Electrophysiological effects of right and left vagal nerve stimulation on the ventricular myocardium

Abstract

Vagal nerve stimulation (VNS) has been proposed as a cardioprotective intervention. However, regional ventricular electrophysiological effects of VNS are not well characterized. The purpose of this study was to evaluate effects of right and left VNS on electrophysiological properties of the ventricles and hemodynamic parameters. In Yorkshire pigs, a 56-electrode sock was used for epicardial (n = 12) activation recovery interval (ARI) recordings and a 64-electrode catheter for endocardial (n = 9) ARI recordings at baseline and during VNS. Hemodynamic recordings were obtained using a conductance catheter. Right and left VNS decreased heart rate (84 ± 5 to 71 ± 5 beats/min and 84 ± 4 to 73 ± 5 beats/min), left ventricular pressure (89 ± 9 to 77 ± 9 mmHg and 91 ± 9 to 83 ± 9 mmHg), and dP/dt_max (1,660 ± 154 to 1,490 ± 160 mmHg/s and 1,595 ± 155 to 1,416 ± 134 mmHg/s) and prolonged ARI (327 ± 18 to 350 ± 23 ms and 327 ± 16 to 347 ± 21 ms, P < 0.05 vs. baseline for all parameters and P = not significant for right VNS vs. left VNS). No anterior-posterior-lateral regional differences in the prolongation of ARI during right or left VNS were found. However, endocardial ARI prolonged more than epicardial ARI, and apical ARI prolonged more than basal ARI during both right and left VNS. Changes in dP/dt_max showed the strongest correlation with ventricular ARI effects (R² = 0.81, P < 0.0001) than either heart rate (R² = 0.58, P < 0.01) or left ventricular pressure (R² = 0.52, P < 0.05). Therefore, right and left VNS have similar effects on ventricular ARI, in contrast to sympathetic stimulation, which shows regional differences. The decrease in inotropy correlates best with ventricular electrophysiological effects.

the autonomic nervous system plays a significant role in the genesis and persistence of ventricular arrhythmias (54, 59). Sympathetic activation is proarrhythmic (16, 32, 53), whereas parasympathetic activation is thought to be cardioprotective (17, 31). The vagal nerve trunk provides important cardiomotor efferent fibers to the heart and also carries afferent signals from the heart. Vagal nerve stimulation (VNS) has been shown to decrease infarct size (48), reduce the ventricular fibrillation (VF) threshold (39), and decrease the incidence of ventricular arrhythmias and mortality during ischemia (13, 27, 38, 52). Furthermore, a preserved parasympathetic reflex has been reported to be protective during myocardial infarction (46). Stimulation of the right vagal nerve (RVN) has shown benefits in a series of patients with cardiomyopathy and is undergoing evaluation in clinical trials (20, 47). The mechanisms of the antiarrhythmic effects of VNS are less clear and are thought to be multifactorial, with a decrease in heart rate (HR) (15), release of nitric oxide (9), and antagonism of the sympathetic nervous system all thought to play a role (8, 30, 49).

Modulation of repolarization by sympathetic nerve stimulation has been well characterized (1, 25, 41, 55, 58). However, the effects of parasympathetic regulation of ventricular repolarization have not been extensively studied. Indeed, parasympathetic innervation of the ventricular myocardium was considered minimal for many years. Histological studies have now confirmed evidence of abundant ventricular parasympathetic innervation (24, 26, 51). Previous studies of ventricular repolarization during VNS were limited to using a small number of electrodes (35), extrastimulus pacing (33), or VF intervals (41): techniques that can alter autonomic tone (23) and do not provide detailed spatial data. Furthermore, differences between the RVN and left vagal nerve (LVN) on regional repolarization remain unclear and may be important given the presence of disease processes that can affect certain areas of the heart to different degrees. For unilateral stimulation, lack of laterality would also be significant in allowing similar effects of VNS from either side.

The aim of this study was to assess the effects of right and left VNS on cardiac function and global and regional ventricular repolarization of the endocardium and epicardium.

METHODS

Animal protocol.

All animal experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the University of California-Los Angeles Institutional Chancellor's Animal Research Committee.

Yorkshire pigs (n = 12, 20–50 kg) were sedated with telazol (8–10 mg/kg im), intubated, and ventilated. General anesthesia was maintained with isoflurane (1–1.5%) and intravenous boluses of fentanyl (2–4 μg/kg). A median sternotomy was performed to expose the heart, and the cervical right and left vagosympathetic trunks were isolated. After the completion of surgery, the anesthesia was switched to α-chloralose (10 mg·kg⁻¹·h⁻¹ iv infusion). HR was monitored via electrocardiogram recordings. The right femoral artery was catheterized to monitor systemic blood pressure. Hourly arterial blood gases were obtained, and tidal volume was adjusted and/or infusions of sodium bicarbonate were given to maintain acid-base homeostasis.

Hemodynamic assessment.

Pressure-volume loops, left ventricular (LV) end-systolic pressure (LVESP), LV end-diastolic pressure (LVEDP), dP/dt_max, dP/dt_min, and the time constant of isovolumic relaxation (τ) were obtained using a 12-pole conductance pressure catheter in the LV (n = 8) connected to a MPVS Ultra Pressure Volume Loop System (Millar Instruments, Houston, TX). τ was calculated using the Weiss method (56) from the pressure-volume loop as a parameter describing the time course of the exponential decay in LV pressure during isovolumic relaxation. The following equation was used to calculate τ: P(t) = A × exp(−t/τ), where P is pressure, A is a constant, referring to the slope of the linear relationship between pressure and exp(−t/τ), and t is time.

VNS.

The RVN and LVN were stimulated separately using bipolar electrodes (Cyberonics, Houston, TX) connected to a Grass S88 Stimulator (Grass Technologies, Warwick, RI). Square stimulation pulses were delivered at 0.5–1.0 ms in duration and 10–20 Hz in frequency. To determine the threshold current, the current was increased starting from 0.2 mA by 0.2-mA intervals until either a 10–20% drop in HR (the percentage of 10% vs. 20% was picked based on the animals' baseline HR so that severe bradycardia or a significant drop in blood pressure was avoided) for both right and left VNS. After the threshold current was reached, the stimulation current was decreased by 0.1 mA to confirm that the threshold was accurate. If a stimulation frequency of 20 Hz led to complete heart block or asystole with right VNS, left VNS, or both, 10 Hz was use as the stimulation frequency for both right and left VNS. Right and left VNS was performed at 1.2 times threshold current for 15 s followed by a 15-min stabilization period to allow for activation recovery interval (ARI) and hemodynamic parameters to return to baseline. To account for any HR effects on ARI, atrial pacing at baseline HR was performed in four animals during right and left VNS. In addition, subthreshold VNS at a level just below the HR response was performed in one animal, and the ARI was analyzed.

ARI recordings.

A 56-electrode nylon sock was placed around the heart (n = 12), and a 64-electrode basket catheter (n = 9) was placed into LV via the left carotid artery. Unipolar electrograms (EGMs) were obtained using a Prucka CardioLab System (GE Healthcare, Fairfield, CT) and a custom-made, 128-channel multiplexor. EGMs were band-pass filtered between 0.05 and 500 Hz. A minimum of 10 EGMs was analyzed from each electrode before and during stimulation. An electrofield electroanatomic mapping system (NavX, St. Jude Medical, St. Paul, MN) was used to assess the basket catheter position. At the end of the experiment, direct incision of the LV was performed to confirm endocardial electrode locations. ARIs were measured and calculated using customized software (ScalDyn, University of Utah, Salt Lake City, UT) as previously described (55). ARI has been shown to correlate well with action potential duration (APD) and allows for multiple simultaneous measurements. Furthermore, changes in ARI at a given site during an intervention correlate well with changes in APD measured from microelectrodes (21, 37). This method allows for measurements of local APD without extrastimulus delivery or induction of VF, methods that can alter autonomic tone. Global dispersion of repolarization (DOR) was calculated using the variance of all mean ARIs recorded in a specific region. EGMs with flattened T waves, fractionation, or noise were excluded.

Regional ARI analysis.

For purposes of this report, anterior refers to the ventral aspect and posterior refers to the dorsal aspect of the animal. Mean ARI in the following regions was analyzed: LV anterior, lateral, and posterior, right ventricular (RV) anterior, lateral, and posterior, RV outflow tract, and LV base and apex. The median number of electrodes in each region was four (range: 3–7). For the assessment of transmural (endocardial vs. epicardial) differences in ARI, electrodes directly across from each other on the anterior, lateral, and posterior aspects of the mid-LV were used.

Three-dimensional sock ARI data were projected onto two-dimensional polar maps using publicly available software (Map3D, University of Utah; http://www.sci.utah.edu/cibc/software/107-map3d.html).

Statistical analysis.

For comparison of continuous variables, a Wilcoxon rank-sum test or Wilcoxon signed-rank test was used. For regional analysis, means and variances in ARI were compared using parametric repeated-measures ANOVA. The P value for a particular pairwise, mean, standardized ARI comparison was considered significant only if the corresponding overall F-statistic was significant. Dispersion in ARI was defined as the variance in the mean ARI recorded from all the electrodes over a given region. To account for baseline differences, the percent change in ARI was also compared. Given the range of values for DOR, the log mean difference was used for statistical analysis of regional differences. Data are presented as means ± SE. SAS 9.1 was used for statistical analysis. P values of <0.05 were considered statistically significant. Correlations between hemodynamics parameters and ARI changes were performed using the Pearson correlation test.

RESULTS

The right VNS current was 4.0 ± 0.8 mA, and the left VNS current was 4.2 ± 0.8 mA (P = 0.6 for right vs. left VNS current). Three animals developed complete heart block with right VNS, and four animals developed complete heart block with left VNS.

Hemodynamic responses to stimulation.

The effects of right and left VNS on hemodynamic parameters are shown in Table 1. Both right and left VNS significantly decreased dP/dt_max and LVESP and increased dP/dt_min and τ. There were no statistically significant differences between right and left VNS on hemodynamic parameters, and, therefore, there was no laterality on the hemodynamic effects of right versus left VNS (Table 1). The increase in dP/dt_min and τ by both right and left VNS suggested a worsening of diastolic function by VNS. The PR interval increased with right VNS from 114 ± 4 to 137 ± 5 ms and increased with left VNS from 112 ± 4 to 140 ± 4 ms. There was no difference in PR interval prolongation between right versus left VNS (P = 0.4), suggesting that both the RVN and LVN provided innervation to the atrioventricular (AV) node.

Table 1.

Responses to RVN and LVN stimulation compared with BL

	RVN		LVN
	BL	During stimulation	BL	During stimulation
Heart rate, beats/min	84 ± 5	71 ± 5*	84 ± 4	73 ± 5*
Systolic blood pressure, mmHg	120 ± 7	110 ± 8*	119 ± 8	111 ± 8*
Diastolic blood pressure, mmHg	83 ± 8	73 ± 8*	83 ± 9	74 ± 9*
End-systolic pressure, mmHg	89 ± 9	77 ± 9*	91 ± 9	83 ± 9*
End-diastolic pressure, mmHg	4 ± 1	6 ± 1	6 ± 2	6 ± 2
dP/dtmax, mmHg/s	1,660 ± 154	1,490 ± 160*	1,595 ± 155	1,416 ± 135*
dP/dtmin, mmHg/s	−1,511 ± 211	−1,090 ± 208*	−1,520 ± 213	−1,161 ± 231*
Time constant of isovolumic relaxation, ms	41 ± 4	50 ± 7*	41 ± 3	50 ± 5*
PR interval, ms	114 ± 4	137 ± 5*	112 ± 4	140 ± 4*

RVN, right vagal nerve; LVN, left vagal nerve.

^*P < 0.05 vs. baseline (BL).

Effects of right and left VNS on epicardial ARI and DOR.

Right and left VNS prolonged global epicardial ARI from 327 ± 18 to 350 ± 23 ms (7 ± 2%, P < 0.05 vs. baseline) and from 327 ± 16 to 347 ± 21 ms (6 ± 2%, P < 0.05 vs. baseline), respectively, with no significant differences between right or left VNS (P = 0.4). Epicardial DOR was increased by both right and left VNS from 354 ± 53 to 512 ± 121 ms² and from 338 ± 57 to 476 ± 110 ms², respectively (P < 0.05 vs. baseline for both conditions; Fig. 1). With respect to anterior, posterior, and lateral changes on the RV and LV, each region demonstrated ARI prolongation (P < 0.05) compared with baseline (Fig. 2) without statistically significant differences in prolongation between these regions. Regional changes in DOR at baseline and during right and left VNS are shown in Table 2. These changes from baseline were not statistically significant. With regard to regional comparisons, no significant anterior, lateral, or posterior differences in DOR during right or left VNS were observed on the epicardium (Fig. 3). Detailed regional DOR comparisons are provided in Tables 3–6.

Fig. 1.

Global epicardial activation recovery interval (ARI) responses to right vagal nerve (RVN) stimulation (RVNS) and left vagal nerve (LVN) stimulation (LVNS). Top: images showing the isolation of the vagal nerves, bipolar electrodes used for vagal nerve stimulation (VNS probe), and the sock electrode on the heart. Locations of the electrodes used to create ARI polar maps are marked. RV, right ventricle; LV, left ventricle; LAD, left anterior descending artery; BL, baseline; Stim, during stimulation. Bottom: mean epicardial ARI (left) and dispersion of repolarization (DOR; right) showing significant differences during stimulation (Stim) compared with baseline (BL) but no significant differences between RVNS and LVNS. Data are means ± SE. *P < 0.01 vs. BL.

Fig. 2.

Regional epicardial ARI responses to RVNS and LVNS. Top: polar maps obtained from one animal at baseline and during RVNS and LVNS. No significant anterior, lateral, or posterior regional differences in responses were found. Bottom: quantified data for all animals. Data are means ± SE. *P < 0.01 vs. BL.

Table 2.

Regional dispersion of repolarization in response to RVN and LVN stimulation

	RVN			LVN
	BL, ms2	During stimulation, ms2	P value vs. BL	BL, ms2	During stimulation, ms2	P value vs. BL
Epicardial regional dispersion of repolarization
LV
Anterior	158 ± 38	148 ± 35	0.82	131 ± 39	113 ± 42	0.50
Lateral	131 ± 44	143 ± 52	0.72	122 ± 42	176 ± 52	0.45
Posterior	111 ± 49	173 ± 79	0.30	94 ± 44	109 ± 69	0.90
RV
Anterior	110 ± 31	160 ± 68	0.10	104 ± 35	145 ± 41	0.14
Lateral	125 ± 51	183 ± 94	0.90	120 ± 56	220 ± 107	0.05
Posterior	156 ± 67	206 ± 87	0.92	143 ± 57	227 ± 96	0.22
Endocardial regional dispersion of repolarization
LV
Anterior	52 ± 32	41 ± 13	0.56	37 ± 20	66 ± 46	0.60
Lateral	88 ± 42	33 ± 17	0.10	103 ± 52	50 ± 20	0.60
Posterior	84 ± 40	88 ± 41	0.74	34 ± 16	60 ± 34	0.20

LV, left ventricle; RV, right ventricle.

Fig. 3.

Epicardial differences in the change in DOR from BL for anterior, lateral, posterior regions of the LV and RV for RVNS and LVNS. Note that no regional differences in DOR were found. Data are means ± SE.

Table 3.

Comparison of the differences in the change of DOR from BL between regions during right VNS

Region	Versus Region	Log Mean Difference	SE of the Difference	P Value
LV endocardium
Anterior	Lateral	0.98	0.61	0.11
Anterior	Posterior	0.11	0.61	0.86
Anterior	Septum	0.34	0.61	0.58
Lateral	Posterior	−0.87	0.61	0.16
Lateral	Septum	−0.64	0.61	0.30
Posterior	Septum	0.23	0.61	0.70
LV epicardium
Anterior	Lateral	0.20	0.49	0.68
Anterior	Posterior	−0.28	0.49	0.56
Lateral	Posterior	−0.49	0.49	0.32
RV epicardium
Anterior	Lateral	0.58	0.49	0.24
Anterior	Posterior	0.61	0.49	0.24
Anterior	RVOT	−0.02	0.50	0.98
Lateral	Posterior	0.03	0.49	0.96
Lateral	RVOT	−0.60	0.50	0.24
Posterior	RVOT	−0.63	0.50	0.22

DOR, dispersion of repolarization; RVOT, RV outflow tract. The log of the mean difference was used given the range of dispersion observed between animals.

Table 6.

Epicardial versus endocardial regional DOR comparisons from baseline during LVN stimulation

Region	Versus Region	Log Mean Difference	SE of the Difference	P Value
Endocardial LV anterior	Epicardial LV anterior	0.43	0.52	0.41
Endocardial LV anterior	Epicardial RV anterior	−0.26	0.52	0.61
Epicardial LV anterior	Epicardial RV anterior	−0.69	0.46	0.14
Endocardial LV lateral	Epicardial LV lateral	−0.46	0.52	0.38
Endocardial LV lateral	Epicardial RV lateral	−0.86	0.52	0.10
Epicardial LV lateral	Endocardial RV lateral	−0.40	0.46	0.39
Endocardial LV posterior	Epicardial LV posterior	0.59	0.55	0.29
Endocardial LV posterior	Epicardial RV posterior	0.21	0.55	0.70
Epicardial LV posterior	Epicardial RV posterior	−0.37	0.46	0.42

Table 4.

Epicardial versus endocardial regional DOR comparisons from BL during RVN stimulation

Region	Versus Region	Log Mean Difference	SE of the Difference	P Value
Endocardial LV anterior	Epicardial LV anterior	0.17	0.56	0.76
Endocardial LV anterior	Epicardial RV anterior	−0.32	0.56	0.56
Epicardial LV anterior	Epicardial RV anterior	−0.50	0.49	0.31
Endocardial LV lateral	Epicardial LV lateral	−0.60	0.56	0.28
Endocardial LV lateral	Epicardial RV lateral	−0.72	0.56	0.20
Epicardial LV lateral	Endocardial RV lateral	−0.11	0.49	0.82
Endocardial LV posterior	Epicardial LV posterior	−0.22	0.56	0.69
Endocardial LV posterior	Epicardial RV posterior	0.18	0.56	0.76
Epicardial LV posterior	Epicardial RV posterior	0.40	0.49	0.42

Table 5.

Comparison of the differences in the change of DOR from BL between regions during LVN stimulation

Region	Versus Region	Log Mean Difference	SE of the Difference	P Value
LV endocardium
Anterior	Lateral	0.43	0.58	0.46
Anterior	Posterior	−0.40	0.60	0.51
Anterior	Septum	0.61	0.58	0.30
Lateral	Posterior	−0.83	0.60	0.17
Lateral	Septum	0.18	0.58	0.76
Posterior	Septum	1.00	0.60	0.099
LV epicardium
Anterior	Lateral	−0.46	0.46	0.32
Anterior	Posterior	−0.24	0.46	0.60
Lateral	Posterior	0.22	0.46	0.63
RV epicardium
Anterior	Lateral	−0.17	0.46	0.72
Anterior	Posterior	0.08	0.46	0.86
Anterior	RVOT	0.00	0.49	0.99
Lateral	Posterior	0.25	0.46	0.59
Lateral	RVOT	0.17	0.49	0.73
Posterior	RVOT	−0.08	0.49	0.86

The log of the mean difference was used given the range of dispersion observed between animals.

When apicobasal differences were compared, the apex showed a slightly greater prolongation of ARI than the base with both right and left VNS from 339 ± 19 to 366 ± 25 ms and from 339 ± 18 to 363 ± 23 ms, respectively (P < 0.05 vs. baseline for both right and left VNS; Fig. 4). ARI at the base increased from 322 ± 17 to 342 ± 21 ms with right VNS and from 323 ± 16 to 341 ± 20 ms with left VNS. The direction of repolarization, however, was maintained during stimulation as the base of heart had a shorter ARI compared with the apex at baseline before both right and left VNS (P < 0.05 for baseline apex vs. base mean ARI; Fig. 4).

Fig. 4.

Apicobasal differences in response to RVNS and LVNS. RVNS and LVNS had greater effects on the apex than the base without changing the direction of repolarization, as the base of the heart had a shorter ARI at baseline. The effects of RVNS versus LVNS on the apex and base were similar, again showing no laterality in response. Data are means ± SE.

Among the hemodynamic parameters, the effects of VNS on HR, dP/dt_max, and LVESP had the highest correlation with the increase in ventricular ARI (R² = −0.58 for HR, P = 0.002; R² = 0.81, P < 0.0001 for dP/dt_max; and R² = −0.52, P = 0.04 for LVESP). Of note, the parameter that correlated most strongly with ARI effects was dP/dt_max (Fig. 5).

Fig. 5.

Correlation between effects of VNS on ARI and hemodynamic parameters. The decrease in dP/dt_max, and not heart rate (HR) or LV end-systolic pressure (LVESP), had the highest correlation with the increase in ARI, also further illustrating the ventricular myocardial effects of VNS. Plots show combined RVNS and LVNS data.

Effects of right and left VNS on endocardial ARI and DOR.

On the endocardium, right and left VNS prolonged ARI from 281 ± 16 to 298 ± 17 ms (P < 0.001) and from 276 ± 18 to 297 ± 19 ms (P = 0.04) versus baseline, respectively. DOR of the LV endocardium changed from 84 ± 22 to 127 ± 28 ms² with right VNS and from 115 ± 35 to 139 ± 49 ms² with left VNS, although these differences from baseline were not statistically significant (P = 0.2 for right VNS and P = 0.4 for left VNS). Similar to the epicardium, no regional differences in the increase in endocardial ARI across the anterior, lateral, and posterior regions of the LV were found during right or left VNS or between right and left VNS (Fig. 6). Similar to the epicardium, no significant anterior, lateral, or posterior differences in DOR during right or left VNS were observed on the endocardium (Fig. 7). Detailed regional DOR comparisons are shown in Tables 3–6.

Fig. 6.

Global and regional mean endocardial ARIs in response to RVNS and LVNS. Note that all endocardial regions (anterior, lateral, posterior, and septal LV) demonstrated a similar increase in ARI with no significant differences in the magnitude of the response for each region. No differences between RVNS versus LVNS were found. Data are means ± SE. *P < 0.01 vs. BL.

Fig. 7.

Endocardial comparisons in the change in DOR from BL for anterior, lateral, and posterior regions of the LV and RV for RVNS and LVNS. Note that no statistically significant regional differences in the change of DOR from BL were found. Data are means ± SE.

Transmural differences in ARI.

During right VNS, mean LV endocardial ARI in the midregion of the LV increased from 281 ± 16 to 298 ± 17 ms (6 ± 2%, P = 0.04). LV epicardial ARI increased from 313 ± 11 to 324 ± 13 ms (4 ± 1%, P < 0.01). During left VNS, LV endocardial ARI prolonged from 276 ± 18 to 297 ± 19 ms (8 ± 2%, P < 0.01). LV epicardial ARI prolonged from 312 ± 12 to 325 ± 12 ms (4 ± 1%, P < 0.01). Therefore, the endocardium showed a greater prolongation of ARI than the epicardium during both right and left VNS (P < 0.01 both conditions; Fig. 8). The greater increase in ARI on the endocardium was consistent across each region.

Fig. 8.

Epicardial and endocardial differences in ARI during RVNS and LVNS. Both RVNS and LVNS increased endocardial ARI more than epicardial ARI. *P < 0.01 vs. BL; #P < 0.05 vs. the epicardium.

Effect of HR on ventricular ARI by VNS.

Atrial pacing was performed during VNS at the same HR as baseline in four animals. Global ARI prolonged from 362 ± 30 to 389 ± 32 ms by right VNS alone. During pacing, ARI still prolonged to 370 ± 30 ms (mean ± SE). Left VNS also demonstrated a prolongation of ARI from 357 ± 23 to 379 ± 30 ms and remained prolonged during atrial pacing (367 ± 24 ms).

The PR interval before right VNS was 123 ± 4 ms, increased to 133 ± 8 ms during VNS, and further increased to 194 ± 10 ms (means ± SE), likely due faster pacing during VNS leading to a greater decremental conduction in the AV node. Before left VNS, the PR interval was 128 ± 8 ms, which increased to 142 ± 13 ms during VNS and further increased to 198 ± 10 ms (means ± SE) with atrial pacing during VNS.

Subthreshold VNS in one animal showed that despite a lack of change in HR, right VNS prolonged ARI from 423 ± 4 to 444 ± 5 ms and left VNS prolonged ARI from 424 ± 4 to 431 ± 4 ms.

DISCUSSION

Major findings.

The major findings of the present study are as follows:

1. Both right and left VNS increased ventricular ARI. This increase was strongly correlated with dP/dt_max. HR and LVESP showed moderate correlations. Parameters of diastolic function were not improved with VNS.

2. During both right and left VNS, epicardial ARI at the apex prolonged more than ARI at the base of the heart. VNS prolonged ARI to a similar degree on the anterior, posterior, and lateral walls of the LV and RV.

3. The endocardium showed a greater prolongation in ARI compared with the epicardium. These effects were similar between right and left VNS.

4. There were no differences between right and left VNS on hemodynamic responses or regional repolarization of the epicardium or endocardium. Therefore, no significant laterality to functional innervation of these nerves was observed.

Global and regional increases in ARI due to right and left VNS.

A histological study (14) has shown that there may be small differences in the parasympathetic innervation of the RV and LV; however, few studies have analyzed functional regional differences in detail. Martin et al. (35) demonstrated that left VNS showed a slightly greater prolongation of the LV epicardial posterior wall (mean difference of 1.2 ms) compared with right VNS in a canine model. Furthermore, they found that VNS did not prolong APD on the anterior epicardial RV (35). In our study, no significant regional differences between right and left VNS on ARI of the LV or RV across anterior, posterior, and lateral regions were noted. A clear increase in endocardial ARI compared with epicardial ARI was observed with both right and left VNS, and both increased ARI at the apex more than the base. Furthermore, a significant prolongation in ARI from baseline with both right and left VNS on the anterior RV epicardium was seen. The difference in the results of this study may be due to a more detailed assessment of repolarization, with multiple electrodes in each region, the fact that ARI may be a more accurate surrogate of APD, and interspecies differences in innervation between canine and porcine models.

From a physiological and anatomic perspective, the lack of anterior, lateral, and posterior regional differences and laterality of the effects of right and left VNS on repolarization may be due to the type of nerve fibers (preganglionic rather than postganglionic fibers) in the vagosympathetic trunk and their obligatory synapse within the ganglia of the intrinsic cardiac nervous system (3, 44). Postganglionic fibers to the myocardium arise from these intrinsic cardiac ganglia (43). The intrinsic cardiac nervous system is known to regulate hemodynamic effects of VNS via a complex, integrated neural network. Therefore, elimination of a single ganglion may reduce but does not eliminate chronotropic or dromotropic responses (43). The fact that both right and left VNS have similar effects suggests that the neurons of the intrinsic cardiac ganglia mitigate and “smooth out” any electrophysiological laterality between the two sides, distributing the postganglionic effects to all regions in a more homogenous manner. These results are further confirmed by a microdialysis study (2) that demonstrated that the level of LV epicardial ACh release is similar with right versus left VNS. This is unlike sympathetic ganglia, where the left stellate ganglion's postganglionic fibers provide greater functional innervation to the posterior walls of the ventricles and the right stellate ganglion provides greater innervation to the anterior walls of the ventricles (29, 55, 58). These results support the value of unilateral VNS for cardiac therapeutic purposes as the net effects are likely to be distributed uniformly to the ventricles.

Apicobasal differences in ARI.

In the present study, significant apicobasal differences in response to VNS were noted, with the apex demonstrating a greater prolongation of ARI than the base. Right or left VNS, however, did not lead to a reversal in the direction of repolarization as the base had a shorter ARI at baseline. Mantravadi et al., using an optical mapping study in a decentralized Langendorff model, showed that bilateral VNS prolonged APD more at the apex than the base, consistent with this study. Mantravadi and colleagues (32a), however, reported that the repolarization at the apex was shorter than the base at baseline. This difference with our study may be due to the type of ex vivo preparation versus the location of the optical mapping performed, as certain regions of the anterior base of the LV have been reported to have a longer APD than the apex or the posterior base of the heart (6, 50). In the present study, the base overall had a shorter ARI at baseline, consistent with rabbit myocyte, canine, porcine, and human mapping studies showing a shorter APD, functional refractory period, and ARI at the base under control conditions (10, 11, 40, 42, 55). The shorter ARI at baseline in vivo could be due to the distribution of K⁺ channels such as a greater concentration of slowly activating delayed rectifier K⁺ current (I_Ks) channels at the base as well as a greater concentration of sympathetic fibers in this region (11, 26). The apicobasal differences during VNS were surprising and cannot be attributed to HR alone, as bradycardia would have affected all regions. Furthermore, they cannot be solely attributed release of other cotransmitters that may prolong APD, such as vasoactive intestinal peptide, as this would have lead to a greater prolongation of ARI at the base rather than the apex, similar to differences observed on the endocardium versus the epicardium. Therefore, the apicobasal differences may be due to reflex sympathetic activation (via activation of I_Ks, which is more prominent at the base, leading to shorter APD) or the distribution of ACh-dependent K⁺ channels, which maybe more densely distributed at the base, leading to a greater shortening of ARI at the base compared with the apex.

Transmural differences in ARI.

Overall, both right and left VNS caused small increases in DOR on the epicardium. This may be due to the bradycardia caused by VNS (28) or to the more sparse density of parasympathetic nerve fibers on the epicardium (51). In fact, functional parasympathetic innervation of the endocardium was greater than the epicardium, consistent with a histological study (51) showing that the endocardium has a greater density of parasympathetic fibers. Furthermore, parasympathetic fibers are thought to run from the endocardium to the epicardium, as supported by the observation that denervation of the epicardium does not reduce the endocardial ventricular response to VNS (34).

Hemodynamic correlates of VNS.

Cervical VNS has been shown to result in a negative chronotropic and inotropic response (4, 18, 22, 36, 57), although specific branches of the thoracic vagus nerve intermingle with sympathetic fibers and can have varied localized chronotropic and inotropic effects (5). Previous studies (12, 19) have suggested that the RVN may provide greater innervation to the sinoatrial (SA) node, whereas the LVN may provide greater innervation to the AV node in decentralized hearts. However, Armour et al. (4, 5) showed that both right and left VNS affect the SA node in a canine model with intact VNS. In the present study, we aimed to keep a similar sinus rate between right and left VNS to eliminate differences between right and left VNS that maybe due to HR alone. However, we did not observe a difference on PR interval and, therefore, AV node conduction between right versus left VNS. With regard to LV pressure, both right and left cervical VNS decrease LV pressure to a similar degree in a decentralized rabbit Langendorff model (7). Analogously, our study showed a significant decrease in LVESP and dP/dt_max. However, dP/dt_max showed the strongest correlation with repolarization effects. This finding is potentially significant in that when assessing ventricular effects of stimulation, particularly in clinical trials, HR may not be the best marker of the appropriate level of VNS.

Effects of VNS on lusitropy and diastolic function are more controversial. Xenopolous et al. (57) demonstrated no changes in LVESP or τ (57), whereas Henning et al. (22) showed that VNS decreased LVESP and increased τ in a canine model. In this porcine model, both right and left VNS decreased LVESP without affecting LVEDP, increased dP/dt_min, and increased τ, suggesting that VNS does not improve, and may in fact worsen, diastolic function.

Limitations.

General anesthetics can suppress nerve activity; however, we were able to reliably record a cardiomotor response during VNS. In addition, the drug concentrations were maintained at a constant level in this study. Furthermore, to reduce effects of inhaled anesthetics, an α-chloralose infusion was used during ARI recordings and VNS. For global and regional analyses, ARIs were not corrected for HR, as any HR effects on DOR are physiologically important. Atrial and ventricular pacing were not performed in all animals given the effect of pacing on altering autonomic tone. Finally, HR would not affect comparisons of regional differences within right and left VNS, and we were able to achieve a similar mean HR response during both right and left VNS. β-Blockers were not given during these experiments. Therefore, the effects of VNS in the setting of sympathetic blockade cannot be assessed from these experiments. Finally, VNS is currently being performed in various clinical trials in different ways without a clear standard; therefore, the results of this study may not completely duplicate what may be seen in clinical studies.

Conclusions.

In the present study, the detailed assessment of regional ARI and hemodynamic parameters showed a lack of laterality between the effects of right versus left VNS. In addition, the functional effects of VNS were greater at the apex than the base and greater on the endocardium than the epicardium, likely reflecting the functional distribution of parasympathetic innervation. The ventricular electrophysiological effects of VNS correlate best with the decrease in ventricular inotorpy and, specifically, dP/dt_max. Our results have significant implications given the advent of neuromodulation therapies using unilateral VNS for ventricular arrhythmias and cardiomyopathy.

GRANTS

This work was supported by American Heart Association National Fellow to Faculty Transition Award 11FTF755004 (to M. Vaseghi) and National Heart, Lung, and Blood Institute Grant R01-HL-084261 (to K. Shivkumar).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: K.Y., E.L.S., P.S.R., J.D.H., N.M., and M.V. performed experiments; K.Y., P.S.R., J.D.H., N.M., and M.V. analyzed data; K.Y., E.L.S., A.M., K.S., and M.V. interpreted results of experiments; K.Y., E.L.S., and M.V. prepared figures; K.Y. and M.V. drafted manuscript; K.Y., E.L.S., P.S.R., J.D.H., N.M., A.M., K.S., and M.V. approved final version of manuscript; E.L.S., P.S.R., J.D.H., N.M., A.M., K.S., and M.V. edited and revised manuscript; K.S. and M.V. conception and design of research.