Effects of Experimental Heart Failure on Atrial Cellular and Ionic Electrophysiology
Background—Congestive heart failure (CHF) is frequently associated with atrial fibrillation (AF), but little is known about the effects of CHF on atrial cellular electrophysiology.
Methods and Results—We studied action potential (AP) properties and ionic currents in atrial myocytes from dogs with CHF induced by ventricular pacing at 220 to 240 bpm for 5 weeks. Atrial myocytes from CHF dogs were hypertrophied (mean±SEM capacitance, 89±2 pF versus 71±2 pF in control, n=160 cells per group, P<0.001). CHF significantly reduced the density of L-type Ca2+ current (ICa) by ≈30%, of transient outward K+ current (Ito) by ≈50%, and of slow delayed rectifier current (IKs) by ≈30% without altering their voltage dependencies or kinetics. The inward rectifier, ultrarapid and rapid delayed rectifier, and T-type Ca2+ currents were not altered by CHF. CHF increased transient inward Na+/Ca2+ exchanger (NCX) current by ≈45%. The AP duration of atrial myocytes was not altered by CHF at slow rates but was increased at faster rates, paralleling in vivo refractory changes. CHF created a substrate for AF, prolonging mean AF duration from 8±4 to 535±82 seconds (P<0.01).
Conclusions—Experimental CHF selectively decreases atrial Ito, ICa, and IKs, increases NCX current, and leaves other currents unchanged. The cellular electrophysiological remodeling caused by CHF is quite distinct from that caused by atrial tachycardia, highlighting important differences in the cellular milieu characterizing different clinically relevant AF substrates.
Congestive heart failure (CHF) is a common cause of clinical atrial fibrillation (AF),1 but how it causes the arrhythmia is unclear. Atrial diseases associated with increased atrial pressure are accompanied by abnormal action potential (AP) properties in humans.2 Data regarding the effects of CHF on atrial ionic currents are limited. Atrial myocytes from hearts with terminal failure showed decreased L-type Ca2+ current (ICa) compared with relatively normal hearts.3 Koumi et al4 noted reduced atrial inward rectifier current (IK1) in patients with CHF; however, Le Grand et al5 found no change in IK1 in dilated human atria, whereas ICa and transient outward K+ current (Ito) were reduced. These data are limited by a wide range of heart diseases and variable drug therapy. Many patients also had AF, which can directly alter atrial ionic properties.6 7 Thus, it is difficult to separate CHF-induced ionic abnormalities in patients from those caused by underlying heart disease, concomitant drug therapy, or AF. Rapid ventricular pacing has been used to study experimental CHF-induced ventricular ionic current changes8 9 ; however, the ionic changes caused by experimental CHF at the atrial level are unknown and were the object of the present study.
Preparation of Animal Model
To produce CHF, the right ventricle was stimulated at 240 bpm for 3 weeks, followed by 2 weeks at 220 bpm. On study days, dogs were anesthetized with morphine (2 mg/kg SC) and α-chloralose (120 mg/kg IV load, 29.25 mg · kg−1 · h−1 infusion), and a median sternotomy was performed. Right atrial (RA) appendage effective refractory period (ERP) was measured with 15 basic (S1) stimuli, followed by a premature (S2) stimulus applied with 5-ms decrements (ERP=longest S1S2 failing to capture, mean of 3 determinations). S1 and S2 were twice-threshold, 2-ms pulses. AF was induced 10 times by burst pacing (10-Hz, 2-ms pulses, 4× threshold currents), and mean duration was noted. If persistent AF requiring cardioversion occurred on 2 occasions, further AF inductions were not performed. Hearts were excised and atrial cells isolated from RA pectinate muscles as previously described.10 In vivo studies were performed in the first 8 control and 18 CHF dogs. Because performing both in vivo and in vitro studies in each dog greatly prolonged the experiments, subsequent dogs were used for in vitro study only, after it had been verified that in vitro results were similar whether or not they were preceded by in vivo study.
Cells were isolated from 28 control and 28 CHF dogs. Cells in a 1-mL bath were superfused (3 mL/min, 35°C, or for the ultrarapid delayed rectifier potassium current [IKur,d], room temperature11 ). The standard pipette solution contained (in mmol/L) potassium aspartate 110, KCl 20, MgCl2 1, GTP 0.1, magnesium ATP 5, sodium phosphocreatine 5, EGTA 10, and HEPES 10. For AP recording, the EGTA concentration was decreased to 50 μmol/L. For ICa recording, K+ was replaced by Cs+. The standard extracellular solution contained (in mmol/L) NaCl 136, KCl 5.4, CaCl2 2, MgCl2 0.8, NaH2PO4 0.33, dextrose 10, and HEPES 10. For ICa recording, tetraethylammonium chloride and CsCl replaced NaCl and KCl, respectively. For K+ current measurement, CdCl2 (200 μmol/L) or (for studies of IK) 5 μmol/L nifedipine was added to block ICa and ICl,Ca. When a holding potential (HP) negative to −50 mV was used in studies of K+ currents, Tris-HCl replaced NaCl. The storage solution contained (in mmol/L) KCl 20, KH2PO4 10, glutamic acid 70, β-hydroxybutyric acid 10, taurine 10, EGTA 10, and dextrose 10, plus albumin 1% (pH adjusted to 7.4 with KOH or with CsOH for K+-free solutions). For IK studies, 2 mmol/L 4-aminopyridine was used to block Ito and IKur,d. Ito was studied in the presence of 10 mmol/L tetraethylammonium and 1 μmol/L dofetilide to inhibit IKur,d and IKr, respectively. IK1 was quantified as 0.5 mmol/L Ba2+-sensitive currents at the end of 300-ms pulses from −40 mV.
For Na+/Ca2+ exchanger (NCX) current recording, the extracellular solution contained (in mmol/L) NaCl 140, CaCl2 1, MgCl2 1, dextrose 10, and HEPES 10, plus niflumic acid 50 μmol/L. The EGTA-free pipette solution contained (in mmol/L) CsCl 130, NaCl 5, Mg-ATP 4, and HEPES 10. NCX current was activated as previously described12 by repolarization of the cell to various potentials after 5-ms depolarizing pulses to +20 mV from an HP of −70 mV.
Tight-seal patch-clamp was used to record currents (voltage-clamp mode) and APs (current-clamp mode). Electrode resistances were 1 to 2 MΩ for current recording and 3 to 5 MΩ for AP recording. Cells with normal resting potentials (negative to −70 mV) were selected for AP recording after it had been verified that resting potential was not altered by CHF. APs (elicited by 2-ms twice-threshold pulses) were analyzed at steady state at each frequency. Recordings were low-pass–filtered at half the sampling frequency (2 kHz for IKr and slow delayed rectifier current [IKs] and 10 kHz for APs, ICa, Ito, and IKur,d).
Junction potentials (6 to 10 mV) were corrected for AP recordings only. Seal resistance averaged 8.6±1.2 GΩ. Series resistance and capacitive time constant averaged 7.6±0.3 MΩ and 506±13 μs before and 1.6±0.04 MΩ and 122.4±6.2 μs after compensation, respectively. The mean maximum series resistance voltage drop was ≈3 mV. Similar numbers of cells from each heart were studied with each protocol. Cell capacitance (the time integral of capacitance current divided by the voltage drop during 5-mV steps) was increased in CHF cells (89.2±1.9 versus 70.5±1.5 pA/pF in control, n=160 each, P<0.001), so currents are presented in terms of densities.
Western Blot Analysis of NCX Expression
Western blot analysis of NCX expression has been used to evaluate ventricular NCX changes in CHF.13 To evaluate NCX protein expression, fast-frozen left atrial samples were placed on ice in radioimmunoprecipitation (RIPA) buffer containing Igepal 1%, sodium deoxycholate 0.5%, SDS 0.1%, β-mercaptoethanol 10 mmol/L, sodium orthovanadate 100 mmol/L, aprotinin 1 μg/mL, leupeptin 1 μg/mL, pepstatin 1 μg/mL, and PMSF 1 mmol/L in PBS. Samples were then minced and incubated/rotated (1 hour, 4°C). The homogenate was then centrifuged at 4°C and the supernatant stored at −70°C. Protein concentrations were measured by Bradford assay.
Solubilized membrane proteins were fractionated on 7.5% SDS-polyacrylamide gels at 200 V for 60 minutes, then transferred electrophoretically to Immobilon-P PVDF membranes (Millipore) in 25 mmol/L Tris base, 192 mmol/L glycine, and 5% methanol (18 hours, 0.07 A). Membranes were blocked with 5% nonfat dry milk (NFDM) in TBS (Tris-HCl 50 mmol/L, NaCl 500 mmol/L; pH 7.5) containing 0.05% Tween-20 (TTBS) (2 hours, room temperature), and then incubated overnight (4°C) with NCX antibody (Research Diagnostics; 1:1000 in TTBS with 1% NFDM). Membranes were then washed 3 times in TTBS and reblocked in 1% NFDM/TTBS for 10 minutes, followed by incubation with peroxidase-conjugated goat anti-mouse IgM (1:6000) in 5% NFDM/TTBS (30 minutes) and 3 TTBS washes. Antibody was detected with Western blot Chemiluminescence Reagent Plus (NEN Life Science Products). Band density was quantified with Quantity One software (PDI) incorporating background subtraction algorithms. Total NCX density was calculated as the sum of the densities of each of the 3 isoforms.14
Group data are presented as mean±SEM. Nonlinear curve fitting was performed with Clampfit in pCLAMP6 (Chebyshev algorithm). Nonpaired t tests (2-tailed) were used for statistical comparisons between control and CHF cells.
In Vivo and AP Findings
All dogs with CHF had increased RA (12±1 versus 3.0±0.3 mm Hg, P<0.01) and left ventricular (19±1 versus 2±1 mm Hg, P<0.01) diastolic pressures, evident ascites and pulmonary congestion, and increased respiratory rates (56±2 versus 24±1 breaths per minute, P<0.001). Atrial rates in CHF dogs averaged 139±5 bpm with and 138±4 bpm without ventricular pacing, versus 138±6 bpm in controls. Mean AF duration was 8±4 seconds in controls, versus 535±82 seconds in CHF dogs (P<0.01), with persistent AF (>30 minutes, requiring electrical cardioversion for termination) in 0 of 8 controls versus 10 of 18 CHF dogs.
Resting potential averaged −74.0±0.9 and −72.4±0.4 mV in 20 control and 20 CHF myocytes used for AP analysis, respectively (P=NS). AP amplitude (1 Hz) averaged 118.8±4.1 mV in control and 116.0±3.8 mV in CHF myocytes (P=NS). CHF prolonged AP duration (APD) increasingly as frequency increased (Figure 1⇓). APDs in isolated cells paralleled ERP in vivo (eg, at 3 Hz, APD90=138±4 ms in controls and 145±6 ms in CHF; ERP=131±8 ms in controls and 149±7 ms in CHF; at 6 Hz, APD90=97±3 ms in controls and 117±6 ms in CHF; ERP=93±4 ms in controls and 122±7 ms in CHF). We did not observe delayed afterdepolarizations (DADs) during AP recording.
Changes in K+ Currents
Ito was decreased by CHF (Figure 2A⇓ through 2C). CHF did not alter the form of the I-V relation (Figure 2D⇓) or voltage dependence (Figure 2E⇓) or the kinetics (Figure 2F⇓) of Ito activation and inactivation. Ito reactivation (Figure 2G⇓) was monoexponential (time constants for control versus CHF were 28.1±0.3 versus 30.1±0.5 ms at −70 mV, 59.6±8.5 versus 65.3±2.4 ms at −60 mV, and 88.9±9.3 versus 93.4±11.9 ms at −50 mV; n=7 per group, P=NS) and, along with frequency dependence (Figure 2H⇓), was unaffected by CHF.
To study IK, currents were recorded as shown in Figure 3⇓ 10 minutes after membrane rupture and then again at least twice at 10-minute intervals to exclude cells with >10% rundown over 20 minutes. Dofetilide (1 μmol/L) was then added to block IKr. The drug-resistant component (IKs, Figure 3A⇓ and 3B⇓) was reduced by CHF. CHF reduced both step (Figure 3C⇓) and tail (Figure 3D⇓) current densities by ≈30%. The voltage dependence of IKs activation (tail-current analysis, Figure 3E⇓) was not altered by CHF (mean V1/2, +21.3±2.5 mV in CHF and +22.1±2.5 mV in control; n=15 per group), nor were the kinetics of IKs activation (Figure 3F⇓).
Examples of dofetilide-sensitive IKr in control and CHF cells are shown in Figure 4A⇓ and 4B⇓. IKr density (Figure 4C⇓) and activation voltage dependence (Figure 4D⇓) were not changed by CHF. Neither IK1 nor IKur,d were altered by CHF, as shown in Figure 5⇓.
Changes in Ca2+ Currents
Typical L-type ICa recordings in control and CHF cells are shown in Figure 6A⇓ and 6B⇓. ICa density decreased significantly (Figure 6C⇓) in CHF myocytes (eg, from −16.6±1.4 to −11.5±0.1 pA/pF at +10 mV, n=35 cells each, P<0.01), without any alteration in the form of the I-V curve (Figure 6D⇓). CHF did not alter the voltage dependence of ICa inactivation or activation (Figure 6E⇓). The time course of ICa inactivation was biexponential and voltage dependent but was not altered by CHF (Figure 6F⇓). ICa recovery was similar in CHF and normal myocytes (Figure 6G⇓), with monoexponential time constants averaging 28.6±1.5 and 30.9±3.5 ms for control and CHF, respectively. CHF similarly did not alter ICa frequency dependence (Figure 6H⇓).
T-type ICa was separated from L-type by subtraction of currents recorded at an HP of −50 mV from those at −90 mV, as illustrated in Figure 7A⇓ and 7B⇓. ICa,T was present in 40% of atrial myocytes (20 of 50 for both control and CHF). As shown in Figure 7C⇓ and 7D⇓, CHF did not alter ICa,T density.
NCX recordings in control and CHF cells are shown in Figure 8A⇓. Substitution of Li+ for extracellular Na+ strongly suppressed inward currents, typical for the NCX.12 14 CHF substantially increased the density of NCX current (Figure 8B⇓). For example, at −70 mV (close to the mean resting potential), NCX averaged −1.1±0.1 in control cells versus −1.6±0.2 pA/pF in CHF cells (P=0.02). NCX protein expression was also increased (Figure 8C⇓), by an average of 87% in atrial tissues from CHF (n=7 blots from 1 heart each) compared with control dogs (n=4).
In the present study, we found that experimental CHF decreases Ito, L-type ICa, and IKs and increases NCX in atrial myocytes, without altering T-type ICa, IKr, IKl, or IKur,d. The atrial cellular electrophysiological substrate produced by CHF is quite different from that seen with atrial tachycardia–induced remodeling, although both experimental paradigms promote AF.
Comparison With Previous Studies of Cellular Electrophysiology Associated With Atrial Disease and Arrhythmias
It is well recognized that CHF frequently causes AF in humans. Alterations in atrial cellular electrophysiology have been reported in patients2 3 4 5 6 and experimental animals7 15 16 17 with atrial disease and/or dilation. Similar to our observations at slower rates, APD was not substantially altered in dogs with tricuspid insufficiency15 and chronic mitral regurgitation.16 Cats with endogenous cardiomyopathy had APDs that were either unchanged or (in more severe cases) increased in the left atrium.17 Decreases in Ito5 and ICa3 5 have been reported in dilated human atria. IK1 was unchanged in dilated atria5 but reduced in atria of patients with symptomatic CHF.4 Both Ito and the sustained outward current are reduced in patients with CHF.5 It is clear that atrial pathology affects cellular electrophysiology and that the latter changes contribute to arrhythmogenesis. However, the information about ionic alterations obtained from studies in human tissues is limited by various disease states, duration of disease, drug therapy, and numbers of patients with AF in each study. For example, in the study by Le Grand et al5 of ionic currents in patients with dilated atria, 25% of patients also had AF. In the study by Van Wagoner et al6 of K+ current changes in patients with AF, all the tissue samples were from patients with significant dilation of ≥1 atria.
Sustained atrial tachycardia causes a substrate that supports AF and at the cellular level reduces APD and APD rate adaptation.7 Although CHF also produced a substrate for AF in the present study, APD was unchanged at slower rates and increased at faster rates. APD adaptation was reduced, but not as much as with atrial tachycardia. The differences in APD remodeling between the CHF and atrial tachycardia models are associated with differences in ionic current changes. Atrial tachycardia reduces ICa by ≈70% without affecting IK.7 CHF reduced ICa to a lesser extent (by ≈30%), significantly reduced IK by reducing IKs (to the same extent as ICa), and increased the inward current carried by the NCX. NCX expression is not affected by rapid atrial pacing,18 and we are not aware of previous studies of atrial NCX current in patients with CHF. In the atrial tachycardia model, strong reductions in ICa largely account for decreases in APD and APD rate adaptation.7 19 In CHF cells, the ICa reductions are offset by decreases in IK and increases in NCX, which probably accounts for the lack of APD alterations at slow rates. The attenuated rate adaptation caused by ICa reduction may account for the increased APD at faster rates.
Comparison of Atrial Electrophysiological Abnormalities With Those in Ventricular Myocytes Caused by CHF
We are unaware of other studies of the impact of experimental CHF on atrial cellular electrophysiology. There is, however, an extensive literature regarding ventricular cellular abnormalities caused by CHF. Ventricular Ito is decreased in dogs with pacing-induced CHF8 and in cardiomyopathic Syrian hamsters.20 Conflicting results have been reported for ventricular L-type ICa, with either a reduction9 20 or no change8 21 reported. Ventricular APD is generally prolonged by CHF.8 20 Ventricular IK1 is reduced in both experimental8 20 and clinical22 CHF. Ventricular NCX is increased by experimental CHF and probably plays an important role in APD prolongation and ventricular tachyarrhythmia promotion at the ventricular level.13
The APD decrease caused by atrial tachycardia plays an important role in leading to multiple-circuit reentrant AF.7 23 The lack of APD decrease in CHF cells suggests other mechanisms of AF promotion. Previous work from our laboratory suggests that structural changes, particularly interstitial fibrosis, cause important local conduction abnormalities that may play a central role in the reentrant substrate of CHF-related AF.23 24 The NCX increases that we noted in the present study may be quite important in atrial arrhythmia promotion by CHF. The NCX is known to carry transient inward currents that cause DADs and triggered activity.14 25 Stambler and coworkers have observed atrial tachycardias with properties of triggered arrhythmias in dogs with CHF,25A to which increased NCX could well be an important contributor. AF is readily induced in CHF dogs by atrial burst pacing but not by single extrastimuli.24 Therefore, atrial tachycardias caused by NCX-related triggered activity may be the trigger needed to initiate AF in the presence of the CHF-induced structural substrate. We did not observe triggered activity or DADs during AP recording in vitro. Additional conditions present in vivo in the setting of CHF, such as increased catecholamines, tissue stretch, and metabolic abnormalities, may therefore be necessary to produce triggered activity.
Our results could have potential implications for understanding the effects of antiarrhythmic drug therapy on AF. Most antiarrhythmic drugs act by inhibiting cardiac ion channels, so any change in the relative importance of currents flowing during the AP could have a major impact on the action of antiarrhythmic drugs. For example, we have found dofetilide to be unusually effective in terminating AF in dogs with experimental CHF.25B This greater efficacy may be due to CHF-induced downregulation of atrial IKs, leaving repolarization more dependent on IKr.
The development of the atrial tachycardia model of AF was an important advance in studying the arrhythmia.23 26 27 It provided a reliable and reproducible animal model of AF, in which the cellular and ionic mechanisms underlying the arrhythmia substrate could be studied.7 The present study adds a new element to our understanding of the cellular electrophysiological basis of AF by showing that experimentally induced ventricular failure produces a substrate for AF with different AP and ionic changes from those caused by atrial tachycardia. Both of these experimental paradigms are important for our understanding of the cellular basis of AF in humans. When AF begins in patients with CHF, the underlying cellular electrophysiological milieu is established by the ionic current abnormalities caused by CHF. Should AF persist, the resulting atrial tachycardia will cause further atrial cellular remodeling, superimposing tachycardia-induced changes on the underlying CHF substrate. This dynamic nature of the cellular substrate for AF is important to appreciate in order to understand the factors controlling the occurrence and perpetuation of the arrhythmia.
IK is very sensitive to isolation procedures,10 and therefore, changes in isolation technique could contribute to differences in IK when several groups are compared. To minimize possible effects of time-dependent changes in enzymes, isolation procedure, etc, animals from each group were studied concurrently in an alternating fashion. Rundown can be a problem in studies of IK and ICa. All current protocols were applied in the same order for control and CHF cells. Cells with significant rundown (>10% over 20 minutes) were rejected.
We studied a specific animal model of CHF, which has a clinical counterpart in tachycardia-induced ventricular cardiomyopathies28 but which may be different in specific aspects of pathophysiology from other forms of clinical CHF. Furthermore, such issues as the rate of development of CHF, the neuroendocrine response, and the distribution of hemodynamic load may be important determinants of ionic changes. Ionic heterogeneity is significant in the canine atrium,29 and it is therefore essential to compare cells from the same region(s) in CHF dogs versus normal dogs. We studied cells from right atrial pectinate muscles to compare with previous studies of tachycardia-induced ionic remodeling performed with right atrial pectinate myocytes.7 Further work on regional variability in remodeling would be interesting but is outside the scope of this study.
We have found that experimental CHF promotes AF and causes discrete changes in atrial AP properties and currents. NCX overexpression may be particularly important by promoting the occurrence of arrhythmogenic afterdepolarizations. The differences between CHF-induced atrial cellular electrophysiological remodeling and that caused by atrial tachycardia point to the potentially variable ionic pathophysiology associated with different substrates for AF. These results are of fundamental importance for our understanding of basic atrial arrhythmia mechanisms and the response to antiarrhythmic drugs.
This work was supported by the Medical Research Council of Canada and the Quebec Heart Foundation. Danshi Li is a Heart and Stroke Foundation of Canada (HSFC) fellow, and Peter Melnyk is an HSFC student. The authors thank Emma De Blasio and Mirie Levi for technical assistance and France Thériault for secretarial help.
- Received October 12, 1999.
- Revision received December 14, 1999.
- Accepted December 22, 1999.
- Copyright © 2000 by American Heart Association
Mary-Rabine L, Albert A, Pham TD, et al. The relationship of human atrial cellular electrophysiology to clinical function and ultrastructure. Circ Res. 1983;52:188–199.
Koumi SI, Arentzen CE, Backer CL, et al. Alterations in muscarinic K+ channel response to acetylcholine and to G protein–mediated activation in atrial myocytes isolated from failing human hearts. Circulation. 1994;90:2213–2224.
Van Wagoner DR, Pond AL, McCarthy PM, et al. Outward K+ current densities and Kv1.5 expression are reduced in chronic human atrial fibrillation. Circ Res. 1997;80:772–781.
Yue L, Feng J, Gaspo R, et al. Ionic remodeling underlying action potential changes in a canine model of atrial fibrillation. Circ Res. 1997;81:512–525.
Kääb S, Nuss HB, Chiamvimonvat N, et al. Ionic mechanism of action potential prolongation in ventricular myocytes from dogs with pacing-induced heart failure. Circ Res. 1996;78:262–273.
Yue L, Feng J, Li GR, et al. Transient outward and delayed rectifier currents in canine atrium: properties and role of isolation methods. Am J Physiol. 1996;270:H2157–H2168.
O’Rourke B, Kass DA, Tomaselli GF, et al. Mechanisms of altered excitation-contraction coupling in canine tachycardia-induced heart failure, I: experimental studies. Circ Res. 1999;84:562–570.
Blaustein M, Lederer WJ. Sodium/calcium exchange: its physiological implications. Physiol Rev. 1999;79:763–854.
Boyden PA, Hoffman BF. The effects on atrial electrophysiology and structure of surgically induced right atrial enlargement in dogs. Circ Res. 1981;49:1319–1331.
Boyden PA, Tilley LP, Albala A, et al. Mechanisms for atrial arrhythmias associated with cardiomyopathy: a study of feline hearts with primary myocardial disease. Circulation. 1984;69:1036–1047.
Yue L, Melnyk P, Gaspo R, et al. Molecular mechanisms underlying ionic remodeling in a dog model of atrial fibrillation. Circ Res. 1999;84:776–784.
Courtemanche M, Ramirez RJ, Nattel S. Ionic targets for drug therapy and atrial fibrillation-induced electrical remodeling: insights from a mathematical model. Cardiovasc Res. 1999;42:477–489.
Sen L, Smith TW. T-type Ca2+ channels are abnormal in genetically determined cardiomyopathic hamster hearts. Circ Res. 1994;75:149–155.
Beuckelmann DJ, Nabauer M, Erdmann E. Alterations of K+ currents in isolated human ventricular myocytes from patients with terminal heart failure. Circ Res. 1993;73:379–385.
Li D, Fareh S, Leung TK, et al. Promotion of atrial fibrillation by heart failure in dogs: atrial remodeling of a different sort. Circulation. 1999;100:87–95.
Benardeau A, Hatem SN, Rucker-Martin C, et al. Contribution of Na+/Ca2+ exchange to action potential of human atrial myocytes. Am J Physiol. 1996;271:H1151–H1161.
Fenelon G, Manders T, Stambler BS. Atrial tachycardia in dogs with ventricular pacing-induced congestive heart failure originates from multiple foci in the crista terminalis and pulmonary veins: experimental evidence supporting the “atrial ring of fire” hypothesis. Circulation. 1997;96:I-237. Abstract.
Li D, Benardeau A, Nattel S. Contrasting efficacy of dofetilide in differing experimental models of atrial fibrillation. Circulation. In press.
Wijffels MC, Kirchhof CJ, Dorland R, et al. Atrial fibrillation begets atrial fibrillation: a study in awake chronically instrumented goats. Circulation. 1995;92:1954–1968.
Morillo CA, Klein GJ, Jones DL, et al. Chronic rapid atrial pacing: structural, functional, and electrophysiological characteristics of a new model of sustained atrial fibrillation. Circulation. 1995;91:1588–1595.
Feng J, Yue L, Wang Z, et al. Ionic mechanisms of regional action potential heterogeneity in the canine right atrium. Circ Res. 1998;83:541–551.