Direct Demonstration of Sinus Node Reentry in the Rabbit Heart

Transcription

1 SINUS NODE REENTRY/AUessie and Bonke 557 Reuter H, Scholz H: A study of the ion selectivity and the kinetic properties of the calcium-dependent slow inward current in mammalian cardiac muscle. J Physiol (Lond) 264: 17-47, 1977 Singer DH, James TN, Harris P, Malm JR, Hoffman BF: Electrophysiology of chronic atrial fibrillation (abstr). Circulation 36: 239, 1967a Singer DH, Lazzara R, Hoffman BF: Interrelationships between automaticity and conduction in Purkinje fibers. Circ Res 21: , l%7b Singer DH, Ten Eick RE: Aberrancy: Electrophysiological aspects. Am J Cardiol 28: 381^401, 1971 Singer DH, Ten Eick RE, DeBoer A: Electrophysiological basis of human atrial tachyarrhythmia. In Cardiac Arrhythmias: The Hannemann Symposium, edited by L Dreifus, W Likoff. New York, Grune & Stratton, 1973, pp Singer DH, Ten Eick RE, Elson J, Bonnar, J: Electrophysiologic properties of diseased human atrium (abstr). Fed Proc 36: 415, 1977 Sleator W Jr, DeGubareff T: Transmembrane action potentials and contractions of human atrial muscle. Am J Physiol 206: , 1964 Ten Eick RE, Singer DH: Human cardiac arrhythmia: Mechanisms and models. In Cardiac Arrhythmias, edited by J Han. Springfield, Illinois, CC Thomas, 1972, pp 3-37 Ten Eick R, Nawrath H, McDonald TF, Trautwein W: On themechanism of the negative inotropic effect of acetylcholine. Pfluegers Arch 361: , 1976 Trautwein W, Kassebaum DG, Nelson RM, Hecht HH: Electrophysiological studies of human heart muscle. Circ Res 10: , 1962 Trautwein W: Membrane currents in cardiac muscle fibers. Physiol Rev 63: , 1973 Trautwein W, McDonald TF, Tripathi, O: Calcium conductance and tension in mammalian ventricular muscle. Pfluegers Arch 354: 55-74, 1975 Wit AL, Hoffman BF, Cranefield PF: Slow conduction and reentry in the ventricular conduction system: I. Return extrasystole in canine Purkinje fibers. Circ Res 30: 1-10, 1972 Direct Demonstration of Sinus Node Reentry in the Rabbit Heart MAURITS A. ALLESSIE AND FELIX I.M. BONKE SUMMARY In spontaneously beating isolated right atria of the rabbit, ectopic premature beats of varying prematurity were elicited. In some preparations, very early premature beats were followed by abnormal extra responses which were suspected to be sinus echo beats. By simultaneous recording of multiple atrial electrograms and consecutive impalement of as many sinus node fibers as possible during repeated induction of sinus echoes, we tried to follow the pathway of the impulse in the sinus node during this phenomenon. In one case we succeeded in obtaining a complete and detailed picture of the electricial behavior of the sinus node, revealing some basic mechanisms underlying the occurrence of sinus echoes. Thus the following could be established. (1) Analogous to longitudinal dissociation in the atrioventricular node, in the sinoatrial node, also, dissociation in conduction may occur. (2) Circus movement within the sinus node is possible. However, the dimensions of such an intranodal circuit are extremely small, the diameter being between 1 and 2 mm; the average conduction velocity in the circuitous pathway being 2.5 cm/sec. (3) No anatomical or pathological obstacle was involved in the present observation of sinus node reentry. The fibers in the center of the circuit, showing completely normal characteristics during sinus rhythm, were kept depolarized by electrotonic depolarizing current during circulation of the impulse around them. (4) After termination of sinus node reentry, a temporary shift of the dominant pacemaker occurred. This pacemaker shift was due to the fact that, at the moment the circulating impulse was blocked, the group of fibers lying distal from the site of block, while now not being discharged by their neighboring fibers, got the opportunity to reach threshold themselves, thus temporarily taking over the role of pacemaker of the heart Circ Res 44: , 1979 THE POSSIBILITY of a "circus rhythm" involving the sinus node was first suggested as early as 1943 by Barker et al. In recent years there has been increasing interest in sinus node reentry as a possible explanation for atrial echo beats and certain types of supraventricular tachycardia in man (Childers et al., 1973; Paulay et al., 1973a, 1973b; Narula, 1974; Wu et al, 1975; Weisfogel et al., 1975; From the Department of Physiology, Biomedical Center, University of Limburg, The Netherlands. Address for reprints: MauriU A. Allessie, M.D., Department of Physiology, University of Limburg, P.O. Box 616, 6200 MD Maastricht, The Netherlands. Received May 4, 1978; accepted for publication November 9, Pahlajani et al., 1975; Dhingra et al., 1975; Curry and Krikler, 1977; Breithardt and Seipel, 1978; Ticzon et al., 1975). However, because until now no one has been able to record directly the electrical activity of the sinus node in man, the diagnosis of sinus node reentry must be based solely on indirect evidence. Criteria such as P wave morphology, configuration of atrial electrograms, sequence of activation between high and low atrium, initiation and termination of paroxysms of tachycardia by a single stimulus, the effect of carotid sinus massage, the effect of shortening the cycle length by atrial pacing (Dhingra et al., 1975; Ticzon et al., 1975), the administration of verapamil (Curry and Krikler, 1977),

2 558 CIRCULATION RESEARCH VOL. 44, No. 4, APRIL 1979 and very recently, endocardial mapping procedures (Breithardt and Seipel, 1978), all have been used in an attempt to identify sinus node reentry in patients. However, despite these efforts, the diagnosis of sinus node reentrant tachycardia still must be considered speculative. This is caused not only by the unavoidable limitations of the methods that can be used in patients, but certainly to no less degree by the fact that, also, in animal experiments, reentry within the sinus node could not as yet be demonstrated with certainty. There are only few studies in which someone has tried to identify directly a reentrant circuit in the sinus node (Han et al., 1968; Bonke et al., 1971; Klein et al., 1973; Strauss et al., 1977). With a single microelectrode at hand, these investigators have attempted to follow the pathway of the activation wave within the sinus node after the induction of an atrial premature beat which was followed by a subsequent beat suspected to be based on sinus node reentry. The most complete data undoubtedly have been obtained by Han et al. (1968). These authors recorded the transmembrane potentials at 18 different sites in the isolated sinus node of the rabbit and the neighboring atrial tissue. Although this study indeed strongly suggests that sinoatrial reciprocation is possible, the authors themselves state that their results cannot be taken as absolute proof of the existence of sinus node reentry, because: "Definitive proof of a circuit within the sinus node would require a number of simultaneous impalements of intranodal cells, and a point-to-point demonstration of the course of the activation front during the whole interval between primary and reentrant atrial responses." The purpose of the present study was to achieve this high level of resolution in an attempt to meet the criteria for a material proof of sinus node reentry. Methods Young adult New Zealand rabbits were anesthetized by intramuscular injection of Hypnorm (10 mg fluanisone mg fentanyl base/kg). Half an hour after the injection, a tracheotomy was made and artificial positive pressure respiration was applied. The thorax was opened by a midsternal incision and the heart was quickly removed. After being immersed in Tyrode's solution, the ventricles were separated from the atria along the atrioventricular groove. Then, the left atrium, the interatrial septum, and the right appendage were dissected and discarded. The preparation that remained consisted of the wall of the vena cava, including the sinus node, and part of the adjacent right atrial musculature. After the vena cava had been opened by an incision along its long axis at the ventral side, the preparation was fixed in a tissue bath with the endocardial side uppermost. The superperfusion fluid had the following composition (in mm): NaCl, 130; KC1, 5.6; CaCl 2, 2.2; MgCl 2, 0.6; NaHCO 3, 24.2; NaH 2 PO 4,1.2; glucose, 11; and sucrose, 13. The fluid was saturated with a mixture of 95% O 2 and 5% CO 2 ; ph was 7.35 ± 0.05 and temperature was 37 ± 0.1 C. The rate of supervision was 100 ml/min. Stimulation technique, microelectrode recording, and data processing were the same as described previously (Bonke et al., 1971; Allessie, 1977). However, a new system for multiple surface recordings was used. To increase the maximum number of electrograms that could be registered simultaneously, we developed a system that can record up to 192 surface electrograms. In the study reported in this paper, 32 unipolar electrograms were recorded simultaneously from the right atrium using a regularly spaced array of electrodes (Teflon-coated silver wires; diameter, 0.25 mm; distance between individual electrodes, 1.4 mm). A large Ag-AgCl plate, placed in the tissue bath, served as a common indifferent electrode. To allow storage of all these signals, the parallel inputs were multiplexed and digitized (sample rate 2000 HZ) using a pulse code modulation (PCM) system (Kayser K ). The single serial output of this system was recorded on one channel of a 14-channel instrumentation recorder with wide band specifications (Ampex PR 2230), running at a speed of 60 inches/sec. A reference electrogram, microelectrode recordings from sinus node fibers, and the applied test stimuli were stored on separate channels of the magnetic tape. After the experiment, the tape was played back at a slower speed (15 inches/sec) and the single serial output, containing the information of 32 different endocardial recordings, was separated again in 32 parallel outputs with the aid of a PCM decoder unit (Kayser K and K ). The electrograms were written on a strip chart recorder (Schwarzer RS 266, paper speed 100 mm/sec) in sets of eight, together with the reference electrogram, the transmembrane potential of a sinus node fiber, and the applied test stimulus. This procedure resulted in a time resolution of better than 2 msec. The preparations were allowed to beat spontaneously. After each 15th spontaneous beat, a single premature stimulus (duration 1 msec, intensity 4 X diastolic threshold) was applied to the atrium. The moment of the premature stimulus was varied in steps of 1-5 msec and the whole spontaneous cycle length was scanned. In this way, a complete relationship between the curtailed atrial cycle (Ai-A 2 ) and the postextrasystolic cycle (A 2 -A3) was obtained. In some experiments a very early premature beat was followed by a response which came much earlier than expected. This response was suspected to be based on sinus node reentry. In one experiment we succeeded in reproducing such spontaneous early responses during a period of almost 1.5 hour. During this period, the qualities and timing of the premature stimulus were kept constant and the interval between the induced ectopic premature beat and the following reentrant atrial activation varied only in trifling degree (less than 5%). With the heart beating at a rate of 150/min

3 SINUS NODE REENTRY/Allessie and Bonke 559 and the premature stimulus given after every 15th beat, in this experiment identical sinus echoes could be reproduced for almost a thousand times. This gave us the opportunity not only to map the spread of activation in the right atrium, but also to follow the formation and conduction of the impulse within the sinus node, both during normal sinus rhythm, the premature ectopic beat, and the early extra response. With a single microelectrode, the transmembrane potentials of as many sinus node fibers as possible were recorded consecutively. The position of the impaled fibers was determined by reading the vernier scales of the manipulator in which the exploring microelectrode was mounted (accuracy better than 0.01 mm). In the most complete experiment, the transmembrane potentials of 130 different sinus node fibers were recorded. From the unipolar surface electrograms, the rapid intrinsic deflection was taken as the moment of activation, whereas the activation times of the sinus node fibers were measured at the 50% level of phase 0 depolarization, both for quickly and slowly rising potentials. Results Figures 1 and 2 illustrate the type of response which was suspected to be based on sinus node reentry. In Figure 1 an atrial electrogram is shown during the application of a single stimulus at four different prematurities. In Figure 2 the values of all curtailed cycles (Ai A 2 ) are plotted against the first and second postextrasystolic intervals (A 2 A3 and r\ * ^^O **! A3 At, respectively). Late atrial premature beats were followed by a full compensatory pause (Ai A 2 + A 2 A3 = 2 X Ai Ai), indicating that the normal impulse formation in the sinus node was not influenced. At shorter coupling intervals, the A 2 A3 cycles were still longer than Ai Ai, but started to become less than compensatory. This is caused by retrograde activation of the sinoatrial node by the early ectopic impulse, resulting in reset of the dominant pacemaker fibers (Bonke et al., 1969; Strauss et al., 1973). At the induction of very early premature beats, however, a particular response occurred. Instead of being followed by an A 2 A3 interval which is longer than the normal cycle length, the premature beat now suddenly was followed by an A 2 A3 interval which was markedly shorter than normal (see second and third trace of Fig. 1). If in such a situation the sum of the Ai A 2 and the A 2 A3 interval is about the same as a single Ai Ai interval, the most probable explanation is that the premature beat has been unable to invade the sinus node. The extrasystole then appears as an interpolated beat. In this case, however, the sum of the Ai A 2 and A 2 A3 intervals was markedly shorter than a single A] Ai cycle. In the curve given in Figure 2, the respective data points are clearly falling below the line of interpolation. Such early atrial responses are generally suspected to be based on sinus echoes (Childers et al., 1973; Ticzon et al., 1975; Dhingra et al., 1975; Strauss et al., 1976). The range of Ai A 2 intervals in which the premature beat was followed by a second, early response was 390 FIGURE 1 Induction of sinus node echoes by a single premature stimulus. The four traces show the different responses of a spontaneously beating atrium during subsequent administration of a single stimulus of increasing prematurity. In the upper trace, a premature beat with an Ai A-i interval of 101 msec results in reset of the sinus node, the A 2 Ai interval of 500 msec being longer than a normal Ai Ai interval but less than compensatory. At shorter A\ Ai intervals (second and third trace), after the induction of the Ai response an early A* response occurs (A\-A 2 + A2 A3 < A\ A\). These A3 responses were suspected of being based on sinus node reentry. When the stimulus is given at still shorter coupling intervals, it falls in the refractory period of the atrium, and the spontaneous sinus rhythm is not affected at all (lower trace).

4 560 CIRCULATION RESEARCH VOL. 44, No. 4, APRIL 1979 RESET,"3L.CO*IPENSATOHY SINUS ECHOES A, A, FIGURE 2 Effect of a single ectopic beat (Ai) on sinus rhythm. The curtailed cycle (Ax Ai) is plotted against the first and second postextrasystolic cycles (A2 A3 and A1 A4). Instead of plotting the absolute values of these cycles, the intervals are normalized by dividing them by the previous interval of the normal sinus rhythm (A\ A\). Although this procedure is not really necessary in isolated heart preparations because sinus rhythm is very stable, this presentation was chosen to facilitate comparison with clinical studies. The upper and lower oblique lines on the graph represent the compensatory line (A, A 2 + A 2 A* = 2 X A\ - A\) and the line of interpolation (A\ A 2 + A 2 A 3 = A\ A\), respectively. quite narrow, ranging no more than a few milliseconds, the lower limit being determined by the refractory period of the atrium. Note that the short A2 A3 intervals were associated with clearly prolonged A3 A( intervals. The reason for this prolongation will be clarified in one of the next paragraphs. 1.0 Multiple Intracellular Recordings during Sinus Node Echoes In an attempt to demonstrate the underlying mechanism for the early responses described in the previous paragraph, we tried to map the excitation of the sinus node area during repeated induction of this phenomenon. This turned out to be not an easy task. In most of the experiments we completely failed to evoke sinus node echoes. In other preparations, sinus echoes could be produced occasionally, but they were not stable enough to allow completion of the mapping procedure. In still others, the mapping procedure was prevented by technical difficulties such as difficulty in getting rapid and good impalements of sinus node fibers or, most frequently, because the tip of the microelectrode was broken before sufficient recordings could be made. In the face of such frequent and common misfortune associated with this kind of experiment, it should be considered highly fortunate that in one experiment we managed to obtain a complete excitation map of the sinus node during repeated induction of a stable sinus node echo. Figure 3 shows the sites in the atrium where local unipolar surface electrograms were recorded and the position of the sinus node fibers from which intracellular recordings were obtained. A total of 32 atrial electrograms and 130 sinus node fibers were studied. Whereas the atrial surface electrodes were spaced in a regular array, the recorded sinus node fibers were unequally distributed over the sinus node area. In the upper and lower parts of the node, the distance between the different impalements was 0.7 mm. In the center of the sinus node, a higher resolution was obtained by making intracellular recordings every 0.35 mm. In addition, there was a very distinct area where we considered it desirable to achieve an even higher spatial resolution. There the distance between the recorded sinus node fibers amounted to only 0.1 mm. 2 mm FIGURE 3 Sketch of the isolated right atrium of the rabbit. SVC = superior vena cava, IVC = inferior vena cava, CT = crista terminalis. The large dots on the atrial myocardium indicate the 32 sites where unipolar surface electrograms were recorded. The small dots on the sinus node area indicate the position of the 130 different sinus node fibers from which intracellular recordings were obtained.

5 SINUS NODE REENTRY/Allessie and Bonke 561 In Figure 4, the transmembrane potentials of 16 sinus node fibers are shown, together with a reference atrial electrogram. The first two rows of action potentials are the last two beats of a series of normal sinus beats. From the moment of discharge and the configuration of the action potentials during sinus rhythm, the different cell types of the sinus node can be recognized. There is a group of fibers, such as fibers 5, 6, 7, and 8, which show all the features of dominant pacemaker fibers (early discharge, high rate of diastolic depolarization, gradual transition into phase 0, low values of dv/dt max, low amplitude and long duration of the action potential). There are others which are clear representatives of latent of subsidiary pacemakers (i.e., fibers 1, 15, and 16). These fibers exhibit practically no diastolic depolarization, are discharged considerably later, show a sudden transition into phase 0, and generate action potentials with a rapid upstroke and a high amplitude. Most of the sinus node fibers (i.e., fibers 2, 3, 4, 9, 10, 11, 12, 13, and 14) have qualities which are somewhere in between the characteristics of these two different cell types. At the moment indicated by the dotted line, a single premature stimulus was applied to the atrium. The delay between the last spontaneous activation and the application of the stimulus was 80 msec. Instead of evoking merely a single premature beat, this stimulus was followed by two rapidly succeeding responses. The curtailed cycle and the postextrasystolic cycle of the atrium were respectively 98 and 229 msec. Looking at the action potentials recorded from the different sinus node fibers during this primary and secondary response, one can observe several interesting phenomena. First of all, some fibers, such as fibers 11, 12, 13, and 14, failed to respond to the premature impulse. In those fibers, initially only local responses of low amplitude, slow rate of rise, and short duration were recorded during the premature beat. On the other hand, there are other fibers (i.e., fibers 15 and 16, and fibers 1 to 10) that did respond to the premature retrograde activation wave. This occurrence of partial sinoatrial entrance block demonstrates the possibility of dissociation within the sinus node, comparable to the longitudinal dissociation observed in the atrioventricular node (Mendez and Moe, 1966; Janse et al., 1971). In those areas where the premature impulse was propagated, the conduction velocity was markedly reduced, as can be concluded from the reduction in amplitude and rate of rise of the premature action potentials. This was especially true for the area of the dominant pacemaker (fibers 5, 6, 7, and 8). This loss in stimulating efficacy of the impulse got even worse during the propagation of the second early response. By then, the amplitude and rate of rise of the action potentials of fibers 6 and 7 had been reduced to such a degree that the propagation beyond these fibers failed. In fibers 8 to 10 only electrotonic "humps" were recorded during the second response, pointing to an extinction of the impulse somewhere in this part of the sinus node. sinus rhythm 500 msec ectopic beat JT. sinus echo FIGURE 4 Intracellular recordings of 16 different sinoatrial node fibers together with a reference atrial electrogram during the induction of a sinus node echo. The different recordings were not made simultaneously but, instead, the figure is composed of separate intracellular recordings obtained with the same microelectrode during repeated induction of identical sinus echoes. The tracing starts with the last two action potentials of a series of 15 undisturbed spontaneous sinus beats. The vertical dotted line indicates the moment of the premature stimulus applied to the atrium. This single stimulus is followed by an ectopic premature beat and a subsequent sinus echo beat. From the individual responses of the different sinus node fibers, the pathway of the impulse within the sinus node between the early ectopic beat and the subsequent extra beat can be traced. The exact location of the present 16 sinus node fibers and the circuitous pathway within the sinus node which is recognized from these recordings are given in Figure 5.

6 562 CIRCULATION RESEARCH VOL. 44, No. 4, APRIL 1979 Mapping of Sinus Node Echoes Figure 5 gives the sequence of excitation of the 16 fibers presented in Figure 4 in a ladder diagram. Above the diagram the preparation is given schematically. In the left sketch, the positions of the 16 different sinus node fibers are marked, whereas in the right scheme the pathway of the premature impulse within the sinus node is indicated. When the preparation was beating spontaneously, the imslnus rhythm JT sinus echo FIGURE 5 Bottom: Ladder diagram belonging to the selection of 16 sinus node fibers shown in Figure 4. During sinus rhythm, the earliest moment of discharge that could be found in the sinus node (fibers 5, 6, and 8) has been taken as zero reference. The moments of activation during the ectopic premature beat and the sinus echo refer to the moment of stimulation as indicated by the vertical broken line. Activation times are given in milliseconds. In the upper part of the figure, a sketch of the isolated right atrium is given. At the left, the exact locations of the 16 sinus node fibers, the atrial reference electrode, and the site of stimulation are indicated. At the right, a schematic presentation is given of the circuitous pathway of the impulse within the sinus node as it could be identified to occur during the sinus echo. SVC = superior vena cava, IVC = inferior vena cava, CT = crista terminalis. Double bars indicate conduction block. pulse arising in the area of fibers 5 to 8 spread into all directions, the atrium being activated after 30 msec. After application of the atrial stimulus, the sinus node was invaded retrogradely. Fibers 15 and 16 lying at the lower part of the sinoatrial border were activated with a delay of 20 and 30 msec, respectively. However, fibers lying a little bit deeper in the tail of the sinus node (fibers 11 to 14) only showed electrotonic humps, obviously because they were not able to respond to the premature impulse. Thus, the lower part of the node showed the phenomenon of entrance block. In the upper part of the sinus node, however, such sinoatrial entrance block did not occur. There the retrograde activation wave was able to penetrate further into the sinus node and was conducted slowly from fiber 1 to 2 to 3 to 4, etc. After the impulse entered the sinus node at the upper part, it then traveled downward within the node, going from fiber 7 to fiber 11. As a result, the fibers lying deep in the lower part of the node (fibers 8, 9, and 10) were activated by the ectopic impulse in a roundabout way, respectively 135, 160, and 168 msec after the premature stimulus. Also fibers 11 to 14, which initially exhibited only local responses, were then truly activated, the impulse now coming from the opposite direction, at 188,195, 203, and 210 msec after the stimulus. As a result of this circuitous route taken by the premature impulse, fibers 1 and 15, which already had been activated directly by the ectopic wave front, were reexcited by the turning impulse at t = 225 msec. Obviously the time required for the impulse to turn around in the sinus node had been sufficient for these fibers to restore their excitability. Since fibers 1 and 15 were located at the sinoatrial border, the reexcitation of these fibers allowed the impulse to escape again from the sinus node, and from this point the atrium was reentered. The surface electrode located low on the crista terminalis recorded the second atrial activation at t = 247 msec. In our experiments, we observed only single sinus node echoes and no sustained sinus node reentry. The reason for this can be derived from Figures 4 and 5. Halfway in its second circus movement the impulse died out somewhere beyond fiber 7. Therefore, instead of being trapped in a sustained sinus node reentry, the impulse described only one and a half revolution, resulting in a single sinus echo beat. The area where the circulating impulse was blocked coincides with the area of the dominant pacemaker fibers. It is known that these fibers have both a relatively low safety factor and a long refractory period. On its first roundtrip, the impulse already had difficulty in propagating through this area. The premature action potentials recorded from fibers 5 to 7 were markedly reduced in amplitude and upstroke velocity. However, conduction then still was favored by the relatively long time interval which elapsed between the last spontaneous discharge and the premature activation of these fibers. As can be seen from the ladder diagram, especially in fibers 6

7 SINUS NODE REENTRY/Allessie and Bonke 563 and 7, the second activation came after a much shorter interval. At the moment these fibers were excited again, they were far from restored from the foregoing activation. This resulted in termination of the circus movement. Figure 6 shows the complete excitation maps of the sinoatrial region. During sinus rhythm there was slightly preferential conduction from the pacemaker center to the cranial part of the crista terminalis, antegrade sinoatrial conduction time being 20 msec. In the direction of the interatrial septum, the impulse propagated extremely slowly, resulting in an area of conduction block at the border between the sinus node and the interatrial septum. Comparison of these data with other mapping studies of sinoatrial conduction (Sano and Yamagishi, 1965; Bouman et al., 1978; Steinbeck et al., 1978) indicates that this pattern of antegrade conduction must be considered as completely normal for the rabbit. Thus, during sinus rhythm, conduction was not depressed and there were no areas of abnormal local conduction block. The propagation of the ectopic beat was associated with both intra-atrial and intranodal conduction block. As a result, the retrograde impulse did not succeed in traversing the node, but was blocked along a line as indicated on the sketch below the map. At the lower part of the node, the impulse already had died out 30 msec after the stimulus, at a distance of about 1 mm from the crista terminalis. At the middle part of the node, the impulse traveled in two directions: (1) slowly upward, parallel to the FIGURE 6 Maps of impulse formation and conduction in the sinoatrial region during normal sinus rhythm (A), an induced early ectopic premature beat (B), and the resulting sinus echo (C). The maps are composed from time measurements of 32 simultaneously recorded atrial electrograms and 130 consecutively impaled sinus node fibers. The exact location of the different recordings is indicated in Figure 3. The individual moments of activation at the respective sites are indicated on the maps. During normal sinus rhythm, the moment of the earliest discharging fibers is taken as t 0. The maps of the ectopic beat and the sinus echo beat are related to the moment of the stimulus. The delay between the last normal sinus beat and the application of the stimulus was 80 msec. Local conduction block has been indicated on the maps by the use of a minus sign. Because of shortage of space in the center of the sinus node, not all measured values could be plotted. Instead, only some representative values are given there. In the sketches below, the spread of activation during the three beats is indicated schematically. Double bars indicate conduction block. See text for further description.

8 564 CIRCULATION RESEARCH VOL. 44, No. 4, APRIL 1979 crista terminalis, until about 70 msec after the stimulus the impulse extinguished high in the upper part of the node, and (2) even more slowly in the direction perpendicular to the crista terminalis, penetrating the center of the sinus node. In the map of the ectopic beat, isochrones have been drawn up to t = 90 msec. From this moment, up to t = 200 msec, conduction in the sinus node was so slow and so fragmentated that only some representative individual activation times are given. From these activation times, the narrow and erratic pathway of the impulse within the sinus node can be followed, and it can be seen that the deepest part of the sinus node is activated at about t = 170 msec. Until that time the fibers located caudal to this area were not yet activated, because direct retrograde invasion of the tail of the sinus node had failed. These fibers now were activated in an antegrade direction by the impulse which had penetrated deep in the sinus node. In the map of the sinus echo, one can see the result of this turning back of the impulse. On its way back to the atrium, the impulse traveled through the fibers in the tail and mid-lower portion of the sinus node, in the time between t = 170 and t = 230 msec. After the impulse had made a complete loop, it left the sinus node again at about the same site where it had entered. The critical moment at which the first fibers in the sinoatrial border were reexcited by the turning impulse was t = 230 msec. Since these fibers already had been activated antegradely at t = 30 msec, the time for recovery of their excitability was about 200 msec. This relatively long period is caused by the extremely slow conduction of the premature impulse in the intranodal circuit. Since the diameter of the circuit can be estimated between 1 and 2 mm, and the revolution time was 200 msec, the mean conduction velocity of the circulating impulse can be calculated to be about 2.5 cm/sec. After the impulse had reexcited the atrium at t = 240 msec, it continued its circuitous route within the sinus node. However, halfway its second roundtrip, the impulse was blocked and the circus movement was terminated. Intracellular Recordings from the Center of Circus Movement within the Sinus Node Figure 7 depicts the transmembrane potentials of the sinus node fibers, located in the center of the intranodal circuit. As mentioned earlier, in this area a very high spatial resolution was obtained by recording action potentials at distances of only 0.1 mm. At the left part of the figure, the location of the circuit is given together with at a higher mag- 2mm sinus echo 500ms«c FIGURE 7 Electrical behavior of the center of the intranodal circuit. At the left part of the figure, the localization and dimensions of the circuit within the sinus node are depicted. Immediately to the right, the circuit is given at a higher magnification. In this scheme, three rows of recording sites (marked A, B, and C) in the center of the circuit are indicated. Each dot represents a successful impalement of a sinus node fiber at that particular point. The distance between the different rows was 150 fim, whereas the distance between the recording sites at each row was 100 tun. In the right side of the figure, the transmembrane potentials recorded at the corresponding points during normal sinus rhythm and the induction of a sinus echo are given. These different recordings were not obtained simultaneously, but with a single microelectrode numerous fibers were impaled consecutively during repeated induction of identical sinus echoes. The recordings were time aligned, using one of the atrial electrograms as a reference. Each tracing starts with the last action potential of a series of 15 normal sinus beats. The broken vertical line indicates the moment of the stimulus applicated to the atrium. During sinus rhythm, perfectly normal action potentials were recorded. There are no signs of abnormally depressed conduction or areas of local conduction block. During the circuitous pathway of the impulse around this area, strong electrotonic interaction becomes apparent. The circulating depolarization wave exerts a clear electrotonic depolarizing effect on the fibers within its center, leading to either a prolonged depolarization (fibers 1-5) or fusion responses (fibers 6-10). B

9 SINUS NODE REENTRY/Allessie and Bonke 565 nification the actual sites of recording from the center of the circuit. The three rows of impalements that were made are marked A, B, and C, and the individual fibers at a single row are numbered. In the right part of the figure, the corresponding intracellular recordings, as obtained during the sinus echo, are shown. A comparison of all these transmembrane potentials, recorded from different but closely adjacent fibers, provides evidence that there was strong electrotonic interaction between these fibers. In fact, one can conclude that the rather complex course of the transmembrane potential in the center of the circuit is determined mainly by electrotonic influences of the depolarization wave turning around this area. The fibers at the upper part of the center (fibers 1 to 5) exhibit a prolonged depolarization in the time interval between the premature impulse and the sinus echo response. This shift of the diastolic membrane potential to less negative values coincides with the roundtrip of the impulse along the upper side of the center. In other words, as a result of electrotonic interaction, the fibers in the "eye" of the vortex are kept constantly depolarized by the depolarization wave, circulating around it. One of the clearest examples of this phenomenon is fiber 5 of row B, which was located in the very center of the circuit. The fibers in the lower part of the center (fibers 6 to 10) exhibit action potentials during the sinus echo, which are markedly prolonged. These long action potentials consist of two components. The first response coincides with the descending limb of the circuit and the second component is synchronous with the impulse traveling in the ascending limb of the circuit. This again demonstrates that there is electrotonic interaction between the two limbs. The resulting fusion response, recorded in the area between the two pathways, is caused by the small distance between the two limbs in this part of the circuit and by the relatively short delay between the activation of the two limbs. Pacemaker Shift in the Sinus Node after a Sinus Echo As already noted, the first spontaneous sinus beat after the occurrence of a sinus echo came later than expected. The post-return cycle (A3 A<), associated with sinus node reentry, was clearly prolonged (see Fig. 2). It has been suggested that, for the differentiation between a sinus echo and an interpolated beat, the finding of a subsequent prolonged A3 Ai cycle is in favor of sinus node reentry (Strauss and Geer, 1977). Our results support this suggestion. In Figure 8, the maps during normal sinus rhythm and the first spontaneous beat after the sinus echo are compared. They show a distinct shift of the dominant pacemaker within the sinus node, the first beat after the sinus echo originating about 2 mm lower in the sinus node than the normal sinus impulses. This sudden shift of the pacemaker in the direction of the tail of the sinus node is further documented in the lower part of the figure. Here the transmembrane potentials recorded from the two different pacemaker centers (marked A and B) are shown. The action potentials during normal sinus rhythm and during the first post-reentry sinus beat, recorded at sites A and B, are superimposed, Normal Sinus Rhythm atrium First Sinus Beat After Sinus Echo 5 mm FIGURE 8 Demonstration of pacemaker shift after the occurrence of sinus node reentry. The two maps at the top of the figure show the site of impulse formation in the sinus node (indicated by asterisks) and the spread of activation toward the atrium, during normal sinus rhythm and the first spontaneous sinus beat after the sinus echo. Isochrones are drawn every 10 msec, the first isochrone around the site of impulse formation being that of t = 5 msec. During the first sinus beat after termination of sinus node reentry, the dominant pacemaker has been shifted over a distance of about 2 mm in the direction of the tail of the sinus node. Shown in the lower part of the figure are the transmembrane potentials of the two sinus node fibers which alternately acted as the dominant pacemaker. Using the atrium electrogram as a reference, the action potentials recorded during normal sinus rhythm and the first spontaneous sinus beat after the sinus echo are superimposed. The unbroken tracings were recorded during normal sinus rhythm, whereas the interrupted tracings belong to the sinus node "escape beat." It is clear that, during normal sinus rhythm, fiber A takes the lead, whereas during the first sinus beat after termination of sinus node reentry, fiber B is discharging considerably earlier than fiber A.

10 566 CIRCULATION RESEARCH VOL. 44, No. 4, APRIL 1979 using the atrial electrogram as common time reference. During normal sinus rhythm (unbroken tracings), fiber A belonged to the group of dominant pacemaker fibers, discharging 30 msec before the activation of the atrium. Fiber B had all the characteristics of a subsidiary pacemaker fiber. During the first spontaneous sinus beat after the sinus echo (dotted tracings), this relationship was completely reversed. Now fiber B belonged to the first discharging fibers (33 msec before the atrium), whereas fiber A had become a latent pacemaker fiber, being activated only 13 msec before the atrium. In Figure 9, an explanation for this intranodal pacemaker shift is offered. In the left part of this figure, the sites of the two pacemaker centers are indicated in relation to the intranodal circuit. The upper hatched area is the pacemaker area during normal sinus rhythm, whereas the lower hatched area indicates the locus of impulse formation during the first sinus beat after termination of sinus node reentry. At the right side of the figure, the transmembrane potentials of three different fibers are shown during the induction of a sinus node echo. Fiber 1 is representative of the group of dominant pacemaker fibers during normal sinus rhythm. Fiber 3 is the fiber found to discharge earliest during the first beat after termination of sinus node reentry, and fiber 2 is taken from the area where the circulating excitation died out. It is important to note that the normal pacemaker center was located on that part of the circuit FIGURE 9 Explanation of the mechanism behind the pacemaker shift as demonstrated in Figure 8. At the left, the sinus node circuit during the echo beat is given, together with the areas of pacemaking during normal sinus rhythm (upper hatched area) and the first spontaneous sinus beat after the echo beat (lower hatched area). At the right, the responses of three different fibers are shown during the A\, Ai, A& and A t beats. Fiber 1 is a dominant pacemaker fiber during normal sinus rhythm and fiber 3 is the dominant pacemaker fiber during the A A beat, whereas fiber 2 is lying on the sinus node circuit in the region where the circulating impulse is blocked. The intervals prior to the SAN* response, which are different for the different fibers, are indicated in milliseconds. See text for further discussion. which was activated twice during the one and a half revolution of the impulse. In contrast, the area that temporarily took over the dominancy of pacing was activated only once during the sinus echo. In other words, the original pacemaker was located proximal to the site where the circulating impulse died out, whereas the new pacemaker was situated distal to the area of block. This offers a very simple explanation for the occurrence of a pacemaker shift: in fibers located distal to the site of block of the circulating impulse (like fiber 3) the SAN 3 response is missing. Therefore, although the intrinsic rate of impulse formation in these fibers may be lower than in the original pacemaker fibers, their earlier start will allow them to discharge earlier and take over the dominance from the original pacemaker. The intrinsic cycle length of impulse formation of fiber 3 turned out to be 440 msec. This is only 40 msec longer than the cycle length of the dominant pacemaker during undisturbed sinus rhythm. Since the last depolarization prior to SAN< in fiber 3 occurred at t = 188 msec and in fiber 1 at t = 265 msec, fiber 3 had a lead of 77 msec in the process of impulse formation. Since this start was considerably greater than the 40 msec which fiber 3 needed more for its impulse formation compared with fiber 1, fiber 3 was the first to reach threshold and consequently acted as dominant pacemaker for the first spontaneous beat after sinus node reentry. This intranodal pacemaker shift was only temporary. During the subsequent sinus beats, the pacemaker shifted back again to its original site. From the recordings in Figure 9, it also can be understood why fiber 2 did not become the new pacemaker. Although this fiber was also located distal to the site of block of the reentrant impulse, and the interval prior to the SAN< response was even longer than in fiber 3 (490 vs. 440 msec); still fiber 2 was subsidiary to fiber 3 during the A< response. The explanation is that fiber 2 was located so close to the site where the circulating impulse died out that it was still under the electrotonic influence of the blocked impulse. Consequently, in direct junction with the SAN2 response, an electrotonic hump was recorded in this fiber. This largely prolonged the effective duration of the SAN2 response, thus preventing the fiber from taking over the pacemaker function. Discussion Our results must be compared with those of other studies in which the investigators tried to demonstrate sinus node reentry with the aid of microelectrode recordings (Han et al., 1968; Bonke et al, 1971; Klein et al., 1973; Strauss and Geer, 1977). Among these, the study of Han, Malozzi, and Moe was not only the earliest, but also the most successful in this respect. As in our experiments, they used an isolated right atrium of the rabbit and a single microelectrode for consecutive recording of action

11 SINUS NODE REENTRY/AUessie and Bonke 567 potentials from the sinus node and the neighboring atrial tissue. An important difference between their study and ours is that Han et al. elicited sinus echoes during constant regular pacing of the atrium, whereas we allowed the preparations to beat spontaneously. Another essential difference of course is the number of sinus node fibers that could be impaled during the period in which stable reciprocal responses were obtained. Han, Malozzi, and Moe succeeded in recording the time relations at 18 different sites in the node and the sinoatrial border. We were fortunate to establish the moments of activation during a sinus echo at as many as 130 different sinus node fibers and 32 points of the adjacent atrial myocardium. There are marked similarities between their and our results. For instance, the transmembrane potentials that Han et al. showed as examples of slow and fractionated conduction of a premature impulse within the sinus node are identical with the configuration of action potentials we recorded from certain sinus node fibers during sinus node reentry.* Also, the three possible responses of sinus node fibers during a sinus echo, depending on their relative position on the circuit, as recognized by Han et al. (see their Fig. 4), could be demonstrated. We found one group of fibers exhibiting only a single response (SAN 2 ), which was sandwiched between the primary and secondary atrial responses (A2 and A3). These fibers responded to the atrial ectopic beat, but failed to be reentered by the circulating impulse (no SAN3). Another group of fibers also showed a single response, which, however, occurred simultaneously with or after reentry of the atrium. These fibers, which were located at a greater distance from the sinoatrial border, were not reached by the early wave front invading the sinus node retrogradely (no SAN 2 ), but they were secondarily activated after the impulse had made one complete circus movement in another area of the sinus node. Thus, these fibers exhibited only a late SAN 3 response. The third group of sinus node fibers showed, like the atrium, a double response SAN 2 + SAN 3. This area of the sinus node was discharged directly by the retrograde activation wave, as well as activated a second time by the circulating impulse. In fact, it was only in these fibers that reentry of the sinus node took place. However, despite these similarities between the data provided by Han et al. and our present results, we come to a somewhat different interpretation. Han et al. explained their findings as a reciprocation between the atrium and the sinus node: "Early premature atrial responses which entered the sinus node, frequently emerged to reexcite the atrium as 'echoes.' When the premature atrial beats discharged some but not all areas, the atrial echoes, in ' Compare the response of the sinus node fiber presented in Figure 2, panels C and D, of the paper by Han et al, with the responses of sinoatrial fibers 11 to 14 presented in Figure 4 of this paper. turn, appeared to reenter the sinus node." However, because of the high spatial and temporal resolution achieved in our study, we were able to show that: 1. The circuit was located completely within the sinus node itself. The atrium was not an essential link in the reciprocal circuit. 2. The dimensions of the intranodal circus movement were extremely small, the diameter being in the order of 1-2 mm only. 3. The reentrant pathway was located in the center of the sinus node, at about the same area where during normal rhythm the sinus impulses were generated. If, with these conclusions in mind, we look again at the observations of Han et al., it seems possible that also in their case such a small intranodal reentrant circuit was responsible for the observed phenomena, instead of the supposed larger reciprocating circuit involving both the sinus node and the atrium. It also explains that, unless the sinus node is mapped with a very high spatial resolution, the presence of an intranodal circuit easily can be overlooked. Is the Dissociation in Sinoatrial Conduction Based on Anatomical or Electrophysiological Properties? It is still an unanswered question whether the longitudinal dissociation occurring in the atrioventricular node is a result of the existence of anatomically defined, so-called a and /? pathways, or whether this is due to purely functional differences in electrophysiological properties in different parts of the node (Mendez and Moe, 1966; Janse et al., 1971). On the basis of our electrophysiological measurements, we could not identify any anatomical or pathological substrate in the sinus node, which served as an obstacle for the impulse to circulate around. On the contrary, the fibers in the center of the circuit showed completely normal action potentials during spontaneous sinus rhythm (see Fig. 7). During sinus node reentry, their transmembrane potentials were largely determined by electrotonic interaction with the depolarization wave circulating around these central fibers. This strongly suggests that there was no electrical insulation between the different "pathways" when dissociation in conduction occurred, and that there was (passive) electrical cross-talk between the legs of the circuit during sinus node reentry. As a result of this electrical continuity across the center of the circuit, the fibers in the very center of the circus movement were kept constantly depolarized as long as the depolarization wave circulated around them. The fibers located somewhat more eccentrically in the central area showed a fusion response (see Fig. 7). Similar observations have been made by Strauss and Geer (1977). These authors recorded the transmembrane potential of a single sinus node fiber during repeated induction of sinus echoes with atrial premature

12 568 CIRCULATION RESEARCH VOL. 44, No. 4, APRIL 1979 depolarizations elicited at different A1-A2 intervals. When the prematurity of the atrial ectopic beat was gradually increased, they observed that the diastolic membrane potential between the SAN 2 and SAN 3 response was gradually reduced, and the interval between the SAN2 and SAN3 response progressively decreased. Finally, this culminated in a complete fusion response in the particular fiber. Strauss and Geer explained this sequence of events by assuming that the atrial ectopic beat induced a reentrant circuit in the sinus node, which moved closer and closer to the recording site as the atrial ectopic beat was elicited earlier. Our results seem to give a firm basis for this speculation. Together they strongly point toward the conclusion that the dissociation in sinoatrial conduction and the resulting sinus node reentry is based purely on electrophysiological characteristics of the sinus node and do not depend on the existence of anatomically defined pathways. On the other hand, Bouman et al. (1978) have found that, at the junction between the sinus node and the atrial septum, there is an area where, already during normal slow sinus rhythm, conduction of the impulse is blocked. This area of local block at the sinoseptal border indeed is a constant finding and was also present in our preparations (see Figure 6, left panel, and Steinbeck et al., 1978). Nevertheless, it did not participate in the circuit which could be identified during sinus node echoes. However, this does not exclude the possibility that under different circumstances such a locus of highly depressed conductivity may act as central obstacle for circus movement of the impulse. More generally, we wish to emphasize that under different circumstances the size and location of a reentrant circuit in the sinoatrial region may differ from the one identified in the present study. With an anatomical obstacle involved, the chances for a sustained reentrant mechanism seem to be higher. It may be speculated then that, instead of single sinus node echoes, shorter or longer periods of sustained sinus node reentrant tachycardia may occur. References Allessie MA: Circulating Excitation in the Heart. Maastricht, University Press, 1977 Barker PS, Wilson FN, Johnston FD: The mechanism of auricular paroxysmal tachycardia. Am Heart J 26: , 1943 Bonke FIM, Bouman LN, Van Rijn HE: Change of cardiac rhythm in the rabbit after an atrial premature beat. Circ Res 24: , 1969 Bonke FIM, Bouman LN, Schopman FJG: Effect of an early atrial premature beat on activity of the sinoatrial node and atrial rhythm in the rabbit. Circ Res 29: , 1971 Bouman LN, Mackaay AJC, Bleeker WK, Becker AE: Pacemaker shifts in the sinus node. Effects of vagal stimulation, temperature and reduction of extracellular calcium. In The Sinus Node. Structure, Function and Clinical Relevance, edited by FIM Bonke. The Hague, Martinus Nijhoff Publishing Co., 1978, pp Breithardt G, Seipel L: Sequence of atrial activation in patients with atrial echo beats. In The Sinus Node. Structure, Function and Clinical Relevance, edited by FIM Bonke. The Hague, Martinus Nijhoff, Publishing Co., 1978, pp Childers RW, Arnsdorf MF, Fuente DJ, Gambetta M, Svenson R: Sinus nodal echoes. Am J Cardiol 31: , 1973 Curry PVL, Krikler DM: Paroxysmal reciprocating sinus tachycardia. In Reentrant Arrhythmias. Mechanism and Treatment, edited by HE Kulbertus, Lancaster, MTP, 1977, pp Dhingra RC, Wyndham C Amat-y-Leon F, Denes P, Wu D, Rosen KM: Sinus nodal responses to atrial extra stimuli in patients without apparent sinus node disease. Am J Cardiol 36: , 1975 Han J, Malozzi AM, Moe GK: Sinoatrial reciprocation in the isolated rabbit heart. Circ Res 22: , 1968 Janse MJ, van Capelle FJL, Freud GE, Durrer D: Circus movement within the AV node as a basis for supra ventricular tachycardia as shown by multiple microelectrode recordings in the isolated rabbit heart. Circ Res 28: , 1971 Klein HO, Singer DH, Hoffman BF: Effects of atrial premature systoles on sinus rhythm in the rabbit. Circ Res 32: , 1973 Mendez C, Moe GK: Demonstration of a dual AV conduction system in the isolated heart. Circ Res 19: , 1966 Narula OS: Sinus node reentry. A mechanism for supraventricular tachycardia. Circulation 50: , 1974 Pahlajani DB, Miller RA, Serratto M: Sinus node reentry and sinus node tachycardia. Am Heart J 90: , 1975 Paulay KL, Varghese PJ, Damato AN: Sinus node reentry. An in vivo demonstration in the dog. Circ Res 32: , 1973a Paulay KL, Varghese PJ, Damato AN: Atrial rhythms in response to an early atrial premature depolarization in man. Am Heart J 85: , 1973b Sano T, Yamagishi S: Spread of excitation from the sinus node. Circ Res 16: , 1965 Steinbeck G, Allessie MA, Bonke FIM, Lammers WJEP: Sinus node response to premature atrial stimulation in the rabbit studied with multiple microelectrode impalements. Circ Res 43: , 1978 Strauss HC, Saroff AL, Bigger JT, Giardina EGV: Premature atrial stimulation as a key to the understanding of sinoatrial conduction in man. Circulation 47: 88-93, 1973 Strauss HC, Bigger JT, Saroff AL, Giardina EGV: Electrophysiologic evaluation of sinus node function in patients with sinus node dysfunction. Circulation 53: , 1976 Strauss HC, Geer MR: Sinoatrial node reentry. In Reentrant Arrhythmias. Mechanism and Treatment, edited by HE Kulbertus, MTP, Lancaster, 1977, pp Ticzon AR, Strauss HC, Gallagher JJ, Wallace AG: Sinus nodal function in the intact dog heart evaluated by premature atrial stimulation and atrial pacing. Am J Cardiol 35: , 1975 Weisfogel GM, Batsford WP, Paulay KL, Josephson ME, Ogunkelu JB, Akhtar M, Seides SF, Damato AN: Sinus node reentrant tachycardia in man. Am Heart J 90: , 1975 Wu D, Amat-y-Leon F, Denes P, Dhingra RC, Pietras RJ, Rosen KM: Demonstration of sustained sinus and atrial reentry as a mechanism of paroxysmal supraventricular tachycardia. Circulation 51: , 1975