Dominant frequency increase rate predicts transition from paroxysmal to long-term persistent atrial fibrillation.

BACKGROUND: Little is known about the mechanisms underlying the transition from paroxysmal to persistent atrial fibrillation (AF). In an ovine model of long-standing persistent AF we tested the hypothesis that the rate of electric and structural remodeling, assessed by dominant frequency (DF) changes, determines the time at which AF becomes persistent. METHODS AND RESULTS: Self-sustained AF was induced by atrial tachypacing. Seven sheep were euthanized 11.5±2.3 days after the transition to persistent AF and without reversal to sinus rhythm; 7 sheep were euthanized after 341.3±16.7 days of long-standing persistent AF. Seven sham-operated animals were in sinus rhythm for 1 year. DF was monitored continuously in each group. Real-time polymerase chain reaction, Western blotting, patch clamping, and histological analyses were used to determine the changes in functional ion channel expression and structural remodeling. Atrial dilatation, mitral valve regurgitation, myocyte hypertrophy, and atrial fibrosis occurred progressively and became statistically significant after the transition to persistent AF, with no evidence for left ventricular dysfunction. DF increased progressively during the paroxysmal-to-persistent AF transition and stabilized when AF became persistent. Importantly, the rate of DF increase correlated strongly with the time to persistent AF. Significant action potential duration abbreviation, secondary to functional ion channel protein expression changes (CaV1.2, NaV1.5, and KV4.2 decrease; Kir2.3 increase), was already present at the transition and persisted for 1 year of follow up. CONCLUSIONS: In the sheep model of long-standing persistent AF, the rate of DF increase predicts the time at which AF stabilizes and becomes persistent, reflecting changes in action potential duration and densities of sodium, L-type calcium, and inward rectifier currents.

Panoramic Endocardial Optical Mapping Demonstrates Serial Rotors Acceleration and Increasing Complexity of Activity During Onset of Cholinergic Atrial Fibrillation.

Background Activation during onset of atrial fibrillation is poorly understood. We aimed at developing a panoramic optical mapping system for the atria and test the hypothesis that sequential rotors underlie acceleration of atrial fibrillation during onset. Methods and Results Five sheep hearts were Langendorff perfused in the presence of 0.25 µmol/L carbachol. Novel optical system recorded activations simultaneously from the entire left and right atrial endocardial surfaces. Twenty sustained (>40 s) atrial fibrillation episodes were induced by a train and premature stimuli protocol. Movies obtained immediately (Initiation stage) and 30 s (Early Stabilization stage) after premature stimulus were analyzed. Serial rotor formation was observed in all sustained inductions and none in nonsustained inductions. In sustained episodes maximal dominant frequency increased from (mean±SD) 11.5±1.74 Hz during Initiation to 14.79±1.30 Hz at Early Stabilization (P<0.0001) and stabilized thereafter. At rotor sites, mean cycle length (CL) during 10 prerotor activations increased every cycle by 0.53% (P=0.0303) during Initiation and 0.34% (P=0.0003) during Early Stabilization. In contrast, CLs at rotor sites showed abrupt decreases after the rotors appearances by a mean of 9.65% (P<0.0001) during both stages. At Initiation, atria-wide accelerations and decelerations during rotors showed a net acceleration result whereby post-rotors atria-wide minimal CL (CLmin) were 95.5±6.8% of the prerotor CLmin (P=0.0042). In contrast, during Early Stabilization, there was no net acceleration in CLmin during accelerating rotors (prerotor=84.9±11.0% versus postrotor=85.8±10.8% of Initiation, P=0.4029). Levels of rotor drift distance and velocity correlated with atria-wide acceleration. Nonrotor phase singularity points did not accelerate atria-wide activation but multiplied during Initiation until Early Stabilization. Increasing number of singularity points, indicating increased complexity, correlated with atria-wide CLmin reduction (P<0.0001). Conclusions Novel panoramic optical mapping of the atria demonstrates shortening CL at rotor sites during cholinergic atrial fibrillation onset. Atrial fibrillation acceleration toward Early Stabilization correlates with the net result of atria-wide accelerations during drifting rotors activity.

Mechanisms by Which Ranolazine Terminates Paroxysmal but Not Persistent Atrial Fibrillation.

BACKGROUND: Ranolazine inhibits Na+ current (INa), but whether it can convert atrial fibrillation (AF) to sinus rhythm remains unclear. We investigated antiarrhythmic mechanisms of ranolazine in sheep models of paroxysmal (PxAF) and persistent AF (PsAF). METHODS: PxAF was maintained during acute stretch (N=8), and PsAF was induced by long-term atrial tachypacing (N=9). Isolated, Langendorff-perfused sheep hearts were optically mapped. RESULTS: In PxAF ranolazine (10 µmol/L) reduced dominant frequency from 8.3±0.4 to 6.2±0.5 Hz (P<0.01) before converting to sinus rhythm, decreased singularity point density from 0.070±0.007 to 0.039±0.005 cm-2 s-1 (P<0.001) in left atrial epicardium (LAepi), and prolonged AF cycle length (AFCL); rotor duration, tip trajectory, and variance of AFCL were unaltered. In PsAF, ranolazine reduced dominant frequency (8.3±0.5 to 6.5±0.4 Hz; P<0.01), prolonged AFCL, increased the variance of AFCL, had no effect on singularity point density (0.048±0.011 to 0.042±0.016 cm-2 s-1; P=ns) and failed to convert AF to sinus rhythm. Doubling the ranolazine concentration (20 µmol/L) or supplementing with dofetilide (1 µmol/L) failed to convert PsAF to sinus rhythm. In computer simulations of rotors, reducing INa decreased dominant frequency, increased tip meandering and produced vortex shedding on wave interaction with unexcitable regions. CONCLUSIONS: PxAF and PsAF respond differently to ranolazine. Cardioversion in the former can be attributed partly to decreased dominant frequency and singularity point density, and prolongation of AFCL. In the latter, increased dispersion of AFCL and likely vortex shedding contributes to rotor formation, compensating for any rotor loss, and may underlie the inefficacy of ranolazine to terminate PsAF.

Effect of sustained-mild and transient-severe hyperglycemia on ischemia-induced blood-brain barrier opening.

The purpose of this study was to examine what levels of hyperglycemia cause blood-brain barrier (BBB) disruption during permanent and transient middle cerebral artery occlusion in the rat and when the adverse effects of hyperglycemia occur. Cerebrovascular function was assessed by measuring the influx rate constant (K(i)) for (3)H-inulin and by measuring cerebral plasma ((14)C-inulin) and (51)Cr-labeled red blood cell (RBC) volume. Different glucose protocols were used to produce mild sustained hyperglycemia (blood glucose approximately 150 mg/dL) or transient-severe hyperglycemia (with a spike in blood glucose of approximately 400 mg/dL). As expected, transient-severe hyperglycemia at the time of occlusion induced marked BBB disruption in animals undergoing 2 h of ischemia with 2 h of reperfusion (25-fold increase in permeability compared with the contralateral core). However, the mild hyperglycemia model induced similar disruption. Similarly, after permanent occlusion, both hyperglycemia models enhanced disruption and they both produced marked ( approximately 50%) reductions in cerebral plasma volume. Apparent cerebral RBC volume also decreased when measured during the final 5 mins of 2 h of ischemia with transient-severe hyperglycemia. However, there was no decrease if the (51)Cr-labeled RBCs were circulated for the whole 2 h, indicating RBC trapping. The spike in blood glucose in the severe hyperglycemia model was used to examine when hyperglycemia induced BBB disruption. Hyperglycemia shortly after occlusion caused severe disruption. In contrast, hyperglycemia after 90 mins of occlusion caused little disruption. These results suggest that mild hyperglycemia has a profound effect on BBB function and that very early correction of hyperglycemia is necessary to prevent adverse effects.

Eplerenone Reduces Atrial Fibrillation Burden Without Preventing Atrial Electrical Remodeling.

BACKGROUND: The aldosterone inhibitor eplerenone (EPL) has been shown to reduce the incidence of atrial fibrillation (AF) in patients with systolic heart failure, but the mechanism is unknown. OBJECTIVES: This study hypothesized that by reducing atrial dilation and fibrosis in the absence of heart failure, EPL also reduces AF burden and prevents AF perpetuation. METHODS: The authors conducted a randomized controlled study in 34 sheep that were atrially tachypaced (13 ± 1 week). They compared daily oral EPL (n = 19) versus sugar pill (SP) treatment (n = 15) from the start of tachypacing. The endpoint was a continuous 7-day stretch of persistent AF (n = 29) or completion of 23 weeks tachypacing (n = 5). RESULTS: EPL significantly reduced the rate of left atrial dilation increase during AF progression. Atria from EPL-treated sheep had less smooth muscle actin protein, collagen-III expression, interstitial atrial fibrosis, and cell hypertrophy than SP-treated sheep atria did. However, EPL did not modify the AF-induced increase in the rate of dominant frequency and ion channel densities seen under SP treatment, but rather prolonged the time to persistent AF in 26% of animals. It also reduced the degree of fibrillatory conduction, AF inducibility, and AF burden. CONCLUSIONS: In the sheep model, EPL mitigates fibrosis and atrial dilation, modifies AF inducibility and AF complexity, and prolongs the transition to persistent AF in 26% of animals, but it does not prevent AF-induced electrical remodeling or AF persistence. The results highlight structural remodeling as a central upstream target to reduce AF burden, and the need to prevent electrical remodeling to avert AF perpetuation.

The effects of cerebral ischemia on the rat choroid plexus.

Although the blood-brain barrier effects of cerebral ischemia have been extensively examined, less attention has focused on ischemia-induced damage to the choroid plexuses that form the blood-cerebrospinal fluid (CSF) barrier (BSCFB). This study examined the rat lateral ventricle choroid plexuses (LVCP) in three ischemic models, bilateral common carotid artery occlusion (2VO)+hypotension with or without reperfusion and permanent middle cerebral artery (MCA) occlusion with or without a tandem common carotid artery occlusion. Blood flow was assessed using [(14)C]-N-isopropyl-p-iodoamphetamine, and LVCP injury by tissue edema, alterations in [(14)C]glutamine transport and BSCFB disruption (measured with [(3)H]inulin). 2VO+hypotension caused an 87% reduction in LVCP blood flow (P<0.01) and a progressive reduction in LVCP glutamine transport. In contrast to cortex, there was no LVCP hyperemia or delayed hypoperfusion on reperfusion, but there was marked BSCFB disruption. After 30 mins of 2VO+hypotension with 6 h of reperfusion, the [(3)H]inulin entry into CSF was increased threefold (P<0.05). Blood-CSF barrier rather than blood-brain barrier disruption appeared to be the main cause of enhanced [(3)H]inulin entry into hippocampus. Middle cerebral artery occlusion with and without a tandem common carotid artery occlusion only caused 53% and 38% reductions in LVCP blood flow but induced LVCP edema. Results suggest that the LVCP is selectively vulnerable to ischemic injury in terms of the absolute blood flows or, for the MCA occlusion models, the % reductions in flows required to induce injury. BCSFB disruption early after ischemia may enhance the movement of compounds from blood to areas close to the ventricular system and participate in delayed neuronal death.

Galectin-3 Regulates Atrial Fibrillation Remodeling and Predicts Catheter Ablation Outcomes.

OBJECTIVES: To determine whether Gal-3 mediates sustained atrial fibrillation (AF)-induced atrial structural and electrical remodeling and contributes to AF perpetuation. BACKGROUND: Galectin-3 (Gal-3) mediates extracellular matrix remodeling in heart failure, but its role in AF progression remains unexplored. METHODS: We examined intracardiac blood samples from patients with AF (N=55) to identify potential biomarkers of AF recurrence. In a sheep model of tachypacing-induced AF (N=20), we tested the effects of Gal-3 inhibition during AF progression. RESULTS: In patients, intracardiac serum Gal-3 levels were greater in persistent than paroxysmal AF and independently predicted atrial tachyarrhythmia recurrences after a single ablation procedure. In the sheep model, both Gal-3 and TGF-ß1 were elevated in the atria of persistent AF animals. The Gal-3 inhibitor GM-CT-01 (GMCT) reduced both Gal-3 and TGF-ß1-induced sheep atrial fibroblast migration and proliferation in vitro. GMCT (12 mg/kg twice/week) prevented the increase in serum procollagen type III N-terminal peptide seen during progression to persistent AF, and also mitigated atrial dilatation, myocyte hypertrophy, fibrosis, and the expected increase in dominant frequency of excitation. Atria of GMCT-treated animals had significantly less TGF-ß1-Smad2/3 signaling pathway activation and expression of α-smooth muscle actin and collagen than saline-treated animals. Ex-vivo hearts from GMCT-treated animals had significantly longer action potential durations and fewer rotors and wavebreaks during AF, and myocytes had lower functional expression of inward rectifier K+ channel (Kir2.3) than saline-treated animals. Importantly, GMCT increased the probability of spontaneous AF termination, decreased AF inducibility and reduced overall AF burden. CONCLUSIONS: Inhibiting Gal-3 during AF progression might be useful as an adjuvant treatment to improve outcomes of catheter ablation for persistent AF. Gal-3 inhibition may be a potential new upstream therapy for prevention of AF progression.

Cell-selective arrhythmia ablation for photomodulation of heart rhythm.

Heart disease, a leading cause of death in the developed world, is overwhelmingly correlated with arrhythmias, where heart muscle cells, myocytes, beat abnormally. Cardiac arrhythmias are usually managed by electric shock intervention, antiarrhythmic drugs, surgery, and/or catheter ablation. Despite recent improvements in techniques, ablation procedures are still limited by the risk of complications from unwanted cellular damage, caused by the nonspecific delivery of ablative energy to all heart cell types. We describe an engineered nanoparticle containing a cardiac-targeting peptide (CTP) and a photosensitizer, chlorin e6 (Ce6), for specific delivery to myocytes. Specificity was confirmed in vitro using adult rat heart cell and human stem cell-derived cardiomyocyte and fibroblast cocultures. In vivo, the CTP-Ce6 nanoparticles were injected intravenously into rats and, upon laser illumination of the heart, induced localized, myocyte-specific ablation with 85% efficiency, restoring sinus rhythm without collateral damage to other cell types in the heart, such as fibroblasts. In both sheep and rat hearts ex vivo, upon perfusion of CTP-Ce6 particles, laser illumination led to the formation of a complete electrical block at the ablated region and restored the physiological rhythm of the heart. This nano-based, cell-targeted approach could improve ablative technologies for patients with arrhythmias by reducing currently encountered complications.

Intracerebral hirudin injection attenuates ischemic damage and neurologic deficits without altering local cerebral blood flow.

There has been considerable interest in the use of thrombin inhibitors to reduce the occurrence of stroke or to potentiate tissue plasminogen activator-induced reperfusion. However, there is growing evidence that thrombin may also have extravascular effects that influence ischemic brain injury. Male Sprague-Dawley rats were subjected to either 90 minutes of temporary middle cerebral artery (MCA) occlusion or sham operation to examine thrombin and protease activated receptor-1 (PAR-1) expression. In another set of rats, the MCA was occluded for 90 minutes and 10 U of hirudin or the same volume of vehicle was injected into the caudate followed by reperfusion for up to 28 days, to test the effects of local thrombin inhibition on ischemic damage, neurologic outcome and cerebral blood flow (CBF). Thrombin immunoreactivity was increased in the ischemic caudate at 4 and 24 hours, whereas PAR-1 expression was unchanged. Hirudin reduced infarct volume in the caudate at 24 hours (79 +/- 41 vs. 115 +/- 20 mm3, P < 0.05) and resulted in a larger residual tissue volume in the caudate at 28 days (17.6 +/- 3.9 vs. 11.8 +/- 6.3 mm3, P < 0.05). Hirudin treatment also had a beneficial effect on body weight and ameliorated neurologic deficits tested by forelimb placing and forelimb use asymmetry during 28 days survival. These beneficial effects of hirudin were not associated with improved regional CBF during reperfusion. These results suggest that, in addition to their effects on coagulation and circulation, thrombin inhibitors also have direct neuroprotective properties and may be considered in stroke therapy.

Glutamine transport at the blood-brain and blood-cerebrospinal fluid barriers.

Glutamine has multiple physiological and pathophysiological roles in the brain. Because of their position at the interface between blood and brain, the cerebral capillaries and the choroid plexuses that form the blood-brain barriers (BBB) and blood-cerebrospinal fluid (CSF) barriers, have the potential to influence brain glutamine concentrations. Despite this, there has been a paucity of data on the mechanisms and polarity of glutamine transport at these barrier tissues. In situ brain perfusion in the rat, indicates that blood to brain L-[14C]glutamine transport at the blood-brain barrier is primarily mediated by a pH-dependent, Na(+)-dependent, System N transporter, but that blood to choroid plexus transport is primarily via a pH-independent System N transporter and a Na(+)-independent carrier that is not System L. Transport studies in isolated rat choroid plexuses and primary cultures of choroid plexus epithelial cells indicate that epithelial L-[14C]glutamine transport is polarized (apical uptake>basolateral) and that uptake at the apical membrane is mediated by pH dependent System N transporters (identified as SN1 and SN2 by polymerase chain reaction) and the Na(+)-independent System L. Blood-brain barrier System N transport is markedly effected by cerebral ischemia and may be a good marker of endothelial cell dysfunction. The multiple glutamine transporters at the blood-brain and blood-CSF barriers may have role in meeting the metabolic needs of the brain and the barrier tissues themselves. However, it is likely that the main role of these transporters is removing glutamine, and thus nitrogen, from the brain.

Intraventricular infusion of vascular endothelial growth factor promotes cerebral angiogenesis with minimal brain edema.

OBJECTIVE: Therapeutic cerebral angiogenesis, i.e., using angiogenic factors to enhance collateral vessel formation within the central nervous system, is a potential method for cerebral revascularization. Vascular endothelial growth factor (VEGF) is a potent endothelial cell mitogen that also increases capillary permeability, particularly in ischemic tissue. The purpose of this study was to assess the angiogenic and capillary permeability effects of chronic intraventricular infusion of exogenous VEGF in nonischemic brain tissue, because many patients with impaired cerebrovascular reserve do not exhibit chronic cerebral ischemia. METHODS: Recombinant human VEGF(165) was infused into the right lateral ventricle of rats at a rate of 1 microl/h for 7 days, at concentrations of 1 to 25 microg/ml, with osmotic minipumps. Control animals received vehicle only. Vessels were identified in laminin immunohistochemical analyses. Capillary permeability and brain edema were assessed with Evans blue extravasation, [(3)H]inulin permeability, and brain water content measurements. RESULTS: Vessel density was dose-dependently increased by VEGF(165) infusions, with significant increases occurring in animals treated with 5 or 25 microg/ml, compared with control animals (P h 0.01). Significant enlargement of the lateral ventricles was observed for the highest-dose group but not for animals treated with other doses. Capillary permeability was assessed in animals treated with a dose of 5 microg/ml. An increase in capillary permeability in the diencephalon was identified with Evans blue extravasation and [(3)H]inulin permeability assessments; however, the brain water content was not significantly increased. CONCLUSION: Chronic intraventricular infusions of VEGF(165) increased vascular density in a dose-dependent manner. There seems to be a therapeutic window, because infusion of VEGF(165) at a concentration of 5 microg/ml resulted in a significant increase in vessel density with minimal associated brain edema and no ventriculomegaly.

Long-term frequency gradients during persistent atrial fibrillation in sheep are associated with stable sources in the left atrium.

BACKGROUND: Dominant frequencies (DFs) of activation are higher in the atria of patients with persistent than paroxysmal atrial fibrillation (AF), and left atrial (LA)-to-right atrial (RA) DF gradients have been identified in both. However, whether such gradients are maintained as long-term persistent AF is established remains unexplored. We aimed at determining in vivo the time course in atrial DF values from paroxysmal to persistent AF in sheep and testing the hypothesis that an LA-to-RA DF difference is associated with LA drivers in persistent AF. METHODS AND RESULTS: AF was induced using RA tachypacing (n=8). Electrograms were obtained weekly from an RA lead and an implantable loop recorder implanted near the LA. DFs were determined for 5-second-long electrograms (QRST subtracted) during AF in vivo and in ex vivo optical mapping. Underlying structural changes were compared with weight-matched controls (n=4). After the first AF episode, DF increased gradually during a 2-week period (7±0.21 to 9.92±0.31 Hz; n=6; P<0.05). During 9 to 24 weeks of AF, the DF values on the implantable loop recorder were higher than the RA (10.6±0.08 versus 9.3±0.1 Hz, respectively; n=7; P<0.0001). Subsequent optical mapping confirmed a DF gradient from posterior LA-to-RA (9.1±1.0 to 6.9±0.9 Hz; P<0.05) and demonstrated patterns of activation compatible with drifting rotors in the posterior LA. Persistent AF sheep showed significant enlargement of the posterior LA compared with controls. CONCLUSIONS: In the sheep, transition from paroxysmal to persistent AF shows continuous LA-to-RA DF gradients in vivo together with enlargement of the posterior LA, which harbors the highest frequency domains and patterns of activation compatible with drifting rotors.

High-resolution endocardial and epicardial optical mapping in a sheep model of stretch-induced atrial fibrillation.

Atrial fibrillation (AF) is a complex cardiac arrhythmia with high morbidity and mortality.(1,2) It is the most common sustained cardiac rhythm disturbance seen in clinical practice and its prevalence is expected to increase in the coming years.(3) Increased intra-atrial pressure and dilatation have been long recognized to lead to AF,(1,4) which highlights the relevance of using animal models and stretch to study AF dynamics. Understanding the mechanisms underlying AF requires visualization of the cardiac electrical waves with high spatial and temporal resolution. While high-temporal resolution can be achieved by conventional electrical mapping traditionally used in human electrophysiological studies, the small number of intra-atrial electrodes that can be used simultaneously limits the spatial resolution and precludes any detailed tracking of the electrical waves during the arrhythmia. The introduction of optical mapping in the early 90's enabled wide-field characterization of fibrillatory activity together with sub-millimeter spatial resolution in animal models(5,6) and led to the identification of rapidly spinning electrical wave patterns (rotors) as the sources of the fibrillatory activity that may occur in the ventricles or the atria.(7-9) Using combined time- and frequency-domain analyses of optical mapping it is possible to demonstrate discrete sites of high frequency periodic activity during AF, along with frequency gradients between left and right atrium. The region with fastest rotors activates at the highest frequency and drives the overall arrhythmia.(10,11) The waves emanating from such rotor interact with either functional or anatomic obstacles in their path, resulting in the phenomenon of fibrillatory conduction.(12) Mapping the endocardial surface of the posterior left atrium (PLA) allows the tracking of AF wave dynamics in the region with the highest rotor frequency. Importantly, the PLA is the region where intracavitary catheter-based ablative procedures are most successful terminating AF in patients,(13) which underscores the relevance of studying AF dynamics from the interior of the left atrium. Here we describe a sheep model of acute stretch-induced AF, which resembles some of the characteristics of human paroxysmal AF. Epicardial mapping on the left atrium is complemented with endocardial mapping of the PLA using a dual-channel rigid borescope c-mounted to a CCD camera, which represents the most direct approach to visualize the patterns of activation in the most relevant region for AF maintenance.

Targeting atrioventricular differences in ion channel properties for terminating acute atrial fibrillation in pigs.

AIMS: The goal was to terminate atrial fibrillation (AF) by targeting atrioventricular differences in ionic properties. METHODS AND RESULTS: Optical mapping was used to record electrical activity during carbachol (0.25-0.5 µM)-induced AF in pig hearts. The atrial-specific current, I(Kur), was blocked with 100 µM 4-aminopyridine (4-AP) or with 0.5 µM DPO-1. Hearts in AF and ventricular fibrillation (VF) were also subjected to increasing levels of extracellular K(+) ([K(+)](o): 6-12 mM), compared with controls (4 mM). We hypothesized that due to the more negative steady-state half inactivation voltage for the atrial Na(+) current, I(Na), compared with the ventricle, AF would terminate before VF in hyperkalaemia. Mathematical models were used to interpret experimental findings. The I(Kur) block did not terminate AF in a majority of experiments (6/9 with 4-AP and 3/4 with DPO-1). AF terminated in mild hyperkalaemia ([K(+)](o) ≤ 10.0 mM; N = 8). In contrast, only two of five VF episodes terminated at the maximum ([K(+)](o): 12 mM [K(+)](o)). The I(Kur) block did not terminate a simulated rotor in cholinergic AF because its contribution to repolarization was dwarfed by the large magnitude of the acetylcholine-activated K(+) current (I(K,ACh)). Simulations showed that the lower availability of the atrial Na(+) current at depolarized potentials, and a smaller atrial tissue size compared with the ventricle, could partly explain the earlier termination of AF compared with VF during hyperkalaemia. CONCLUSION: I(Kur) is an ineffective anti-arrhythmic drug target in cholinergic AF. Manipulating Na(+) current 'availability' might represent a viable anti-arrhythmic strategy in AF.