Functional Mechanisms Underlying Tachycardia-Induced Sustained Atrial Fibrillation in a Chronic Dog Model
Background Rapid atrial activation causes electrical remodeling that promotes atrial fibrillation (AF), but underlying mechanisms are incompletely understood. We applied epicardial mapping to evaluate atrial electrophysiology and AF duration in dogs subjected to rapid atrial pacing (400/min).
Methods and Results Dogs paced for 1 (P1, n=7), 7 (P7, n=13), or 42 (P42, n=7) days were compared with sham dogs (P0, n=13). Atrial pacing progressively increased AF duration. Atrial effective refractory period (ERP) and ERP accommodation to rate were significantly decreased by pacing, with near-maximal changes within 7 days. Atrial conduction velocity decreased more slowly, with maximum changes at 42 days, contributing to increases in AF duration after ERP stabilized. Stepwise multilinear regression indicated that both wavelength (P=.02) and duration of pacing (P=.0001) were independent determinants of changes in AF duration. Mean atrial fibrillation cycle length (AFCL) at 112 recording sites decreased with increased duration of rapid pacing (P<.001), and the SD of AFCL increased progressively (P<.0001), together accounting for 72% of the variance in AF duration. Increases in AFCL variability were due to regionally determined differences in AFCL changes caused by rapid pacing. The number of zones of reactivation per cycle of AF increased as AF became more sustained, consistent with multiple-wavelet reentry.
Conclusions Rapid atrial activation causes time-dependent decreases in ERP, conduction velocity, and wavelength, which, along with increased regional heterogeneity, provide a substrate for AF. The conduction abnormalities and increased regional heterogeneity previously noted in patients with AF may be a consequence, as well as a cause, of the tachyarrhythmia.
Atrial fibrillation is the most frequently encountered arrhythmia in clinical practice and is likely to become more common with the aging of the population.1–3 Major limitations remain in our knowledge of the pathological and electrophysiological mechanisms underlying AF. Experimental AF has been studied in normal animal hearts,4–9 with or without acute interventions like topical aconitine4 or vagal nerve stimulation7–9 to promote AF maintenance. Most clinical AF occurs in patients with chronic atrial pathology, which is absent in the short-term AF models described above, limiting their applicability to clinical AF.
Over the past few years, investigators have developed new animal models of AF associated with atrial dilation and ultrastructural changes that resemble those of clinical AF.10,11 These models generally involve rapid atrial activation, induced by either rapid 1:1 atrial pacing11 or electrically maintained AF,10,12 mimicking the “domestication of AF,” a process by which AF causes electrophysiological remodeling that favors its own maintenance (“AF begets AF”).12
Limited information is available on the mechanisms of AF in the rapid activation models. Whereas Wijffels et al12 showed that electrically maintained AF results in a decrease in atrial ERP and in ERP adaptation to rate change,12 ERP changes were near-maximal within 24 hours, while the duration of AF increased more slowly to approach steady-state values over 7 days.12 This implies that additional factors are altered to promote the maintenance of AF. Morillo et al11 showed that 1:1 atrial pacing at 400/min for 6 weeks led to reductions in atrial ERP and to sustained AF in about 85% of dogs, but they did not study the time course of electrophysiological changes or directly measure conduction velocity.
The present study was designed to assess the electrophysiological changes that provide a substrate for sustained AF among dogs subjected to atrial pacing at 400/min for periods of up to 6 weeks. We wished to determine (1) whether the ERP changes produced by rapid 1:1 atrial pacing resemble those previously described during electrically maintained AF, (2) whether atrial conduction is altered, (3) whether the time course of changes in electrophysiological variables can account for the time course of changes in the duration of AF, and (4) whether epicardial mapping of atrial activation during AF can provide clues about the mechanisms of the arrhythmia and the role of underlying electrophysiological alterations.
Pacemaker Insertion and Atrial Pacing
Adult mongrel dogs (weight, 28.2±2.0 kg; n=40) were anesthetized with sodium pentobarbital (30 mg/kg IV, additional doses of 4 mg/kg as needed). Respiration was maintained via an endotracheal tube and a mechanical ventilator. Under sterile conditions, a unipolar screw-in pacing lead (Medtronic Inc) was inserted via the right external jugular vein and fixed in the right atrial appendage under fluoroscopic guidance. The lead was then connected to a Medtronic pacemaker unit (model 8084) in a subcutaneous pocket in the neck.
Twenty-four hours later, the pacemaker was programmed to capture the atrium at 400/min with 4-ms pulses at twice-threshold current. The atrium was stimulated at this rate for a period of 1 (group designated “P1,” n=7), 7 (“P7,” n=13), or 42 (“P42,” n=7) days. The surface ECG was verified after 24 hours and then weekly. Rapidly paced dogs were compared with sham dogs (“P0,” n=13), similarly instrumented but maintained without pacemaker activation for 1 (n=5), 7 (n=5), or 42 (n=3) days. The results of P0 dogs were the same regardless of observation period, so they were grouped together for all analyses.
On study days, dogs were reanesthetized with morphine (2 mg/kg SC) and α-chloralose (120 mg/kg IV bolus, continuous infusion of 29.25 mg · kg−1 · h−1) and ventilated with oxygen-enriched air. Respiratory parameters were adjusted to maintain physiological arterial blood gases. Body temperature was maintained (37°C) with a circulating-water system. Polyethylene catheters were inserted into the left femoral artery and both femoral veins. A median sternotomy was performed, and a pericardial cradle was created. Two bipolar Teflon-coated stainless steel electrodes were inserted into the right atrial appendage for recording and stimulation. A programmable stimulator (Digital Cardiovascular Instruments) was used to deliver 2-ms pulses at twice-threshold current. A P23 1D transducer (Statham Medical Instruments), electrophysiological amplifiers (Bloom Ltd), and a paper recorder (Astromed MT-95000) were used to record blood pressure, surface ECG leads, a right atrial electrogram, and stimulus artifacts.
In paced dogs, the implanted pacemaker was deactivated. Activation maps for CV measurements were obtained after 2 minutes of pacing at the right atrial appendage at each of four BCLs from 150 to 400 ms. The ERP was measured with 15 basic (S1) stimuli followed by a premature (S2) stimulus at an S1S2 interval that was decreased by 10-ms decrements from the BCL, with the ERP defined as the longest S1S2 interval failing to produce a response. The ERP was determined twice at each BCL, and the mean of the ERP values was used for data analysis. If AF occurred during ERP testing, it was cardioverted, and the dog was allowed to rest for 30 minutes. In such cases, incremental S1S2 from below the ERP were used to obtain an initial ERP estimate and minimize the number of S1S2 trials needed.
AF was then induced with 10-Hz, 2-ms stimuli at four times the threshold current. AF was defined as a rapid (>450/min), irregular atrial rhythm with varying atrial electrogram morphology. AF lasting >45 minutes was considered sustained. An 8-second window of activation data was acquired during AF to analyze activation patterns. To estimate mean AF duration, AF was induced 15 times for AF duration <10 minutes and twice for AF duration between 10 and 45 minutes. If AF lasted >45 minutes, no further AF induction was attempted to avoid excessive prolongation of the experiment. AF lasting >45 minutes was terminated by direct current electrical cardioversion. AF inducibility was also assessed with single premature S2 stimuli at a 400-ms BCL. AF was considered to be inducible if induced reproducibly at a given coupling interval.
Five thin silicon plaques containing 112-bipolar electrodes with 1-mm interpolar and 6-mm interelectrode distances were sewn into position to cover the atrial epicardial surface as previously described.7–9 Each electrogram signal was filtered (30 to 400 Hz), digitized (12-bit resolution and 1-kHz sampling rate), and transmitted into an IBM-compatible microcomputer. Software routines were used to amplify, display, and analyze each electrogram signal and to generate activation maps.7–9 Each electrogram was analyzed with computer-determined peak-amplitude criteria and was reviewed manually. Electrogram timing was compared to QRS complexes to exclude ventricular electrograms.
CV was determined by analyzing activation at four electrode sites in the direction of rapid propagation (perpendicular to consecutive isochrones) in the right atrial free wall. Distance from the proximal site was plotted against activation time, and CV was determined from the slope of the best-fit regression line. Activation maps were reviewed to ensure continuous longitudinal propagation, and only data with correlation coefficients >.99 were accepted for analysis. The same sites were used for CV measurements for each experiment. Fig 1⇓ illustrates CV measurements under P0, P1, P7, and P42 conditions. Two other indexes of conduction speed were analyzed. First, CV along Bachmann’s bundle was measured in a fashion similar to that used to calculate CV in the right atrial free wall. Second, total conduction time was assessed by subtracting the mean earliest activation time (average of the three earliest activation times) from the mean latest activation time (average of the three latest atrial activation times during the atrial complex). The wavelength for reentry was calculated as the product of conduction velocity and ERP.13
AFCL was calculated at each recording site by calculating the mean of 10 consecutive activation intervals. The results at all sites were averaged to obtain the overall mean AFCL under each condition. Regional variability of AFCL was assessed two ways. First, overall intersite variability was evaluated by calculating the SD in AFCL at all 112 recording sites. Second, regional differences were assessed by calculating mean AFCL at all sites within each of six zones: the left and right atrial appendage, Bachmann’s bundle in the left and right atria, and right and left atrial free walls.
The activation pattern during AF was studied by constructing sequential atrial activation maps. Zones of reactivation were defined as zones activated early in one cycle that were reactivated at the beginning of the next cycle. The number of reactivation zones was determined for each of three successive AF cycles in each dog by an individual blinded to study group.
Statistical comparisons of multiple group means were obtained by ANOVA with Dunnett’s test. The Kruskal-Wallis test was used for nonparametric comparisons of unpaired measures. Stepwise multilinear regression was used to assess the dependence of a single dependent variable on multiple independent variables, and linear regression was used to analyze single dependent and independent variables. Average results are given as mean±SEM, and a two-tailed value of P<.05 was considered statistically significant.
Changes in AF Duration and Electrophysiological Properties
The duration of AF induced by burst pacing increased progressively with increased duration of rapid atrial stimulation (the Table⇓), increasing from 7.4±1.8 seconds (P0) to 2486±497 seconds after 6 weeks of rapid pacing (P42, P<.0001). The number of dogs in which sustained AF was induced increased progressively, from none under control conditions to 86% of P42 dogs. Pacing decreased ERP, CV, wavelength for reentry, and rate-dependent ERP accommodation.
Fig 2⇓ shows the BCL-dependent changes in atrial ERP, CV, and wavelength produced by rapid atrial pacing. Under control conditions, ERP decreased in response to increased rate, showing typical rate-dependent accommodation (Fig 2⇓, top). Mean ERP values (particularly at longer cycle lengths) and ERP accommodation to rate were strongly diminished by rapid pacing, with decreases in accommodation apparent within 1 day of the onset of pacing and near maximal at 7 days. Rapid stimulation also reduced atrial CV; however, the changes in CV were slower to develop than those in ERP, with relatively small changes after 1 and 7 days of pacing and much more pronounced alterations at 42 days (Fig 2⇓, middle). Alterations in wavelength reflect the time course of changes in both ERP and CV, with a progressive reduction that becomes significant at longer cycle lengths after 7 days and at all cycle lengths at 42 days (Fig 2⇓, bottom).
To confirm the effects of rapid pacing on atrial CV, we obtained two other measures reflecting conduction speed (Fig 3⇓). Pacing led to highly significant decreases (P<.01) in CV over Bachmann’s bundle and increases (P<.01) in atrial conduction time under P42 conditions (Fig 3⇓). These changes were highly correlated with right atrial CV (r2=.94, P<.05 for correlation between CV in Bachmann’s bundle and right atrial free wall; r2=.99, P<.01 for right atrial CV versus conduction time).
AF occurred spontaneously within 1 week in 3 of 13 P7 dogs (23%) and 2 of 7 P42 dogs (29%); by the end of the 6-week observation period, spontaneous and persistent AF had occurred in 4 of 7 P42 dogs (67%). When AF occurred spontaneously, the dog was not cardioverted and was left in the arrhythmia until the scheduled study day. There were no significant differences at electrophysiological study between dogs without spontaneous AF and those with AF: at a cycle length of 150 ms, ERP averaged 87±4 and 76±7 ms, CV averaged 92±4 and 86±7 cm/s, and wavelength was 7.8±0.5 and 7.0±0.6 cm in dogs without and with AF, respectively.
Relation Between Electrophysiological Changes and AF Duration
Fig 4⇓ shows the results of an analysis of the relationship between AF duration and ERP, CV, and wavelength. Individual symbols show results for each dog, and symbols with error bars are the mean±SEM for each group. The correlation was weak between individual values of ERP and AF duration (r2=.136, P=.019), stronger for AF duration as a function of CV (r2=.211, P=.003), and strongest for AF duration as a function of wavelength (r2=.277, P=.0005, n=40 for each comparison). When ERP, CV, wavelength, and duration of pacing were included as covariates in stepwise multilinear regression, the correlation was best explained by a model including wavelength (P=.018) and duration of pacing (P=.0001). These results suggest that both ERP and CV changes contribute to the progressive tendency to maintain AF, that alterations in wavelength appear to reflect well their combined effects, and that pacing duration affects AF duration independently of changes in ERP, CV, and wavelength.
Changes in Atrial Activation During AF
Fig 5⇓ shows an analysis of mean AFCL as a function of pacing interval. AFCL decreased progressively with increased duration of rapid pacing, with significant changes after 7 and 42 days (P<.01 for each). The SD of AFCL (an index of AFCL variability) increased progressively during rapid pacing (Fig 6⇓, left). When the SD of AFCL was expressed as a ratio of mean AFCL in each dog, to determine the variation in AF interval relative to the interval itself, the increases in variability became even more apparent (Fig 6⇓, right). We performed stepwise multilinear regression of AF duration on the duration of rapid pacing, the mean AFCL, and the SD of AFCL in each dog (n=40). Both AFCL and SD of AFCL were highly significant determinants of AF duration (P<.0001 for each), and the model including them accounted for 72% of the variance in AF duration (R2=.721). The duration of rapid pacing did not provide significant information in predicting AF duration independently of mean AFCL and SD of AFCL.
The increase in AFCL variability caused by rapid pacing could be due to a random or highly localized (eg, site by site) increase in variability or could be organized at a regional level. To address the latter possibility, we analyzed the mean AFCL in six different atrial regions, as illustrated in Fig 7⇓. AFCL was first determined at multiple electrode sites in the left atrial appendage, the right atrial appendage, the left and right sides of Bachmann’s bundle, and the right and left atrial free walls. Data from a total of four to nine (mean, seven) electrode sites were available for analysis in each region. Values of AFCL at individual sites are shown for representative dogs in each group in Fig 7⇓. The AF interval decreased in dogs subjected to longer durations of rapid pacing and appeared to decrease to a greater extent in some zones, particularly in the left atrium, than in others. Fig 8⇓ shows the mean (±SEM) AFCL in each of the six regions for each group. In sham dogs, there were no statistically significant regional differences in mean AFCL. After 7 and 42 days of rapid pacing, mean AFCL decreased from control values in all regions, but decreases were larger in some regions than in others, resulting in highly significant regional variability in AFCL. These observations suggest that a substantial proportion of the increased variability in AFCL resulting from sustained rapid pacing is due to regionally determined differences in electrical remodeling.
The final set of analyses addressed the pattern of epicardial activation during AF. Fig 9⇓ shows activation data during three consecutive cycles of AF in a P0 dog (top) and a P42 dog (bottom). In the P0 dog, the first cycle (top left) shows one zone of early activation with a figure-eight pattern of propagation. Electrograms from six sites show consecutive activation to site f. Site a (asterisk) was reactivated from the region of site f (dashed arrow), initiating the next cycle (top middle). The overall activation pattern of this cycle was similar to the preceding one, resulting in initiation of the next cycle by reactivation near site a. In all P0 dogs, AF was characterized by one or two sites of early activation per cycle and a fairly organized pattern of activation. In paced dogs, activation became progressively more complex. For the P42 dog shown in Fig 9⇓ (bottom), the first cycle had early-activating zones at the middle of Bachmann’s bundle, the lateral left atrium, and the medial left atrium. Reactivation of these zones (asterisks, dashed arrows) initiated the next cycle (bottom middle). Three early-activating zones in the middle cycle were also reactivated (dashed arrows) to initiate the third cycle (bottom right). Pacing dogs with sustained AF were characterized by multiple spatially distinct regions initiating each cycle and complex activation patterns. To quantify these activation changes, we calculated the mean number of zones of reactivation per cycle during AF in each dog. For example, the P0 dog illustrated in Fig 9⇓ showed one reactivation zone between cycles 1 and 2 and one between cycles 2 and 3 (asterisks in top left and middle), whereas the P42 dog had three reactivation zones for each cycle. As shown in Fig 10⇓, rapid pacing caused a progressive increase in the number of reactivation zones per cycle of AF.
In the present study, we evaluated the time course of changes in atrial electrophysiology and activation during AF in dogs subjected to rapid atrial pacing. The results suggest that chronic atrial tachycardia causes progressive decreases in atrial ERP, in ERP adaptation to rate and in CV, along with progressive increases in the ability of single atrial extrasystoles to induce AF and in the duration of AF induced by burst pacing. The time course of changes in CV is slower than that of ERP alterations, contributing to increases in AF susceptibility after ERP changes are maximal. Activation studies indicate regional variability in the extent of electrophysiological remodeling (as reflected by the local AFCL), and point toward increases in the number of functional reentry zones during AF as the latter becomes more sustained.
Electrophysiological Changes Related to AF
The present studies provide the first direct evidence (to our knowledge) that chronic rapid atrial activation may lead to abnormalities in atrial conduction. Increases in CV that occurred after ERP changes stabilized likely contributed to delayed changes in AF duration in our dogs and could account for the discrepancies between the time courses of ERP changes and AF duration in the work of Wijffels et al12 (although preliminary reports from the same laboratory argue against conduction slowing in the goat model14). Abnormalities of intra-atrial conduction are known to be associated with clinical AF,15–17 and it has long been known that patients with AF have prolonged P waves when in sinus rhythm. Conduction abnormalities favor AF by reducing the wavelength for reentry5 and are usually considered to be secondary to underlying atrial disease. Our findings raise the interesting possibility that atrial conduction abnormalities can be due to the arrhythmia itself, with AF being both a potential cause and a consequence of atrial conduction abnormalities. Tachycardia-induced conduction slowing may well have contributed to the increases in P-wave duration noted in rapidly paced dogs by Morillo et al11 and more recently by Elvan et al.18
We found evidence for a role of regional differences in electrical remodeling in refractory properties during AF as reflected by local values of the AFCL. The most important changes occurred in the inferoposterior left atrial free wall, implying that regionally variable remodeling may contribute to the reduced left atrial refractoriness suggested by a variety of observations. In patients with AF, complex atrial activity is found during AF more often in the left compared with right atrium.19 Morillo et al11 found that in dogs subjected to 6 weeks of rapid atrial pacing, the inferoposterior left atrium shows particularly rapid activation during AF and is important in arrhythmia perpetuation. Li et al20 also found that rapidly paced dogs have shorter refractory periods in the left atrium, differences not observed in dogs with chronic pericarditis.20,21
The statistical importance of AFCL variability as an independent determinant of AF duration supports the importance of remodeling-induced electrical heterogeneity in the substrate for AF. This finding agrees with published observations supporting a role for atrial electrical heterogeneity in experimental AF.22–24 Heterogeneity in atrial repolarization also appears to be prominent in clinical AF25,26 and is associated with increased variability in AFCL at multiple atrial recording sites.27 Initial analyses did not provide evidence of increased refractoriness heterogeneity in the goat AF model of Wijjffels et al,12 but recently presented data from additional studies point toward increases in regional heterogeneity of refractoriness.28
Increased atrial vulnerability in humans is associated with a reduced atrial ERP adaptation to rate change,29 which resembles the ERP changes in our dogs. We have previously shown that atrial ERP is a major determinant of the ability of single extrasystoles to induce AF, with short local refractory periods facilitating AF induction.30 A loss of ERP adaptation to rate resulted in substantial ERP abbreviation at longer cycle lengths in both our dogs and previous studies,12,29 a factor that may be an important contributor to vulnerability to AF induction by premature beats.
Although multiple wavelet reentry is believed to underlie tachycardia-dependent AF,12 direct evidence from epicardial mapping studies has been lacking. In the present study, we observed changes in activation during AF compatible with multiple wavelet reentry and with an increased number of reentrant wave fronts stabilizing AF in rapidly paced dogs.
Possible Mechanisms Underlying Electrophysiological Changes
The ionic mechanism of ERP alterations in patients with AF is poorly understood. We have found that ERP alterations in dogs subjected to rapid atrial pacing are probably due to alterations in action potential duration caused by decreases in l-type Ca2+ current.31 Action potential duration changes in patients with AF25,26 are similar to those we have noted in the present model,31 and l-type Ca2+ current is reduced in atrial cells from patients with atrial dilation.32,33 We have also obtained data that suggest that atrial myocytes of rapidly paced dogs have reduced INa, possibly accounting for conduction changes.34 There is evidence for reduced connexin 40 expression in goats with AF-induced remodeling,14 suggesting a potential role for intercellular coupling changes in altering conduction.
Potential Clinical Relevance
The ability of AF to alter atrial electrophysiology means that in patients with AF, the atrial substrate contains two elements: the factors that permit the initial occurrence of AF and the changes caused by AF per se. The latter component will vary in importance in any given patient, depending on the duration and incidence of AF episodes. The present study suggests that increased regional heterogeneity in AFCL and the occurrence of conduction slowing, common in patients with AF,27,35 may be caused by rapid atrial activation during AF. The AF-promoting effect of rapid pacing may also be relevant to the known predilection to AF of patients with the Wolff-Parkinson-White syndrome in whom sustained atrial tachycardias may create a substrate that can support AF.36 To date, the therapy of AF has targeted the electrophysiological properties promoting maintenance of the arrhythmia; however, with improved knowledge of the mechanisms underlying development of the substrate for AF, it may become possible to target the substrate directly.
Limitations of Our Findings
The mapping system we used has a variety of limitations that must be considered. The number of electrograms available (maximum, 112) limits spatial resolution. Much better resolution can be obtained with visual imaging systems based on voltage-sensitive dyes.37 Dye-based techniques also have drawbacks, including possible toxic effects of voltage-sensitive dyes and chemicals needed to arrest cardiac contractility, a need for ex vivo perfusion with crystalloid solutions, and a limited field of vision. Possibly because of the limitations of our system, we were unable to define complete reentry circuits at the origin of many activation wave fronts during AF. On the other hand, we were able to identify clear changes in the pattern of activation during AF as a function of the period of atrial pacing. These changes are consistent with an increased number of reentry circuits, in keeping with the observed decreases in the wavelength for atrial reentry, and support the concept that an increased number of simultaneous circuits made possible by decreases in the reentrant wavelength stabilize AF in dogs exposed to chronic atrial tachycardia. Our observations are compatible with those of Gray et al,37 who applied optical mapping with voltage-sensitive dyes to record transmembrane potentials from 20 000 right atrial sites during AF in Langendorff-perfused sheep hearts. These workers observed incomplete reentry circuits, breakthrough patterns, and wave-front collision on the epicardial surface, in agreement with earlier in vitro studies suggesting that reentry can occur in a three-dimensional fashion and that the right atrium cannot be treated fully as a two-dimensional structure.38
We measured CV in the direction of rapid atrial propagation. Fiber orientation is an important determinant of CV in cardiac tissue, with lower junctional resistance in the longitudinal direction (parallel to the long axis of fibers) resulting in much more rapid longitudinal conduction compared with conduction transverse to fiber orientation.39,40 This property (referred to as tissue anisotropy) is particularly important in the atria and can vary at the microcospic level.40 Atrial geometry is complex, and issues of anisotropy, transmural conduction, and structural determinants (like the pectinate muscles and crista terminalis) are undoubtedly relevant to mechanisms underlying AF.37,39,40 These issues, while of great importance, are beyond the scope of the present study.
We considered AFCL to be an index of atrial refractoriness during AF in our AFCL heterogeneity studies. Previous studies have shown a good correlation between AFCL and ERP and have used AFCL as an index of refractory period during AF.27 However, refractory period may not be the only determinant of AFCL, and because of excitable gaps during AF, the true refractory period is likely to be less than the AFCL. Recent in vitro studies suggest that the minimum AF interval correlates well with local ERP and that the latter is less than the mean AFCL.41
Selected Abbreviations and Acronyms
|AFCL||=||AF cycle length|
|BCL||=||basic cycle length|
|ERP||=||effective refractory period|
This work was supported by grants from the Medical Research Council of Canada, the Quebec Heart Foundation, and the Fonds de Recherche de l’Institut de Cardiologie de Montréal. Dr Gaspo is a research fellow from the Medical Research Council of Canada. Dr Bosch is a holder of a fellowship from the Deutsche Forschungsgemeinschaft. The authors wish to thank Emma De Blasio and Mirie Levi for their skilled technical assistance.
- Received June 10, 1997.
- Revision received August 14, 1997.
- Accepted August 27, 1997.
- Copyright © 1997 by American Heart Association
Kannel WB, Wolf PA. Epidemiology of atrial fibrillation. In: Falk RH, Podrid PJ, eds. Atrial Fibrillation: Mechanisms and Management. New York, NY: Raven Press Publishers; 1992:81-92.
Scher D. Studies on auricular tachycardia caused by aconitine administration. Proc Soc Exp Biol Med. 1947;64:223-239.
Rensma PL, Allessie MA, Lammers WJEP, Bonke FIM, Schalij MJ. Length of excitation wave and susceptibility to reentrant atrial arrhythmias in normal conscious dogs. Circ Res. 1988;62:395-410.
Wang Z, Pagé P, Nattel S. The mechanism of flecainide’s antiarrhythmic action in experimental atrial fibrillation. Circ Res. 1992;71:271–287.
Wang J, Bourne GW, Wang Z, Villemaire C, Talajic M, Nattel S. Comparative mechanisms of antiarrhythmic drug action in experimental atrial fibrillation: importance of use-dependent effects on refractoriness. Circulation. 1993;88:1030-1044.
Wang J, Feng J, Nattel S. Class III antiarrhythmic drug action in experimental atrial fibrillation: differences in reverse use dependence and effectiveness between d-sotalol and the new antiarrhythmic drug ambasilide. Circulation. 1994;90:2032-2040.
Salmon DR, McPherson DD, Augustine DE, Holida MD, White CW. A canine model of chronic atrial fibrillation: echocardiographic and electrocardiographic validation. Circulation. 1985;72(suppl III):III-250. Abstract.
Morillo CA, Klein GJ, Jones DL, Guiraudon CM. Chronic rapid atrial pacing: structural, functional, and electrophysiological characteristics of a new model of sustained atrial fibrillation. Circulation. 1995;91:1588-1595.
Wijffels MC, Kirchhof CJ, Dorland R, Allessie MA. Atrial fibrillation begets atrial fibrillation: a study in awake chronically instrumented goats. Circulation. 1995;92:1954-1968.
van der Velden HMW, van Zijverden M, van Kempen MJA, Wijffels MCEF, Groenewegen WA, Allessie MA, Jongsma HJ. Abnormal expression of the gap junction protein connexin 40 during chronic atrial fibrillation in the goat. Circulation. 1996;94(suppl I):I-593. Abstract.
Kumagai K, Akimitsu S, Kawahira K, Kawanami F, Yamanouchi Y, Hiroki T, Arakawa K. Electrophysiological properties in chronic lone atrial fibrillation. Circulation. 1991;84:1662-1668.
Elvan A, Wylie K, Zipes DP. Pacing-induced chronic atrial fibrillation impairs sinus node function in dogs: electrophysiological remodeling. Circulation. 1996;94:2953-2960.
Jaïs P, Haïssaguerre M, Shah DC, Chouairi S, Clémenty J. Regional disparities of endocardial atrial activation in paroxysmal atrial fibrillation. Pace Pacing Clin Electrophysiol. 1996;19(pt II):1998-2003.
Ortiz J, Niwano S, Abe H, Rudy Y, Johnson NJ, Waldo AL. Mapping the conversion of atrial flutter to atrial fibrillation and atrial fibrillation to atrial flutter: insights into mechanisms. Circ Res. 1994;74:882-894.
Liu L, Nattel S. Sympathetic stimulation decreases wavelength for atrial reentry but fails to promote atrial fibrillation in dogs: importance of heterogeneity in refractoriness. Pace Pacing Clin Electrophysiol. 1996;19(pt II):643. Abstract.
Le Heuzy J, Boutjdir M, Gagey S, Lavergne T, Guize L. Cellular aspects of atrial vulnerability. In: Attuel P, Olsson SB, Schlepper M, ed. The Atrium in Health and Disease. Mount Kisco, NY: Futura Publishing; 1989:81-94.
Tieleman RG, De Langen CDJ, Grandjean JG, Van Gelder IC, De Kam PJ, Wijffels MCEF, Crijns HJGM. Regional differences in pacing induced electrical remodeling of the atrium and recovery from electrical remodeling induces increased dispersion of refractoriness. Circulation. 1996;94(suppl I):I-352. Abstract.
Wang J, Liu L, Feng J, Nattel S. Regional and functional factors determining induction and maintenance of atrial fibrillation in dogs. Am J Physiol. 1996;271 (Heart Circ Physiol 40):H148-H158.
Yue L, Feng J, Gaspo R, Li G-R, Nattel S. Ionic remodeling underlying tachycardia-induced atrial fibrillation in dogs. Circulation. 1996;94(suppl I):I-592. Abstract.
Gaspo R, Bosch RF, Bou-Abboud E, Nattel S. Tachycardia- induced changes in sodium current in a chronic dog model of atrial fibrillation. Circ Res. In press.
Sung RJ, Castellanos A, Mallon SM, Bloom MG, Gelband H, Myerburg RJ. Mechanism of spontaneous alternation between reciprocating tachycardia and atrial flutter-fibrillation in the Wolff-Parkinson-White syndrome. Circulation. 1977;56:409-416.
Gray RA, Pertsov AM, Jalife J. Incomplete reentry and epicardial breakthrough patterns during atrial fibrillation in the sheep heart. Circulation. 1996;94:2649-2661.
Schuessler RB, Kawamoto T, Hand DE, Mitsuno M, Bromberg BI, Cox JL, Boineau JP. Simultaneous epicardial and endocardial activation sequence mapping in the isolated canine right atrium. Circulation. 1993;88:250-263.
Spach MS, Kootsey JM, Sloan JD. Active modulation of electrical coupling between cardiac cells of the dog: a mechanism for transient and steady state variations in conduction velocity. Circ Res. 1982;51:347-362.
Spach MS, Dolber PC. Relating extracellular potentials and their derivatives to anisotropic propagation at a microscopic level in human cardiac muscle. Circ Res. 1986;58:356-371.
Kim KB, Rodefeld MD, Schuessler RB, Cox JL, Boineau JP. Relationship between local atrial fibrillation interval and refractory period in the isolated canine atrium. Circulation. 1996;94:2961-2967.