INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE HEART is innervated by the parasympathetic nervous system along two separated pathways. The effect of right vagal stimulation on heart rate is much larger than identical left vagal stimulation (16). Vagal stimulation has minor effects on ventricular refractoriness (22). However, recent research showed a more complicated selectivity (26). Extensive overlap is demonstrated for both right and left vagal projections to the sinoatrial (SA) node and atrioventricular (AV) node regions, with a tendency for predominance of the right vagus at the SA node and the left vagus at the AV node (1). Interconnection between the two branches takes place in the intracardiac ganglia (6, 8). Under normal physiological conditions, the bilateral influence will adapt heart rate and AV node conductivity to the prevailing conditions (15). In case of increased vagal activity, the heart rate slows down and AV conduction velocity diminishes. In the present study, we investigate the effect of differences in timing of vagal activity between the left- and right-sided vagal branches on irregularities in AV conduction.

In the intact animal, vagal pulses arrive in clusters. Clusters of stimuli that originate from medullary neurons will reach target organs via left- and right-sided branches at different times due to differences in conduction velocities of the respective axons and differences in transmission time in the ganglia (28). Differences in anatomy and metabolic state and pathological developments in the autonomic nervous system, especially in the vagal nerve, will enhance these differences in function (11). These effects will increase in importance at increasing vagal activity. The influence of that particular vagal branch is provided by the electrophysiological status of the target cells and thus by the moment at which the stimulation cluster reaches its target (13).

O'Toole et al. (23) showed that, under bilateral continuous vagal stimulation, AV block persisted during stimulation. After left parasympathectomy, this blocking disappeared although right vagal stimulation was still present. Randall et al. (25) showed that, after right parasympathectomy in conscious animals, marked AV blocks occurred without slowing SA rhythmicity when parasympathetic activity was increased.

It is important at what moment in the electrophysiological status of the AV nodal cells the stimulation pulse arrives. In a period of some milliseconds, a large change in conduction velocity will take place (17). Differences between left and right vagal activity can also be due to anatomical differences. The vagal efferent neurons originate from higher nuclei in the brain. In a study on the autonomic nuclei in the brain stem [left and right nuclei ambiguus (NA)] projecting to the cervical vagal efferents, Thompson et al. (33) demonstrated that slowing of the atrial rate was greatest with right NA stimulation. Stimulation of the left NA produced atrioventricular blockade when both vagal nerves were intact. When the left vagal nerve was cut, AV blockade disappeared. These experiments show that the left vagal branch predominantly controls the ventricular electrophysiological properties as a function of vagal stimulation frequency, intensity of the stimulation (stimuli/cluster), and moment of arrival at target (phase response). When the left vagal branch dominates the right branch by increased activity (asymmetry) and decreased conduction velocity (asynchrony), the changes of irregularities in heart rate and in AV conduction velocity increase (17, 21, 23). So asynchronicity, but also asymmetry, between left and right vagal pulses changes the regularity of heart rate.

The influence of vagal stimulation on AV conduction also depends on heart rate. Generally, increased vagal stimulation decreases heart rate and increases AV conduction time (AVCT). However, during atrial pacing, AVCT increases at increasing pacing rate (increasing heart rate). When bilateral vagal stimulation occurred (regardless of continuous or pulsed stimulation), AVCT increased at a fixed pacing interval (12, 20, 34-36).

In the present study, we hypothesize that irregularities in heart rate occur when the left vagal branch is activated earlier than the right branch. When the right branch is activated earlier than or almost simultaneously with the left branch, no irregularities occur. From the mentioned literature, we decided to test this hypothesis for the following two types of vagal stimulation: 1) continuously, with unpaced heart and 2) pulsed, with paced atrium.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Wistar Kyoto rats (330 ± 35 g) were anesthetized with urethan (1.3 mg/kg, ip). The experiments conformed to the internationally recognized standards for animal research published by the American Physiological Society. Tracheotomy was performed to connect the animal to a respirator to secure constant ventilation. Left and right vagal nerves alongside the carotid arteries were isolated from the surrounding tissue. The nerves were cut, and the distal ends were connected to stimulus electrodes. Care was taken that the surrounding tissue did not make contact with the electrodes. Electrocardiogram (ECG) electrodes were placed on the thorax using a bipolar lead, such that the amplitudes of the P and QRST waves were strikingly apparent.

Two types of vagal stimulation protocols were used. The first protocol consisted of continuous vagal stimulation, without atrial pacing. The second protocol consisted of pulsed vagal stimulation and atrial pacing.

In the first protocol, the frequency of continuous vagal stimulation was chosen at 100 Hz at a stimulus duration of 0.3 ms. When blood pressure is raised to higher levels (5), vagal pulses as observed in bursts during a cardiac cycle appear at high frequencies. For that reason, in rat hearts showing high heart rates, we assumed a stimulation frequency during the burst of 100 Hz. To keep the effect of each pulse transient as much as possible, we selected the duration at 0.3 ms. The amplitude of the vagal stimulation pulses was selected as follows: at slowly increasing stimulation amplitude, the level of incidental AV blocking was detected. The amplitude used for further analysis of asynchrony was set at 80% of the amplitude that induced incidental AV blocking. The amplitude differed slightly among animals due to differences in connective resistance.

Asynchrony was introduced by setting the delay between the right and left stimulation unit. We used six delay values between the right and left stimulus (right/left delay): 0, 2, 4, 6, 8, and 10 ms. Because interval duration was 10 ms between two consecutive stimuli, a delay between right and left up to 5 ms is said to be early right vagal stimulation, whereas a delay between 5 and 10 ms is said to be in favor of the left stimulus (early left vagal stimulation). A delay of 10 ms is in fact the same as a delay of 0 ms.

In the second protocol, the atrium was paced. The thorax was opened, and a small double electrode was connected to the right atrium in the region of the SA node. The atrium was paced at different intervals, between 140 and 220 ms. Duration of the stimulus pulse was 0.5 ms. The vagal stimulation was set to pulse at 40 Hz. Only three stimuli of 0.3 ms each were given, starting at 25 ms after the pacing stimulus. Asynchrony between the right and left vagal stimuli was restricted to delaying the pulses 5 ms to each other. For early left vagal stimulation, left vagal stimuli were initiated 5 ms earlier than right stimuli, and for early right vagal stimulation, right vagal stimuli were initiated 5 ms earlier than left stimuli. The amplitude of the vagal stimuli in synchronous stimulation was such that at the longest pacing interval (220 ms) no AV blocking occurred, whereas at the shortest pacing interval (140 ms) some incidental AV blocking was allowed.

The ECG recording and pacing stimuli were sampled at 4 kHz and 12 bits. A software package was developed to detect onset P and onset Q to obtain the P-Q and R-R intervals. R-R interval was calculated from consecutive onsets of the QRS complexes. P-Q duration was obtained from onset P until onset Q as representative for AVCT. To avoid transient response of the onset of stimulation, sampling was started 20 s after onset of the vagal stimulation. The sample period hereafter was also 20 s. A period of 20 s elapsed before another delay between right and left stimulus was applied. A protocol consisted of six different delay values (0, 2, 4, 6, 8, 0 ms or 0, 8, 6, 4, 2, 0 ms to avoid cumulative effects). The duration of a protocol was therefore restricted to ~6 min.

In the pulsed stimulation protocol, five different atrial pacing intervals were chosen (220, 200, 180, 160, and 140 ms). A period of 60 s was sampled in which no vagal stimulation was applied. Hereafter, vagal stimulation was applied for 120 s. The last 60 s were sampled. The delay was set at 0, at 5 ms in favor of the right stimulus (early right stimulation), and at 5 ms in favor of the left stimulus (early left stimulation). At the end, 60 s were recorded during the longest pacing interval (220 ms) without vagal stimulation to test the deterioration of the preparation. When the mean P-Q interval differed >10% from the mean P-Q interval at the beginning, the data were discarded. The sampled signals were stored on disk for further evaluation.

As a result of the not normally distributed R-R and P-Q values, histograms were calculated to quantify the irregularity in R-R interval and P-Q interval. The fraction of the different distributions in the histogram was calculated relative to the total distribution. The fractions of the distribution of the peaks, different from the one at the undisturbed interval, are used to quantify the irregularity due to vagal stimulation. Student's t-test was used to set significancy at P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In each animal, a decrease in R-R interval was recorded after bilateral vagotomy. Although this change was often small and the variance in R-R was influenced by ventilation, the decrease in R-R was significant: before vagotomy, 155 ± 10 ms vs. after vagotomy, 150 ± 7 ms (n = 12). Synchronous vagal stimulation (at 80%, see METHODS) increased the mean R-R interval (173 ± 8 ms; n = 12).

Continuous stimulation. Figure 1A shows the changes in R-R interval during asynchronous stimulation at different delays between right and left vagal stimulus (duration 20 s) for one particular animal. The R-R interval after vagotomy without vagal stimulation is also shown. Different delays between right and left stimulus were used. The synchronously stimulated nerves introduced an increase in R-R interval (not shown). When the synchronous nature of the stimuli was changed, the mean R-R interval remained unchanged, although the number of beats with prolonged R-R interval increased when the delay between right and left was larger than 5 ms (which in fact means that the left vagal nerve is stimulated 10  x ms earlier than the right nerve). The degree to which AV nodal blockade was produced increased when the left-sided nerve was stimulated earlier than the right one. In Fig. 1A for early left stimulation (8 ms), mean R-R interval was 187 ± 57 ms, whereas mean P-P interval was 150 ± 4 ms.

View larger version (3113K): [in this window] [in a new window]   Fig. 1.   A: tachograms of 3 different protocols under continuous stimulation. Delay between right and left vagal stimulus is as follows: , early left (8 ms); , early right (2 ms); , no stimulation. B: histograms of the R-R intervals of the 3 protocols in A. Solid line, no vagal stimulation; broken line, early right vagal stimulation; dotted line, early left vagal stimulation.

Figure 1B shows histograms employed in the three protocols of Fig. 1A. The irregularities are quantified by the fraction of the distribution of the irregular R-R intervals in the histograms. These fractions are plotted versus the delay between right and left stimulus in Fig. 2. It is evident that, when delays longer than 5 ms were used, there was an increased fractional occurrence. The bars represent the SE of the 12 animals used in this part of the study.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Fraction of deviated R-R intervals (%) of the total distribution vs. the delay between right and left stimulus. Stimulation frequency was 1 kHz. Delay between 0 and 5 ms means early right, and delay between 5 and 10 ms means early left vagal stimulation. Bars indicate SE (n = 12 experiments).

Pulsed stimulation. Pulsed stimuli were applied during atrial pacing. At one particular atrial pacing interval (P-P = 180 ms), the duration of the P-Q interval is plotted (Fig. 3) in the case of no vagal stimulation, when right-sided nerves were stimulated 5 ms earlier than left-sided ones, and when left-sided nerves were stimulated 5 ms earlier than right-sided ones.

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Tachogram under atrial pacing conditions. P-Q interval is plotted for (pulsed) early left vagal stimulation (), early right vagal stimulation (), and no vagal stimulation (). Inset: difference in early left (L) and early right (R) vagal stimulation.

The atrial pacing interval is shown at 180 ms (Fig. 3). The onset of the P wave after stimulation was not detectable reliably. When left-sided nerves were stimulated earlier than right-sided ones, the P-Q interval increased gradually until total AV block occurred. The value of the P-Q interval in the beat after the AV block is the same as the P-Q value at early right vagal stimulation.

In Fig. 4A, for one animal, the mean values of the P-Q interval at different vagal stimulation conditions are plotted versus the pacing interval. The bars represent the variance. When no vagal stimulation is applied, the P-Q interval is relatively short and increases significantly with decreasing pacing interval (increasing pacing frequency). At vagal stimulation, P-Q lengthened, and at increased duration (>80 ms) irregularities in AV conduction (predominantly AV blocking) appeared. In the case of AV blocking, as shown in Fig. 3, the P-Q interval was adopted identical to the P-P interval (=180 ms). At any pacing interval, the P-Q interval is longer in the case of earlier left vagal stimulation. In the case of three protocols, no variance is plotted due to the not normally distributed values of the P-Q interval. The histograms of two of these protocols (early left vagal stimulation) are plotted in Fig. 4B, left (160 ms pacing) and right (180 ms pacing).

View larger version (817K): [in this window] [in a new window]   Fig. 4.   A: relationship between P-Q interval under different vagal stimulation conditions and atrial pacing interval. , Early left; , early right; , synchronous; x, no stimulation. Bars indicate SE. No. 1 indicates no SE due to abnormal distribution. B: histograms of the P-Q interval of 2 protocols with early left vagal stimulation in A. Left, pacing at 160 ms; right, pacing at 180 ms.

Calculating the fractions of the deviated P-Q intervals as a measure of irregularity in AV conduction due to vagal stimulation, we plotted the fraction versus the delay between left and right vagal stimulus in Fig. 5. When the delay is positive (5 ms), the left vagal stimuli were applied earlier than the right stimuli. The negative (5 ms) delay indicates that the right stimuli were applied earlier than the left stimuli. At zero delay, the fraction at synchronous stimulation is plotted together with the fraction of the distribution without vagal stimulation. The bars represent the variance due to interanimal differences. All mean values at early left vagal stimulation are significantly (P < 0.05) larger than other types of stimulation. There was no significant difference between the values for early right or synchronous vagal stimulation.

View larger version (10K): [in this window] [in a new window]   Fig. 5.   Fraction of deviated P-Q intervals (%) of the total distribution vs. delay between pulsed left and right vagal stimulus. , Early left (n = 43); , early right (n = 35); , synchronous (n = 40); *, no vagal stimulation (n = 32). Bars indicate SE.

Data from both stimulation protocols are summarized in Table 1.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

From these results, we conclude that greater changes in heart rate and AV nodal conduction occur when parasympathetic efferent preganglionic axons in the two cranial vagi are stimulated asynchronously, especially when the left-sided cervical vagus is stimulated a few milliseconds earlier than the right one.

In this study, we showed that early left vagal stimulation induced a prolongation of the P-Q interval. This prolongation is assumed to be caused by an increase in AVCT, which may be due to momentarily induced hyperpolarization of AV node cells (10, 19). This was not the case when the right vagal stimulus arrived earlier than the left one. Phase dependency, relative to the cardiac cycle of a single stimulus, was reported by Jalife and Michaels (14) for the pacemaker activity of the SA node. They showed the existence of a small time window during which a single (right) vagal pulse exerts its largest differences in pacemaker activity. For the AV node, Martin (17) showed a comparable small window in which the largest differences in AVCT occur. AVCT is also dependent on the R-R interval on a beat-to-beat basis (12, 20, 34). We therefore used two types of stimulation, continuous vagal stimulation (without atrial pacing) and pulsed vagal stimulation (with atrial pacing). Under both types of vagal stimulation protocols employed in the present investigations, AV blockade appeared most frequently when the left-sided nerves were stimulated first. It is already reported (23, 25) that, in the case of unilateral stimulation of the cervical vagal nerves and direct stimulation of ganglia on the heart (4), left-sided stimulation induces lengthening of AVCT, up to blocking for several beats. The present study adds to the finding that bilateral stimulation can do the same if the left side is favored above the right side, especially in a small time window. No irregularity is found in synchronous conditions or when the right side is stimulated earlier than the left side.

In several studies, continuous stimulation is used to characterize vagal influence (10, 16, 21). Only at high blood pressure does vagal efferent neuronal activity display continuous activity patterns. Cerati and Schwartz (5) showed electroneurograms of single vagal efferent neuron activity at different levels of blood pressure in the cat. At high blood pressure, the frequency of the (pulsed) neuron activity is ~100 Hz. To study the effects of stimulation of the cardiac ganglia on canine atria, Butler et al. (4) also used high-frequency stimulation (200 Hz). We have chosen continuous stimulation at 100 Hz to exclude a number of parameters that are also involved in the reaction of the cardiac system on stimulation. In this experiment, we did not choose the moment at which the first stimulus arrives in the (varying) cardiac cycle, the duration of stimuli in each vagal pulse, or the frequency of the stimuli in the right and left vagal pulses. The amplitude of the pulses was such that, under synchronous conditions, no or almost no irregularities were observed. AVCT increased under all stimulated conditions compared with normal (before vagotomy). The majority of the irregularities observed after early left vagal activity were AV blocking.

Because the R-R interval also has a direct influence on AVCT, we used the pacing protocol to fix this influence at a certain level. Atrial pacing frequencies were chosen within the normal physiological range for the rat. In Fig. 4A, we confirm the findings of Nayebpour et al. (20) and Wallick et al. (34) that AVCT increases at decreasing pacing interval and that increased vagal stimulation enhances this increase. In the present study, threshold for AVCT occurred, after which blockade occurred (Figs. 3 and 4A). Slowly increasing AVCT is observed in early left vagal stimulation up to a threshold value, as demonstrated in Fig. 3. Above this value, the AV node is blocked singularly. AVCT after the block drops to the same value as under synchronous vagal stimulation and starts to increase again. Zhao and Billette (37) showed that, because of a decreasing nodal refractory period (after fatigue), Wenckebach periodicity was generated. In our study, this periodicity occurred only during early left vagal stimulation. The increase in AVCT developed in a block during synchronous vagal stimulation if pacing rate increased (Fig. 4A: synchronous at 160 ms). In a study on the origin of neurons influencing the AV node conduction, Massari et al. (18) retrogradely labeled neurons in the AV node ganglion on the heart. They found that mainly cells in the left part of the rostral ventrolateral NA were labeled. Activation of the nuclei by glutamate showed an increased AVCT and occasional heart blocks without changing heart rate. They conclude that anatomically separated and functionally selective preganglionic vagal neurons in the NA independently control AV conduction and heart rate. Butler et al. (4) used high-frequency electrical stimulation in selected parts of the cardiac ganglia. After administration of atropine, stimulation of the right atrial ganglionated plexus but not the left atrial ganglionated plexus induced tachycardia. Hence, the cardiac ganglia contain next to parasympathetic also sympathetic efferent neurons. In our study, we are especially interested in the parasympathetic influences. Thompson et al. (33) showed that, at the level of the NA, already the functional left-right differentiation of the vagal nerves can be observed. Stimulation of the left NA produces AV blocks and sinus bradycardia. This was prevented by left vagotomy. Right vagotomy only prevented sinus bradycardia after stimulation of the right NA. Chiou et al. (6) showed that, between the NA and the two atrial ganglia, another "third fat pad" is located in which efferent vagal neurons travel to the ganglia on the atria. From these anatomical differences between left and right vagal origin and the differences in pathway, it is easy to understand that there are differences between the two vagal branches in all aspects of nerve activity: conduction velocity, firing frequency, timing. The present study shows that asynchronous stimulation of the efferent vagal neurons induces changes in heart rate and AVCT. Early left vagal stimulation induced more irregularities in AVCT (recorded as P-Q interval) than early right vagal stimulation.

We performed this study in relation to research on sudden death in highly trained athletes at rest. Trained athletes show decreased heart rates at rest due to increased vagal activity (30), even lower than 35 beats/min at rest, often accompanied with arrhythmia. Also, the decreased sympathetic activity and cardiac cellular adaptation (changes in intrinsic heart rate; see Refs. 3, 29, and 31) will contribute to the bradycardia. It is questioned if these low heart rates in the long run are a healthy adaptation of the cardiac regulation. Ector et al. (7) investigated 16 intensive athletes with syncope or Adam Stokes syndrome at rest and after exercise. In about one-half of the cases, a pacemaker had to be implanted to relieve the complaints. In almost all of the other athletes, the problems disappeared after stopping intensive exercise. Asystole and/or ventricular fibrillation can be the result of increased AVCT, long Q-T intervals, and late potentials, mentioned as lethal causes (2, 27, 32).

After severe exercise, hyperemic blood supply does not equally recover all cells, keeping cellular metabolic conditions at different levels. The above-mentioned neural conduction parameters will be affected due to this unequally distributed recovery. Therefore, asynchronous activation is a physiological phenomenon. We showed that, when the delay between the two stimuli is only a few milliseconds, already irregularities can occur. Under normal conditions, the baroreflex has to correct the decreased blood pressure by eliminating vagal activity. Frederiks et al. (9) showed that, in trained conditions, baroreflex sensitivity is decreased. Hence, in trained conditions with decreased baroreflex sensitivity, increased vagal activity at rest suddenly can become troublesome.