Author + information
- Received December 12, 1996
- Revision received April 18, 1997
- Accepted July 1, 1997
- Published online October 1, 1997.
- Sanjiv K Gandhi, MDA,*,
- Burt I Bromberg, MDB,
- Mark D Rodefeld, MDA,
- Richard B Schuessler, PhDA,
- John P Boineau, MD, FACCA,
- James L Cox, MD, FACCA and
- Charles B Huddleston, MDA
- ↵*Dr. Sanjiv K. Gandhi, Cardiothoracic Research Laboratory, Box 8234-3308 CSRB, 660 South Euclid Avenue, Saint Louis, Missouri 63110.
Objectives. This study sought to 1) establish whether the atrial flutter (AFL) inducible acutely occurs spontaneously in a chronic canine model, and 2) characterize any reentrant circuits present chronically.
Background. We previously demonstrated, in an acute canine model of the modified Fontan operation, that the lateral tunnel suture line creates a sufficient electrophysiologic substrate for AFL.
Methods. Using cardiopulmonary bypass, a suture line was placed through a right atriotomy in adult dogs (n = 7) to simulate the lateral tunnel of the Fontan operation. Holter recordings were made preoperatively, on the first postoperative day and 2, 4 and 6 weeks postoperatively. At 6 to 8 weeks, through bilateral ventriculotomies, 253-point unipolar atrial electrodes were inserted. AFL was induced using atrial burst pacing, and endocardial activation sequence maps were created.
Results. Preoperatively, all dogs were in sinus rhythm. Spontaneous AFL occurred in all dogs postoperatively, with a mean (±SD) cycle length of 192 ± 22 ms. At 6 weeks postoperatively, of six dogs that survived, four had intermittent AFL, and two had incessant AFL. At reoperation, sustained AFL was inducible in six of six dogs, with a mean cycle length of 194 ± 17 ms. Activation sequence maps demonstrated conduction block at the lateral tunnel suture line, which facilitated unidirectional conduction critical for propagation of the reentrant circuit. The AFL circuit was similar to that observed acutely.
Conclusions. In a chronic canine model of the modified Fontan operation, the lateral tunnel suture line alone, in the absence of atrial stretch or hypertension, provides an electrophysiologic substrate that promotes spontaneous AFL. This model may be useful for evaluating various forms of treatment and prevention of AFL after the Fontan operation.
Atrial flutter (AFL), or intraatrial reentrant tachycardia, is a common early and late postoperative complication of the modified Fontan operation, or total cavopulmonary connection, with an incidence of up to 30% at 5-year follow-up [1–5]. These arrhythmias are not well tolerated in patients with single-ventricle physiology. Management with antiarrhythmic agents or by antitachycardia pacing is often ineffective [6–8].
In an acute canine model, we demonstrated that the Fontan lateral tunnel suture line, used to construct the right atrial baffle, which directs inferior vena caval blood to the superior vena cava (which is previously anastomosed to the pulmonary artery), was a sufficient anatomic substrate to produce AFL . There were several limitations of the acute canine model. The arrhythmias induced acutely may not have accurately represented spontaneous AFL after the Fontan operation. AFL was induced with programmed stimulation, introducing the possibility that the arrhythmias observed were nonspecific findings. It also remained to be determined whether the critical features of the reentrant circuit would persist over time. The objectives of the present study were to 1) establish whether the AFL inducible acutely occurs spontaneously, and 2) characterize any reentrant circuits present chronically.
1.1 Acute phase operative technique.
Seven adult mongrel dogs (25 to 30 kg) free of cardiac disease were anesthetized with intravenous pentobarbital sodium (30 mg/kg body weight). Normothermic cardiopulmonary bypass was initiated through a right posterolateral thoracotomy . Through a longitudinal right atriotomy, a continuous 4-0 polypropylene suture (no baffle) was then placed to simulate the lateral tunnel of the Fontan operation, as previously described (Fig. 1) . After closure of the atriotomy, the animals were weaned off cardiopulmonary bypass, and decannulation and closure were performed.
1.2 Holter monitor analysis.
Twenty-four hour Holter recordings were made on the day before the initial operation, on the first postoperative day and 2, 4 and 6 weeks postoperatively. The percent of time spent in sinus rhythm, junctional escape and AFL for each dog on each Holter examination was quantitated. AFL was identified on the Holter recordings as a fixed-cycle length atrial tachycardia that did not require ventricular participation. Holter recordings were searched for episodes of spontaneous initiation of AFL. Six episodes of spontaneous initiation of AFL per Holter recording were examined when applicable.
1.3 Chronic phase operative technique.
Six to 8 weeks postoperatively, reoperation was performed, again using pentobarbital anesthesia. Normothermic cardiopulmonary bypass was instituted as before. Bipolar pacing and sensing electrodes were sutured to the appendages of the right and left atria, respectively. After excision of the tricuspid and mitral valve leaflets, two form-fitting silicone elastomer (Dow Corning Corp.) atrial unipolar endocardial 253-point mapping electrodes were positioned and anchored to each annulus through bilateral ventriculotomies. An indifferent electrode was attached to the chest wall for unipolar reference.
All animals received humane care in compliance with the “Principles of Laboratory Animal Care” formulated by the National Society for Medical Research and the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Science and published by the National Institutes of Health (NIH Publication 86-23, revised 1985). The study protocol was also approved by the Washington University Animal Studies Committee.
1.4 Pacing protocol.
Programmed extrastimulation and burst pacing were performed using a programmable pulse generator (Bloom Inc.). Stimulus input was set at twice pacing threshold. Attempts to induce AFL were made at reoperation after placement of the mapping electrodes. Sustained AFL was defined as the presence of a stable tachycardia >30 s in duration with a fixed atrial cycle length <250 ms. Termination of AFL was achieved either by overdrive pacing or by premature stimulation. Reproducibility was established by reinduction of the tachyarrhythmia using the same extrastimulus pattern that originally induced it. No pharmacologic agents were utilized for the induction of AFL. After pacing, all animals were killed by termination of cardiopulmonary bypass.
1.5 Activation sequence mapping.
At reoperation, atrial activation sequence data during spontaneous rhythm and during all sustained arrhythmias were obtained by simultaneously recording 253 unipolar electrograms from the endocardial multipoint electrodes. The limb lead electrocardiogram (ECG) and a bipolar left atrial electrogram were simultaneously recorded. Local endocardial activation times were determined, and computer-generated activation sequence maps were reconstructed as previously described [9, 10].
1.6 Statistical analysis.
Results are expressed as mean value ± SD. The statistical significance of paired data was determined by the Student paired ttest. Data with three or more groups were compared using repeated-measures analysis of variance; multiple comparisons were made with contrast. Categoric data were compared using chi-square analysis (SYSTAT 5.0, SYSTAT Inc.). A p value <0.05 was considered statistically significant.
2.1 Spontaneous arrhythmias.
Preoperatively, all dogs were in sinus rhythm, and no identifiable atrial ectopic activity was present. After the simulated Fontan operation, all seven dogs experienced spontaneous paroxysms of sustained AFL (Figs. 2 and 3). ⇓⇓Six dogs survived. Dog 7 died on the first postoperative day. Holter recording demonstrated prolonged episodes of AFL in the immediate postoperative period but ceased before the terminal event. A postmortem examination did not reveal any significant pathologic findings. Six weeks postoperatively, of the six dogs that survived, three had intermittent AFL and two had incessant AFL (Table 1). The mean cycle length of spontaneous AFL was 192 ± 22 ms.
Postoperative bradyarrhythmias indicative of sinus node dysfunction were also common. Episodes of junctional escape rhythm of varying degrees of severity were documented in six of seven dogs (Fig. 4). Six weeks postoperatively, four of six dogs were in a junctional escape rhythm 5% to 40% of the time.
2.2 Spontaneous initiation of AFL.
AFL was initiated from both a junctional escape and sinus rhythm (Fig. 5). More frequently, episodes of AFL resulted from an antecedent junctional escape rhythm (69%) than from sinus rhythm (31%, p = 0.02) (Table 2). Overall, dogs were more often in sinus rhythm postoperatively (47%) than in junctional escape rhythm (25%, p < 0.05). The duration of a particular rhythm did not predict the percent of time AFL was initiated from that rhythm (p < 0.001). The spontaneous mean cycle length increased immediately before the onset of AFL from 559 ± 163 to 674 ± 236 ms (p = 0.001).
In certain instances, it appeared that atrial premature depolarizations may have triggered episodes of spontaneous AFL (Fig. 6). Intervening atrial fibrillation before a stable fixed cycle length atrial tachycardia ensued was observed inconsistently in only two of six dogs and in only 12% of all episodes of spontaneous initiation examined (Fig. 7).
2.3 Activation sequence maps during AFL and sinus rhythm.
At reoperation 6 to 8 weeks postoperatively, sustained AFL could be reproducibly induced in six of six dogs, with a mean cycle length of 194 ± 17 ms. The mean cycle lengths of spontaneous (192 ± 22 ms) and induced AFL (194 ± 17 ms) did not differ significantly (p = 0.56) (Table 3).
In all dogs, the AFL circuit was confined to the right atrium. During AFL, atrial activation sequence maps demonstrated a line of conduction block along the lateral portion of the Fontan suture line on the right atrial free wall. The reentrant circuit depended on an isthmus of myocardium bounded superiorly by this line of conduction block and inferiorly by the tricuspid annulus (Fig. 8). In these examples, a clockwise loop of reentry, as viewed from below the tricuspid valve, is demonstrated. The wavefront courses rostrally up the free wall and then crosses onto the septum. After moving caudally down the septum, the impulse propagates along the floor of the right atrium to complete the circuit. Both clockwise and counterclockwise activation patterns were observed.
During sinus rhythm, the earliest point of atrial activation was near the region of the sinus node. The same line of block present during AFL, along the free wall segment of the lateral tunnel suture line, was observed during sinus rhythm (Fig. 9). The impulse breaks across the suture line at its inferior most aspect and propagates up the free wall of the atrium. It reaches the septal surface by traveling around either side of the inferior vena cava and continues uniformly and rapidly along the septum in a caudocranial direction. The septal and free wall wavefronts collide near the right atrial appendage.
3.1 Electrophysiologic substrates for AFL.
The configuration of a reentrant pathway is determined by the anatomic substrate, which may be congenital or acquired [11–14]. In our model of the modified Fontan operation, the substrate is a structural discontinuity in the atrial myocardium imposed by the lateral tunnel suture line combined with the natural orifices of the vena cava. The running suture line produces tissue necrosis , which results in a complete line of conduction block. In our acute model of the Fontan operation and again in the present study, we demonstrated that hemodynamic effects, such as atrial stretch, hypertrophy and hypertension, which have long been regarded as crucial in the etiology of postoperative atrial arrhythmias in patients who have undergone the Fontan operation [4, 8], are not required for the genesis of an anatomic substrate. That the tachycardia cycle lengths and activation sequence patterns observed in the chronic canine model were identical to those seen acutely (177 ± 31 ms) also provides evidence that any postoperative fibrotic alterations within the atrium did not influence the fundamental anatomic substrate of AFL.
Regions of pathologic slow conduction were not identified within the AFL circuit in the present model. Evidence for the existence of discrete areas of slow conduction crucial to sustaining reentry in patients after Fontan repair, analogous to those described for ventricular tachycardia and spontaneous type I human AFL, is equivocal. Fragmented electrograms and pace mapping have been cited as proof of slow conduction in this setting. However, absolute verification of slow conduction requires demonstration of reduced conduction velocity between simultaneously recorded electrograms. Such verification is difficult to accomplish in clinical studies, where the number of simultaneously recorded electrograms and the plane in which they are recorded are limited. Pace mapping may demonstrate concealed entrainment but does not necessarily prove slow conduction. In our studies we also observed low amplitude fragmented signals along suture line borders that may represent damaged myocytes. However, conduction in adjacent healthy myocardium was normal. It is healthy myocardium within the isthmus between the free wall suture line and the tricuspid annulus that is responsible for propagation of the reentrant wavefront, not the adjacent injured or fibrosed regions.
3.2 Spontaneous initiation of AFL.
AFL results from the interaction of an anatomic substrate with multiple dynamic functional components. The combination of a nonuniform refractory field, anisotropic geometry, critically timed and placed premature atrial depolarizations and the particular spatial and temporal relation of these variables govern the presence of repetitive activity [18, 19]. Although hemodynamic variables may be important considerations clinically, the chronic canine model of the Fontan operation saliently illustrates that the spontaneous initiation of AFL is not dependent on them.
3.2.1 Triggering event.
The occurrence of atrial premature beats appeared to be common in the postoperative period. Preoperatively, no atrial ectopic beats were observed. The likelihood of spontaneous initiation of AFL increases proportionally to the number of atrial premature beats. Clinically, it has been discovered that in patients who have undergone a Fontan repair, the incidence of supraventricular extrasystoles increases with time from repair.
It is possible that those beats that we classified as premature may have in fact been echo beats. AFL may have been precipitated by functional block without premature atrial depolarizations. It is difficult to definitively differentiate these mechanisms solely on the basis of Holter monitor recordings. However, experimentally, reentrant arrhythmias are rarely initiated without premature depolarizations . In addition, acutely and chronically, we never successfully induced AFL by pacing at long cycle lengths. Finally, functional block, demonstrated to initiate AFL in our acute-phase experiments , predominantly occurs at short, premature intervals.
Others have found that AFL develops from an intervening transitional rhythm, usually atrial fibrillation, in which critical areas of slow conduction and unidirectional block occur. In the present study, transitional atrial activity, resembling atrial fibrillation, was not a prominent finding. Perhaps because the line of block defining the pathway is anatomic, it became statistically more likely that an atrial premature beat would initiate reentry. The fixed line of conduction block requires the premature beat to encounter a much smaller region of functional block, which creates the transient unidirectional block necessary for initiating AFL.
3.2.2 Sinus node dysfunction.
There is a high degree of correlation between postoperative AFL and sinus node dysfunction in patients after a Fontan operation . Furthermore, pathologic abnormalities in the sinus node have been linked to supraventricular arrhythmias, where a loss of sinoatrial node fibers is a consistent histologic finding [24, 25]. That sinus node dysfunction was present postoperatively in our study is evidenced by the frequent occurrence of a junctional escape rhythm. The free wall aspect of the lateral tunnel suture line, which lies along the crista terminalis, is in close proximity to the sinus node, increasing the likelihood of direct injury to it [15, 26]. In the current study, AFL was significantly more likely to be initiated from an antecedent junctional escape than sinus rhythm (Table 2). When sinus node dysfunction is present, atrial depolarization is governed by subsidiary pacemakers. We speculate that these latent pacemakers provide multiple foci of atrial stimulation, increasing the probability of initiation of reentry.
3.2.3 Autonomic factors.
Alterations in autonomic tone are prominent in the canine model, which may explain the consistent increase in the spontaneous cycle length immediately before the onset of AFL. Cholinergic suppression has previously been manifested to provoke atrial premature beats and reentry . In addition, vagally mediated slowing of the heart rate may nonuniformly alter the effective refractory period of the atrium, making the tissue more vulnerable to reentry .
3.3 Significance as a model of spontaneous AFL.
Spontaneous AFL is a unique characteristic of the present model. Although many chronic animal models of atrial reentry have been investigated in the past, to our knowledge, none has had the feature of consistent spontaneous AFL. This feature makes our model well suited for studying numerous electrophysiologic and clinical aspects of this reentrant tachycardia. The paroxysmal nature of AFL, well illustrated in the present study, is related to the inherently labile nature of many functional components. Abnormal impulse formation and dispersion of refractoriness can be influenced by a variety of secondary factors. Some of these factors, including alterations in autonomic tone, are more pronounced in the dog; others, including structural changes from atrial enlargement and stretch [29–31], were not present but are extant in patients. Alteration of these functional components under experimental conditions, for example, by interrupting vagal innervation or by inducing right atrial enlargement, may permit delineation of the mechanisms involved in spontaneous initiation.
3.4 Clinical significance and implications.
Clinical intracardiac studies thus far have been unable to elucidate details of the reentrant circuit after the modified Fontan operation. Clinical investigators are often hindered by the absence of true sequential atrial activation sequence mapping with an electrode density sufficient to clearly delineate areas of conduction disturbance. Accurate identification of anatomic barriers is also imprecise. Although animal models of AFL permit a detailed understanding of reentrant circuits, to a large extent previous models have only been of academic interest because of the artificial nature of the barriers utilized. The line of block in this model is the unintentional consequence of a clinically relevant suture line.
The present chronic canine model provides strong evidence of the actual AFL circuit that occurs clinically after the modified Fontan operation. The spontaneous occurrence of AFL eliminates the possibility that the AFL induced acutely was an artifact of the stimulation technique. The consistency of the activation sequence patterns observed acutely and chronically establishes the constant nature of the reentrant circuit over time. Furthermore, the striking similarities in cycle lengths of induced and spontaneous AFL affirms that the spontaneous tachycardia was the same as that induced by premature stimulation and validates the endocardially mapped reentrant circuits. Therefore, we believe that efforts directed toward definitive treatment and prevention of AFL after the modified Fontan operation require consideration of the reentrant circuit described in the present study.
3.5 Limitations of the study.
The animal model used in this study had normal cardiac anatomy. The anatomic and physiologic influences of complex congenital cardiac malformations, which may have influenced both arrhythmia formation and the configuration of the reentrant circuit, were therefore not present. In addition, because spontaneous AFL was never mapped, we cannot be absolutely certain that the reentrant circuit was the same as that induced with programmed extrastimulation.
In a chronic canine model of the modified Fontan operation, the lateral tunnel suture line alone, in the absence of atrial stretch or hypertension, provides an electrophysiologic substrate that promotes spontaneous AFL. This model may be useful for evaluating various forms of treatment and prevention of AFL after this type of Fontan repair.
We thank the technical staff of our cardiac research laboratory. We also thank Joe Loslo for help with the Holter analysis and the Surgical Illustrations Department of Washington University for assistance in preparing the figures.
☆ This study was supported by Grants HL 32257 and HL 33722 from the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland.
- atrial flutter
- Received December 12, 1996.
- Revision received April 18, 1997.
- Accepted July 1, 1997.
- The American College of Cardiology
- Driscoll DJ,
- Offord KP,
- Feldt RH,
- Schaff HV,
- Puga FJ,
- Danielson GK
- Balaji S,
- Gewillig M,
- Bull C,
- de Leval MR,
- Deanfield JE
- Cecchin F, Johnsrude CL, Perry JC, Friedman RA. Effect of age and surgical technique on symptomatic arrhythmias after the Fontan procedure. Am J Cardiol 1195;76:386–91.
- Balaji S,
- Johnson TB,
- Sade RM,
- Case CL,
- Gillette PC
- Branham BH,
- Bi X,
- Cox JL
- Frame LH,
- Page RL,
- Hoffman BF
- ↵Bromberg BI, Rodefeld MD, Huddleston CB, Schuessler RB, Saffitz JE, Boineau JP. Sinus node dysfunction following the modified Fontan: evidence for direct injury in a canine model [abstract]. PACE 1995;18(P II):876.
- Triedman JK,
- Saul JP,
- Weindling SN,
- Walsh EP
- Van Hare GF,
- Lesh MD,
- Stanger P
- Spach MS,
- Miller WT,
- Dolber PC,
- Kootsey JM,
- Sommer JR,
- Mosher CE Jr.
- Spach MS,
- Dolber PC,
- Heidlage JF
- Gewillig M,
- Wyse RK,
- de Leval MR,
- Deanfield JE
- Schuessler RB,
- Boineau JP,
- Bromberg BI,
- Hand DE,
- Yamauchi S,
- Cox JL
- Shimizu A,
- Nokaki A,
- Rudy Y,
- Waldo AL
- Kürer CC,
- Tanner CS,
- Vetter VL
- Sims BA
- Euler DE,
- Jones SB,
- Gunnar WP,
- Loeb JM,
- Murdock DK,
- Randall WC
- Kavey RW,
- Gaum WE,
- Byrum CT,
- Smith FC,
- Kveselis DA
- Schuessler RB,
- Rosenshtraukh L,
- Boineau JP,
- Bromberg BI,
- Cox JL
- Allessi R,
- Nusynowitz M,
- Abildskov JA,
- Moe GK
- Boyden PA
- Boyden PA,
- Hoffman BF
- Schoels W,
- Kuebler E,
- Yang H,
- Gough WB,
- El-Sherif N