White New Zealand rabbits of both sexes (n=17, 3.0–4.9kg) were anesthetized with sodium pentobarbital (60mg/kg, iv). A cannula was inserted below the larynx by tracheotomy for intermittent ventilation with positive pressure and an air/O2 mixture (model 607 respiration pump, Harvard Apparatus, UK; 60cycles/min) after neuromuscular blockade (vecuronium 0.05mg/kg/h, iv) to control respiratory rate and volume. The level of anesthesia was assessed by monitoring blood pressure (BP), heart rate (HR) and respiratory activity, using supplementary drug doses to maintain stable general anesthesia. The femoral artery was cannulated for continuous BP monitoring using a SensoNor 840 transducer and the femoral vein was used for administration of 0.9% saline solution and the drugs. The bladder was catheterized and drained. Rectal temperature was maintained at ∼38°C by means of a heating pad (Harvard Apparatus, UK) placed under the animal. The ECG was obtained with bipolar subcutaneous electrodes placed in all four limbs and the signal was amplified and filtered. The ECG, BP and electrograms were continuously monitored on a polygraph (Lectromed, Hertfordshire, UK) and analyzed (PowerLab software, ADInstrument).

The anesthetized animal was placed in ventral decubitus to expose the dorsal portion of the thorax, and an incision was made at the level of the C6 and T2 vertebrae, using electrocauterization and an automatic retractor to access the T1 vertebra, which was perforated with a 2mm drill to expose the spinal dura mater. With the animal in dorsal decubitus, a median incision was made in the neck and the right cervical vagus nerve was identified; this was then isolated and prepared for bipolar electrical stimulation with platinum electrodes, isolated with soft paraffin. A slight rotation enabled a bipolar electrode to be inserted through the spinal canal using a micromanipulator (M330, WPI) to stimulate thoracic sympathetic flow at the level of T1, which contains mainly preganglionic fibers to the stellate ganglion.21 The electrodes for vagus nerve and sympathetic trunk stimulation were connected to a multi-channel programmable pulse generator (Master-8, AMPI, Israel). The vagus nerve was stimulated (20Hz, 2-ms impulse), with the amplitude adjusted to obtain a ≈50% decrease in HR, and sympathetic stimulation (3Hz, 1-ms impulse) was tested to obtain increases in HR and BP.

Besides monitoring HR and BP, we recorded 30-s series of RR intervals for spectral analysis using fast Fourier transform and MatLab software (MathWorks, Natick, MA, USA) in Origin (OriginLab, Origin Lab Corporation, Northampton, MA, USA) to confirm autonomic stimulation, in accordance with a previously described method.22,23 The spectral frequency was divided into three components: very low-frequency (VLF: 0.01–0.1Hz), low-frequency (LF: 0.1–0.3Hz) and high-frequency (HF: 0.3–1Hz). The HF band was attributed to vagal modulation and the LF band was considered to reflect sympathetic activity.

A median sternotomy was performed and the pericardium opened to expose the heart and for epicardial electrophysiological study of the atria and PVs. Following thoracotomy, an end-expiratory pressure of 1–2cm H2O was maintained.

The experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals, published by the US National Institutes of Health, and were approved by the Ethics Committee of the Faculty of Medicine of Lisbon University.

The electrograms were recorded with four monopolar electrodes (125-μm Teflon-insulated Ag/AgCl), placed in a nylon mesh with 4-mm spacing and applied in epicardial position, from the high lateral right atrium (RA) to the high lateral left atrium (LA) (Fig. 1). A circular electrode (coaxial cable) was placed around the proximal segment of the left PVs for stimulation and electrogram recording. The electrograms were amplified and filtered from 0.5 to 300Hz and from 30 to 250Hz for monopolar and bipolar recordings, respectively (Neurolog, Digitimer, UK). At the end of the procedure, the animals were sacrificed with an overdose of anesthetic.

Figure 1.

Open-chest rabbit heart preparation with nylon mesh supporting epicardial monopolar electrodes from the high lateral right atrium to the high lateral left atrium. The arrows indicate two electrodes with epicardial contact on the left atrium. RAA: right atrial appendage.

(0.34MB).

Epicardial activation was recorded in sinus rhythm, and conduction times from RA to LA and from RA to PV depolarization were measured in baseline conditions and during vagal, sympathetic or combined sympathovagal stimulation. Stimulation to study ERPs and assess AF inducibility was performed with a bipolar electrode (0.1mm, Teflon-insulated silver), positioned with a micromanipulator (M330, WPI) in the RAA and the LAA, and with a circular electrode in the PVs. In stable conditions, the programmed stimulation protocol was performed (Master-8 stimulator, AMPI, Israel), with an extrastimulus being introduced during continuous bipolar pacing with a 300-ms cycle length (or 20ms less than the RR interval for a heart rate of >200bpm), using 1-ms impulses and twice the amplitude of the capture threshold, and a coupling interval 100ms shorter than the baseline pacing cycle, which was then reduced in 5-ms steps until the ERP was reached. The ERP was taken as the longest S1–S2 interval that failed to initiate local depolarization. Conduction times and ERPs were analyzed in baseline conditions and during vagal, sympathetic and combined stimulation. To assess vulnerability for AF induction, burst pacing (50Hz, 10s, suprathreshold intensity) was used, with the bipolar electrode situated in the RAA, LAA and PVs, alone or combined with vagal, sympathetic or sympathovagal stimulation. The induction protocol was applied up to three times at each stimulation site. AF was defined as a rapid disordered atrial rhythm, with variable cycle length, polarity, configuration and amplitude, and lasting more than five cycles.24 AF duration was assessed under each of the experimental conditions.

Continuous variables were expressed as means±standard deviation and categorical variables as frequencies and percentages. Comparisons were made using the ANOVA test and the Student's t test for continuous variables. The Kolmogorov–Smirnov test was used to assess the normality of distribution of continuous variables. ERP values and conduction times, in baseline conditions and during autonomic stimulation, were compared by the paired Student's t test. The chi-square test with Yates correction was used for categorical variables. Results with p<0.05 were considered significant. The statistical package used was GraphPad (GraphPad Software Inc., CA, USA).

Conduction times between the RA and the LA increased with vagal stimulation from 20±4ms to 30±10ms (p<0.05), and decreased to 16±6ms with sympathetic stimulation (p<0.05). Combined sympathovagal stimulation also significantly changed interatrial conduction (from 21±5ms to 31±11ms, p<0.05). Conduction times between the RA and the PVs increased during vagal stimulation (p<0.05) and decreased with sympathetic stimulation (p<0.01), but did not change significantly with sympathovagal stimulation (Table 1).

With regard to ERP values (Table 1), refractoriness was higher in the LAA than in the RAA and PVs (83±16ms vs. 69±16ms and 59±16ms, respectively; p<0.01) in baseline conditions. Vagal and sympathovagal stimulation consistently reduced ERPs significantly in the RAA, LAA and PVs, while sympathetic stimulation resulted in a decrease in ERPs only in the LAA. During autonomic stimulation, ERP shortening was more marked in the LAA than at the other sites. However, with vagal stimulation alone, the differences in ERPs between the PVs and the LAA remained statistically significant (p<0.05).

Autonomic stimulation had a significant effect on the induction and duration of AF (Table 1). Burst pacing at 50Hz in the RAA, LAA and PVs induced AF intermittently in 41%, 35% and 53% of the rabbits, respectively. During vagal stimulation, AF was reproducibly induced in 76%, 65% and 76% of the animals, with burst pacing in the RAA, LAA and PVs, respectively (50Hz vs. 50Hz+vagal stimulation, p<0.05 for the RAA), and AF duration increased significantly with pacing in the RAA (6.8±2.5 vs. 1.8±1.2s, p<0.01) and PVs (3.6±1.6 vs. 1.7±0.7s, p<0.05). In four animals, AF was induced with an extrastimulus combined with vagal stimulation during assessment of ERPs in the RAA.

AF inducibility increased with sympathetic stimulation (71%, 65% and 76% for the RAA, LAA and PVs, respectively) (Fig. 5), without reaching statistical significance compared to burst pacing alone. In two animals, short periods of AF were induced with an extrastimulus at the different sites, and spontaneous premature depolarization originating in the PVs occurred in another two. AF duration increased significantly only with PV stimulation (9.6±3.7 vs. 1.7±0.7s, p<0.05). During combined sympathovagal stimulation, AF induction was reproducible with burst pacing in the RAA, LAA and PVs in 88%, 100% and 75% of the animals, respectively (50Hz vs. 50Hz+sympathovagal stimulation, p<0.05 for the RAA and LAA), and AF duration increased at all sites. AF duration was >10s in 37.5% of the rabbits (only with vagal or sympathovagal stimulation), and the arrhythmia ceased immediately after cessation of vagal stimulation in two-thirds of these cases.

Figure 5.

Inducibility of atrial fibrillation at different stimulation sites (RAA, LAA and PV) with burst pacing alone or combined with autonomic stimulation (vagal, sympathetic or sympathovagal). ap<0.05. SV: sympathovagal. Other abbreviations as in text.

(0.09MB).

In recent years, animal experiments have sought to elucidate the basic mechanisms underlying the pathophysiology of AF and have considerably improved our understanding of electrophysiological concepts such as simultaneous multiple depolarization waves with variable spatial orientation, rapid focal activity with fibrillatory conduction, and rotors resulting in the identification of areas of fragmented atrial electrical activity. However, the development of in vivo models with preserved autonomic innervations that enable dynamic electrophysiological assessment with different types of ANS activity, and analysis of their relation to AF induction and duration, remains a challenge.

In dogs, microreentry, automaticity and triggered activity have been suggested as mechanisms associated with atrial arrhythmias.28,29 Hayashi et al. used cervical vagal stimulation to induce AF with atrial pacing bursts in anesthetized dogs.30 In recent studies, vagal stimulation was associated with shortening of atrial ERPs and AF induction in dogs.17,31 These vagal effects and AF inducibility can be modified by epicardial ablation of the ganglionated plexi.12,15 In patients undergoing AF catheter ablation, recurrence of AF is reduced if the procedure includes selective vagal denervation.32,33 The idea that cardiac denervation can play a role in AF prevention is a complex but promising approach that requires further investigation.

In previous studies, sympathetic stimulation has been less effective than vagal stimulation in influencing atrial refractoriness and AF induction.11,34 However, in a canine model using chronic rapid pacing, sympathetic stimulation induced repetitive activity in the PVs and contributed to initiation of AF.35 Combined sympathetic and parasympathetic stimulation can also affect electrophysiological properties and increase vulnerability for AF in dogs.16,35 The electrophysiological effects of autonomic reflex activity in the atria and PVs may thus play an important role in the pathophysiology of AF.

The sequence of mechanisms responsible for spontaneous termination of AF has yet to be clarified. It may be that resolution (or reversal) of the changes in autonomic tone that accompany initiation of AF also contribute to its termination. Tomita et al. demonstrated that the changes in ANS activity that precede initiation of AF episodes normalize immediately after termination of AF.8 In the present study, using an intact rabbit heart model that has little propensity for AF in baseline conditions, we confirmed that sympathovagal interaction has significant effects on conduction and refractoriness of the atria and PVs and contributes to AF inducibility and duration. Furthermore, vagal activity appeared to exert a direct influence on termination of AF, since in two-thirds of the animals with longer AF, the arrhythmia ceased immediately after cessation of vagal stimulation.

Evidence for the role of the ANS in the genesis of paroxysmal AF has been provided by clinical studies and experimental models. Despite conflicting results of analyses of HR variability in the period preceding spontaneous initiation of AF, recurrent paroxysmal episodes appear to be related to dynamic variations in autonomic tone, possibly the result of synergy between adrenergic flow and vagal activity.5,18,36,37 Further studies are required to elucidate the effect of innervation and the interaction of sympathovagal reflex activity on the complex dynamics of the functional substrate of atrial arrhythmogenesis.

Assessment of refractoriness was based on only three sites (RAA, LAA and PVs). The lengthy and detailed process involved in calculating ERPs in baseline conditions and with different autonomic stimulation would have been impractical, particularly in terms of maintaining the stability of the in vivo preparation. However, the technique of programmed electrical stimulation with an extrastimulus has been widely used to study refractoriness. Future studies using different methods may be warranted to confirm these results. Nevertheless, in the present study electrophysiological changes, and their effect on AF induction, occurred immediately after autonomic stimulation, suggesting a clear relationship between vagal and sympathetic modulation and the results obtained.

Finally, although AF induction reflects vulnerability, the difficulty in sustaining the arrhythmia in the small atrial area of the rabbit heart makes it difficult to study the electrophysiological mechanisms underlying sustained AF with this model.