo. One such experiment is illustrated in Figure 3. Panel A shows records obtained prior to introducing an A 3. In panel B, the atrial septum was pre-excited 100 msec before the predicted AS 2. The atrial response to A 3 failed to conduct to the ventricle, but blocked AS 2 and prevented the echo. At intermediate V1A3 intervals, pre-excitation of the atrial septum blocked AS 2 and resulted in a propagated response which reached the His bundle and ventricle before the expected echo. An experiment which illustrates this phenomenon is presented in Figure 4. Panel A shows records obtained just prior to introducing A s. In panel B, the atrial septum was pre-excited 30 msec before the expected AS 2. The atrial response to A 3 propagated to the His bundle 18 msec ahead of the predicted echo. When the V1A3 interval was adjusted so that the atrial septum was pre-excited just before the expected arrival of AS 2, the atrial response to A 3 blocked AS 2, but failed to con- VJASJ mtec VJVJ was decreased was decreased ASiE m»ec CircuUiion Rti rcb, Vol. XIX, Sfpfmbtr 1966

5 642 MIGNONE, WALLACE FIGURE 3 200m»tc Block of AS S and echo hy early pre-excitation of atrium. Panel A, V { V s followed hy an echo. Panel B, V 1 V 1 followed hy pre-excitation of atrium (arrow on RA lead). See text for details. RAi A Z FIGURE 4 i' I ' AS, ' AS 3H 3 Antegrade propagation ahead of the expected echo, in response to an appropriately timed stimulus applied to atrium. Panel A, V t V t with an echo. Panel B, V,V a followed hy pre-excitation of atrium (arrow on RA lead) which propagates to atrial septum (AS S ), His (H } ) and ventricle (V H ). See text for further details. duct to the ventricle and did not prevent the echo. One such experiment in which the relative positions of V 2, AS 2 and H E remained constant (±2 msec) is shown in Figure 5. In panel A, control recordings of paired stimuli with an echo are shown. In panel B, the atrial septum was pre-excited 22 msec before the expected AS 2. The atrial response to A 3 rend- CircmUtion Rtstrcb, Vol. XIX, Stpltmbtr 1966

6 VENTRICULAR ECHOES 643 RAi T r 200m*»c FIGURE 5 Persistence of an echo despite rendering the atrium Panel B, V,V2 with an echo. Note that in panel B as in panel A, despite pre-excitation (arrow) of the expected ASt. The difference in contour between ASt See text for details. ered the atrium refractory at the time AS^ would have occurred, but did not prevent the echo. That the ventricular response following Vu was an echo is suggested by the fact that the VjHE interval was identical to that in panel A. Further evidence that the ventricular response was indeed an echo is that when the atrial septum was pre-excited 4 msec earlier (i.e. 26 msec in advance of the expected AS;;), a ventricular response occurred 15 msec ahead of the expected echo. These observations were confirmed on repeated trials, and in all animals tested the atrial septum could be pre-excited by 10 to 30 msec without abolishing the echo. To define the temporal relation between atrial complexes recorded by the His electrode and other atrial activity in the region of the AV node, a bipolar probe electrode was used to measure the sequence of excitation over the endocardial surface of right atrium. The interval between the pacing stimulus and local excitation was measured at eleven different points in close proximity to the AV node. The earliest point was arbitrarily established as zero. All other points were expressed as Orc$iUlK)n Rtlcorch, Vol. XIX, Scplembir I960 refractory. Panel A, V1Vl with an echo. the V2 echo interval is exactly the same atrium (AS,) 22 msec in advance of the and AS, is also a result of pre-excttation. deviations from zero. One map is shown in Figure 6, and comparable data were obtained from 2 other dogs. The map shows that during ventricular pacing with retrograde transmission (panel A), earliest atrial activity was recorded near the coronary sinus and approximately 1 cm above the AV ring. Activity reached atrial tissue at the region of the His electrode 3 to 12 msec later. During atrial pacing with antegrade conduction (panel B), the region of the atrium under the His electrode was excited coincident with or no more than 2 msec after activation of the area which was observed to be the earliest during retrograde transmission. Thus, for an atrial stimulus to have rendered refractory all atrial muscle, including that immediately adjacent to the AV node, it would have been necessary to pre-excite the atrial complex recorded from the His electrode during retrograde transmission by a minimum of 3 msec and a maximum of 14 msec. The region of the atrial septum subjacent to the His electrode was preexcited by 10 to 30 msec in 9 dogs without preventing the echo.

7 644 MIGNONE, WALLACE B COR. SINUS COR. SINUS 5 -M 8 5 J2-" S leaflet FIGURE 6 A map of the sequence of atrial excitation in the region of the AV node, during retrograde transmission (panel A) and antegrade transmission (panel B). In panel A, the box enclosed by the dotted line denotes the point of earliest activity during retrograde transmission. Other points are plotted as deviations from the earliest point in milliseconds. In panel B, O denotes the pacing site and each point is plotted as a RA L-n ATRIOVENTRICULAR CONDUCTION FOLLOWING THE ECHO In 6 of the 9 animals in which the V 2 echo interval remained constant, a third stimulus (A s ) was applied to the atrium after AS2. The earliest stimulus which would elicit an atrial response followed AS 2 by 130 to 180 msec. Even the earliest response was observed to propagate to the ventricle with a rapid conduction time equal to AV conduction during a slow spontaneous atrial rhythm. One such experiment is presented in Figure 7. The AS 8 H 3 interval was 50 msec, which was the same as that observed during a spontaneous atrial rhythm and was 45 msec shorter than that observed when paired stimuli were applied to the atrium at an AjA L > interval equal to AS L >A.i. deviation in milliseconds from the pacing site. The enclosed box refers to the point of earliest activity noted in panel A. The standard position of the implanted His electrode is shown. S. leaflet refers to the septal leaflet of the tricuspid valve. See text for further details. 200m»c FIGURE 7 Rapid antegrade propagation immediately after an echo. Tracing shows V } V t followed by an echo (V E ). A stimulus (arrow) was applied to the atrium as soon as the effective refractory period after A, was ended. The response to that stimulus (A s ) propagated antegrade across the AV node (ASjH, interval) with a conduction time 45 msec faster than that observed when the atrium was paced at an A t A t internal equal to the ASfA, interval shown in the figure. See text for details. Circulation Ruarcb, Vol. XIX, Stpumhtr 1966

8 VENTRICULAR ECHOES 645 STUDIES OF THE ECHO PHENOMENON IN AWAKE DOGS Three animals with previously implanted recording electrodes were examined when awake and unmedicated. Figure 8 shows electrograms recorded from 1 of the dogs. The nomenclature is identical to that described previously except that an electrode which recorded local activity from the right bundle branch had been implanted on the right Purkinje-papillary muscle junction (PPJ). Panel A shows electrograms recorded during atrial pacing with a propagation sequence which proceeded from atrial septum (AS) to His bundle (H) and to right Purkinje tissue (P). Panel B illustrates paired ventricular pace with echoes. The activation sequence for the echo was identical to that observed during control supraventricular paced beats. Discussion Recent microelectrode studies by Hoffman and Cranefield7 and Paes de Carvalho and De Almeida8 indicate that the AV node consists of three functionally separable parts; an upper AN region, the middle N region, and a lower NH zone. The AN and NH layers conduct more rapidly than the middle N zone where most of the normal nodal delay occurs. De Almeida9 has presented additional histological evidence of three structural zones within the node which correspond to those described functionally by Hoffman and Paes de Carvalho. The AN and NH zones are composed of more parallel fibers, while the N zone is a more highly branched intercommunicating network in which fibers are separated by larger amounts of connective tissue. In view of the above considerations, it seems likely that conduction would be very complex in the N zone and might involve sequential firing of series elements as well as multiple summation and cancellation. Failure of conduction in one group of cells might, therefore, establish a large distal area which would be inaccessible to the perpendicular spread of electrical activity from neighboring cells. It seems reasonable to postulate that if pathways within the AV node were to become dissociated, it would be most likely in the N zone, where there is a wide divergence of action potential duration and where the safety factor for conduction is lowest.8'1(ti n Wata- L-E 200 msec FIGURE 8 A ventricular echo in an awake unmedicated dog. His = His electrogram, PPJ = electrogram recorded from the right Purkinje-papillary junction (P = Purkinje spike), L-ll = Lead II of ECG. Panel A, a spontaneous atrial beat. Panel B, VtV2 with an echo. See text for details. CircmUliom Resmrch, Vol. XIX, Scplemitr 1966

9 646 MIGNONE, WALLACE nabe et a.]. 1 ' 2 have recently demonstrated that propagation may fail in one part of the node while another continues to conduct impulses successfully. Their studies provide convincing in vitro evidence that the AV node can be dissociated into conducting and nonconducting paths, and that these paths have spatial separation. In their experiments reentry within the node was observed repeatedly. The data presented in this report, as well as recent observations by Mendez et al., 4 demonstrate that during paired pacing of the ventricle with echo beats, it is possible to stimulate the atrium over a considerable interval prior to the expected atrial response to V 2, and observe propagation to the His bundle and ventricle which occurs before the expected echo. From these observations it is apparent that a pathway through the node is available for AV conduction at the same time V 2 is slowly traversing another part of the node in a retrograde direction. Events such as those illustrated in Figure 5 deserve further comment. Pre-excitation of the atrial septum by 22 msec failed to abolish the echo while pre-excitation by 26 msec resulted in a propagated response at the His bundle 15 msec before the expected His response of the echo. As noted previously, we could not ascribe these findings to an unstable preparation, since the relative positions of V 2, AS 2 and H E remained constant throughout the period of observation. How then, can we account for the fact that a stimulus which led to atrial pre-excitation by 22 msec failed to reach the His bundle, while pre-excitation 4 msec earlier not only reached the His bundle, but reached it 15 msec in advance of the echo. A possible explanation can be offered, which is based on a recent analysis of potential mechanisms involved in "supernormal" conduction. 18 As will be discussed subsequently, we believe that there is evidence to support the view that on its return to the ventricles, the echo utilizes certain elements in the lower portion of the node which were used previously by V L. during antegrade propagation. At relatively long ViV 2 intervals, such as those in the experiment shown in Figure 5, these elements are thought to be in a state of relative refractoriness during propagation of the echo. 4 If we assume that the impulse consequent to pre-excitation by 26 msec encountered partially refractory tissue in the upper portion of the node, then conduction could have been so slowed that it reached the lower portion of the node at the end of its refractory period. The impulse could then propagate through the lower node at a substantially faster rate than the echo and reach the His bundle early, as a result of apparent supernormal conduction. Although no proof can be offered to support the validity of this explanation, it is, nevertheless, a reasonable explanation based on current information concerning the functional properties of the AV node. Further support was obtained for dissociated pathways within the node from studies designed to examine the characteristics of AV conduction immediately after an echo. The experiment shown in Figure 7 demonstrates that when the atrium was stimulated as soon as its refractory period following AS 2 had ended, the atrial response propagated rapidly to the ventricle. Propagation of A 8 after the echo was substantially faster than that observed when paired stimuli were applied to the atrium at an AiA 2 interval which equalled the AS 2 A S interval. The apparent supernormal conduction of A s can be accounted for if it is assumed that those elements in the upper node which were used by V 2 during retrograde conduction were then not used by the echo during antegrade transmission. The pathway utilized by V 2 would not only be available for propagation of A 8, but would have had sufficient time to recover fully and thus permit A s to propagate through the node with maximal speed. The occurrence of echoes, the conditions under which they were produced, and the characteristics of AV transmission before and after the echo provide in vivo evidence that premature excitation can dissociate at least a part of the node into functionally separate pathways. In each dog described in this report there was a range of ViV 2 intervals in which V 2 CircmUsion Ruircb, Vol. XIX, Stpumit 1966

10 VENTRICULAR ECHOES 647 was followed by a nonstimulated ventricular echo. The QRS complex of an echo was essentially identical to that of a normal supraventricular beat. Further evidence that the echo was of supraventricular origin was obtained from animals in which electrodes had been implanted on the His bundle and right bundle branch. The echo responses were conducted over the His-Purkinje system in a manner identical to that of supraventricular beats. During paired stimulation of the ventricle with echo responses, VoASo increased and ASo echo decreased as the V1V0 interval was decreased. This observation has been made previously by Mendez et al. 4 and suggests that a part of the return path used by the echo was activated by retrograde propagation during V 2. Mendez and his associates presented convincing evidence that Vo responses which were too early to propagate successfully to the atrium, did, nevertheless, enter the AV node. These responses were referred to as "concealed." It was suggested that the lower node is activated by both Vo and the echo in succession, and that the refractory period after Vo of those elements used for both antegrade and retrograde conduction accounted for the variation of the ASo echo interval. From the above considerations it seems reasonable to postulate that the retrograde impulse consequent to Vo penetrates the lower portion of the AV node. On entering the N region the impulse encounters an electrically nonhomogeneous population of cells differing in the degree to which they are excitable. Those elements of the node which are excitable provided a pathway for transmission to the atrium, but elements in parallel which have not fully recovered are left unexcited. Having traversed a portion of the node in a retrograde direction, the impulse then reaches a connecting bridge and returns antegrade to the ventricle. On its return to the ventricle it first utilizes those elements within the N region which were not excited by V 2 and then arrives at a common path in the lower node which is used by both V 2 and the echo. The question remains, is the connecting link Circulation Rn rch, Vol. XIX, Stpitmttr 1966 within the AV node, or does the bridging circuit involve the atrium. We never observed a ventricular echo during paired pacing of the ventricle unless V L > propagated to the atrium. This observation is consistent with the view that the atrium is essential to the occurrence of a ventricular echo. The possibility exists, however, that under the conditions of our experiments, any Vo which was not extinguished within the N zone would traverse successfully the AN zone and reach the atrium. Moreover, if the connecting bridge for an echo were in the AN zone, then atrial excitation would necessarily precede the echo, but would not participate in the reentry circuit. It was reasoned that if the atrium could be rendered refractory before Vo had traversed the node, and if the echo persisted, this would localize the connecting bridge to within the AV node. Mendez et al. 4 recently published a series of experiments in which pre-excitation of "atrial tissue" by even a few milliseconds resulted in antegrade conduction which blocked the echo. On the basis of this observation the authors concluded that the atrium was a necessary link in the circuit for a ventricular echo. It should be noted, however, that their recording electrodes consisted of needles impaled at the "atrionodal junction." If their electrode was indeed placed in atrial tissue then the results are contrary to those described in this report. If, on the other hand, the needle tips were inadvertently inserted into the superficial aspect of the underlying AV node, then the difference in results is explainable since excitation of the node would be expected to block the echo. They noted that during an echo the interval between their "atrial" recording and the His bundle was identical to that for a stimulated supraventricular beat initiated a few milliseconds in advance of the echo. This observation alone suggests that their electrode was located either precisely at the reentry point or within the AV node. In our experiments the atrial complex recorded from the surface electrode over the His bundle could be pre-excited by 10 to

11 648 MIGNONE, WALLACE 30 msec without preventing the echo. Although the His electrode was in close proximity to the AV node we could not be certain that the atrial activity recorded by this electrode necessarily represented the earliest activity during retrograde transmission. It was determined in three successful mapping experiments that the atrial complex recorded by the His electrode was a minimum of 3 and a maximum of 12 msec later than the earliest activity in atrial muscle during retrograde conduction. It also was established that when a stimulus was delivered to the atrial appendage, activity reached the atrium subjacent to the His electrode coincident with, or no more than 2 msec after it reached the area of earliest retrograde atrial activity. We have concluded from these data that it was possible to render the entire atrium refractory, including that portion immediately adjacent to the AV node, without abolishing the echo phenomenon. If this conclusion is correct, then it follows that the atrium is not an essential link in the reentry circuit used for a ventricular echo. Recent evidence indicates that there are specialized fibers within the atrium which connect the SA node and AV node. These tracts have been identified histologically 14 and they have been shown to conduct the impulse with greater speed than ordinary atrial muscle.* Furthermore, it has been demonstrated that these tracts maintain a functional continuity between the SA and AV node in the presence of serum potassium levels which are sufficient to render ordinary atrial muscle inexcitable. ir> Moe and Mendez 16 have recently shown that ventricular echoes persist when atrial activity is abolished by high potassium levels. These experiments by Moe and Mendez support the conclusion that the atrium per se is not a necessary link in the echo pathway. Although this conclusion seems fully justified, neither the experiments by Moe nor those described in this report exclude the possibility that a specialized group of fibers adjacent to the AV node, but within the 'Unpublished observations, J. W. Holsinger and A. G. Wallace. atrium, participate in the echo circuit. It would seem that this possibility could be examined only by the use of multiple microelectrode insertions in the immediate vicinity of the AV node. The data reported above do not support the view that the atrium per se is an essential link in the circuit used for a ventricular echo. Coupled with the previous observation that atrial echoes can occur without participation of the ventricle, these experiments suggest that echo phenomena can be accounted for by reentry within the region of the AV node. The concept that nonuniform recovery enables an early impulse to dissociate certain elements of the node and thus functionally produce a dual pathway for AV transmission seems justified. References 1. MOE, G. K., PRESTON, J. B., AND BURLINGTON, H.: Physiologic evidence for a dual A-V transmission system. Circulation Res. 4: 357, ROSENBLUTH, A.: Ventricular echoes. Am. J. Physiol. 195: 53, ROSENBLUTH, A.: Two processes for atrioventricular propagation of impulses in the heart. Am. I. Physiol. 194: 495, MENDEZ, C., HAN, J., GARCIA DE JALON, P. D., AND MOE, G. K.: Some characteristics of ventricular echoes. Circulation Res. 16: 562, WALLACE, A. C, AND DACGETT, W. M.: Reexcitation of the atrium. "The echo phenomenon". Am. Heart J. 68: 661, CRANEFIELD, P. F.: Paired pulse stimulation and post-extrasystolic potentiation in the heart. Progr. Cardiovascular Diseases 8: 446, HOFFMAN, B. F., AND CRANEFIELD, P. F.: The Electrophysiology of the Heart. New York, McGraw-Hill, PAES DE CARVALHO, A., AND DE ALMEIDA, D. F.: Spread of activity through the atrioventricular node. Circulation Res. 8: 801, DE ALMEIDA, D. F.: Histologic Aspects of the Atrio-ventricular Node of the Rabbit Heart. Proc. Symp. Specialized Tissues of the Heart, Rio De Janeiro, 1960, p HOFFMAN, B. F., PAES DE CAHVALHO, A., CARLOS DE MELLO, W., ANT) CRANEFIELD, P. F.: Electrical activity of single fibers of the atrioventricular node. Circulation Res. 7: 11, HOFFMAN, B. F.: Physiology of atrioventricular transmission. Circulation 24: 506, WATANABE, Y., DREIFUS, L. S., AND SCOTT, J. C: Inhomogeneous conduction in the AV node. CircuUlicm Rtstrch, Vol XIX, Septtmber 1966

12 VENTRICULAR ECHOES 649 A model for re-entry. Federation Proc. 24: right and left atrium in the human heart. Am. 212, No. 2, part 1, Heart J. 68: 498, MOE, C. K., MENDEZ, C, AND ABILDSKOV, J. A.: 15. VASSALLE, M., AND HOFFMAN, B. F.: Spread of A complex manifestation of concealed AV con- sinus activation during potassium administraduction in the dog heart. Circulation Res. tion. Circulation Res. 17: 285, : 51, MOE, G. K., A.XD MENDEZ, C: Physiologic basis 14. JAAEES, T. N.: Connecting pathways between the of reciprocal rhythm. Progr. Cardiovascular sinus node and A-V node and between the Diseases 8: 461, CircuUtiof Rtimrcb, Vol. XIX, Stpicmbir 1966