Sympathetic responses induced by radiofrequency catheter ablation of atrial fibrillation

Abstract

Radiofrequency catheter ablation (RFCA) is a frequently performed procedure in patients with atrial fibrillation. Prior studies have shown that the RFCA may directly stimulate vagal afferents during the procedure, whereas the vagal tone assessed by heart rate variability (HRV) is lowered weeks after the RFCA procedure. The effects of RFCA performed in the left atrium on sympathetic nerve activity have not been assessed. In the present study, we hypothesized that RFCA would lower muscle sympathetic nerve activity (MSNA) during ablation and would raise MSNA 1 day postablation. A total of 18 patients were studied. In protocol 1 (n = 10), electrocardiogram, blood pressure, and MSNA in the peroneal nerve were recorded through the RFCA procedure performed in the electrophysiology laboratory. In protocol 2, eight patients were studied before the procedure and 1 day postablation. RFCA led to a decrease in MSNA immediately after the procedure (25.4 ± 3.2 to 17.2 ± 3.8 bursts/min, P < 0.05). Cardiac parasympathetic activity was determined using indexes of HRV and increased during the procedure. One day postablation, MSNA was above baseline values (21.3 ± 3.7 to 35.7 ± 2.6 bursts/min, P < 0.05). HRV indexes of cardiac parasympathetic activity fell, and the HRV index of sympathovagal balance was not significantly altered. The results show that RFCA raised cardiac parasympathetic activity and decreased MSNA during the procedure. One day postablation, MSNA rose and cardiac parasympathetic activity fell. In addition, RFCA evokes differentiated sympathetic responses directed to the heart and skeletal muscles.

NEW & NOTEWORTHY The effects of radiofrequency catheter ablation performed in the left atrium on muscle sympathetic nerve activity (MSNA) have not been assessed. The results of this study show that radiofrequency catheter ablation raised cardiac parasympathetic activity and decreased MSNA during the procedure. One day postablation, MSNA rose and cardiac parasympathetic activity fell. We speculate that the partial autonomic afferent denervation induces these effects on autonomic activity.

INTRODUCTION

Atrial fibrillation (AF) is a common clinical problem (24, 36), which affects 2.3 million patients in the United States (19). It is associated with an increased risk for stroke (68) and an increase in morbidity and mortality in patients with structural heart disease (6). Recently, alterations in autonomic nervous system activity have been identified as a trigger of atrial arrhythmias including AF (8, 10). For example, increased vagal tone is frequently noted at the onset of AF (7).

Radiofrequency catheter ablation (RFCA) is an effective and progressively more common strategy for the treatment of drug refractory AF (40, 45). Catheter ablation for AF in patients with heart failure is associated with improved cardiac function, exercise capacity, and quality of life while reducing hospitalizations and death from any cause (26, 40). Although the antiarrhythmic mechanisms responsible for successful ablation are controversial (22, 50–53), cardiac parasympathetic denervation is considered as a likely mechanism for the antiarrhythmic action of RFCA (53).

The radiofrequency ablation procedure is performed by applying radiofrequency energy with an externally irrigated electrode catheter that results in “resistive heating” and coagulation necrosis of the myocardium (67). Early studies have shown that noxious heat-induced activation in cutaneous C-fibers in cats (35). Importantly, a prior report (48) showed that noxious heat to the mechanically excitable field in the wall of the pulmonary artery and left ventricle also activated autonomic afferent fibers in cats. Since RFCA may primarily affect cardiac parasympathetic nerves (53), we speculated that delivering radiofrequency energy (heating-induced coagulation necrosis) would activate cardiac parasympathetic nerves during the procedure. Moreover, it is known that acute burn injury is usually associated with pain and mechanical hyperalgesia in the injured and nearby areas (32). Thus, we speculated cardiac parasympathetic activation could also be observed during the intervals between radiofrequency energy applications. We speculated that this acute effect would fade with time after the procedure and that the effects of cardiac parasympathetic denervation would be observed later (e.g., 1 day postablation). Thus, we speculated that autonomic responses after the procedure (e.g., 1 day) would be different from those during the procedure.

Results from prior reports support this speculation. For example, during the RFCA procedure, transient sinus bradycardia and hypotension are commonly observed (25, 50–53, 64). This suggests that the procedure engages cardiac vagal efferents during the procedure. On the other hand, a number of reports using heart rate (HR) variability (HRV) analysis have shown that RFCA can induce postproceudre (a week to months) vagal efferent withdrawal (25, 42, 53, 56, 59, 69). Interestingly, attenuation of vagal efferent activity may be associated with a reduced rate of AF recurrences (5, 53, 60, 69).

It has been generally accepted that the sympathetic and parasympathetic nerves act reciprocally. When cardiac vagal activity was decreased, muscle sympathetic nerve activity (MSNA) increased (18, 66). When cardiac vagal activity was raised, MSNA decreased (12, 55). Thus, we speculated that when vagal nerves are activated during the RFCA procedure, systemic sympathetic nerve activity (e.g., MSNA) would decrease. Prior studies in dogs have suggested that stimulation of receptors in the atrium led to a decrease in renal sympathetic activity (28). We further speculate that 1 day postablation vagal activities would be decreased and that systemic sympathetic nerve activity (e.g., MSNA) would increase. To our knowledge, there are no prior reports that have examined the effects of RFCA on MSNA in humans.

On the other hand, since autonomic nerves (e.g., both afferent and efferent) in the heart can be ablated with this procedure, we speculated that sympathetic responses in the heart may be different from sympathetic responses directed to other organs (e.g., vascular beds in muscles). Prior reports that used HRV to examine cardiac sympathetic activity have yielded conflicting results (25, 27, 42, 50, 53, 54, 56, 59, 60). These reports examined the HRV a week to several months after the RFCA procedure and showed that RFCA raised (25, 53, 56), did not alter (42, 54, 59), or decreased (27, 50, 60) indexes of cardiac sympathetic tone.

The purpose of the present study was to investigate the sympathetic neural adjustments that accompany the RFCA procedure. We hypothesized that 1) during the procedure, RFCA may directly stimulate the vagal afferents, which would lead to increased cardiac vagal activation and systemic sympathetic withdrawal, and 2) 1 day postablation, the denervation of vagal nerves would lead to a decrease in cardiac vagal tone and a rise in MSNA. Our results support these hypotheses.

METHODS

Patients

A total of 18 patients (16 men and 2 women, 56 ± 3 yr, 180 ± 2 cm, 97 ± 5 kg) who underwent RFCA for paroxysmal AF were studied. All patients underwent the RFCA procedure as part of their standard care treatment. The experimental protocol was approved by the Institutional Review Board of the Milton S. Hershey Medical Center and conformed with the Declaration of Helsinki. Each patient had the purposes and risks of the protocol explained to him or her before written informed consent was obtained.

Measurements

Beat-by-beat HR was monitored from the electrocardiogram (Cardicap/5, Datex-Ohmeda, GE Healthcare). As described in previous reports (13–15), multifiber recordings of MSNA were obtained with a tungsten microelectrode inserted in a peroneal nerve. A reference electrode was placed subcutaneously 2–3 cm from the recording electrode. The recording electrode was adjusted until a site was found in which muscle sympathetic bursts were clearly identified using previously established criteria (65). The nerve signal was amplified, band-pass filtered with a bandwidth of 500–5,000 Hz, and integrated with a time constant of 0.1 s (Iowa Bioengineering, Iowa City, IA).

Catheter Ablation Procedure

Antiarrhythmic drugs were discontinued five half-lives before the procedure. Anticoagulation therapy was continued until the day before the procedure. The ablation procedure was conducted under moderate sedation with propofol or dexmedetomidine. Small doses of midazolam or fentanyl were added as needed. Transesophageal echocardiography was performed in all patients the day before the procedure to rule out atrial or atrial appendage thrombus. The absence of a left atrial thrombus was again ruled out with intracardiac echocardiography just before transseptal puncture (62). Multipolar electrode catheters were introduced percutaneously and positioned in the coronary sinus, right atrial appendage, His bundle region, and right ventricle. An initial electrophysiology study was performed to rule out an accessory arteriovenous pathway and inducible tachycardia. Intracardiac echocardiography and three-dimensional reconstructions of the anatomy of the left atrium, left atrial appendage, and pulmonary veins were obtained with the CartoSound system (29). After administration of heparin (5,000 units iv), two 8.5-F long sheaths (SR0 and SL1, St. Jude Medical, St. Paul, MN) were introduced into the left atrium with transseptal puncture performed under fluoroscopic and intracardiac echocardiographic guidance. A circular mapping catheter (Lasso Biosense Webster, Diamond Bar, CA) was introduced through the SL1 sheath into the left atrium for electrical mapping of the pulmonary veins. An externally irrigated electrode catheter (3.5 mm, D curve, SmartTouch, Biosense Webster) was introduced through the SR0 sheath into the left atrium for mapping and ablation (power: 10–30 W). After transseptal puncture, patients received additional doses of intravenous heparin to maintain an activated clotting time of >320 s. The radiofrequency ablation procedure was performed by delivering lesions in the antra just outside the pulmonary veins without additional lesions being delivered to the left atrium. The end point of these procedures was demonstration of entrance and exit block (capture of pulmonary vein potentials during pacing within the pulmonary veins with dissociation from the atrium). Anticoagulation was continued for at least 3 mo or indefinitely in patients with a CHADS₂ scores ≥2. Antiarrhythmic therapy was restarted immediately after the procedure and continued for at least 3 mo.

Protocols

Protocol 1: responses during the left atrial RFCA procedure.

All experiments were performed in the morning. Responses during the left atrial RFCA procedure were recorded in the cardiac electrophysiology laboratory. Patients (n = 10) were instrumented for electrocardiogram, respiratory movement with a pneumograph, and MSNA. Variables were recorded with a data-acquisition system (MacLab, AD Instruments, Castle Hill, NSW, Australia) at 200 Hz throughout the RFCA procedure. Averaged systolic and diastolic blood pressure (BP) and mean arterial BP were obtained in 5-min segments from a 4-Fr cannula positioned in the left radial artery. Ten minutes of baseline data were collected, and variables were recorded throughout the RFCA procedure.

There was a technical challenge to keep MSNA recording during the RFCA procedure as the procedure table and patients’ body were moved during the procedure. If the MSNA signal was lost during the procedure, the electrodes were left in place but no attempt was made to recover the recording until after the RFCA procedure was completed. An effort to recover the MSNA site was then attempted for 5 min. Regardless of whether the MSNA site was maintained, recovered, or not reobtained, 10 min of immediate postablation data were then collected (i.e., from the 5th to 15th minute from the end of RFCA procedure).

Protocol 2: responses 1 day postablation.

This protocol was performed in a clinical research laboratory within the Penn State Clinical and Translational Science Institute at the Hershey Medical Center. These were different patients from those in protocol 1. Patients (n = 8) were studied on 2 different days. All experiments were performed in the morning. Visit 1 (before RFCA) occurred within 1 wk before the left atrial RFCA procedure. Visit 2 was performed at the time of patient discharge after the RFCA. This was approximately 1 day after the procedure. The two visits were scheduled in the approximate same time of the day (i.e., morning). In working this comparison, we assumed that aside from the ablation procedure itself, other clinical factors would remain constant. During these two study visits, electrocardiogram and MSNA were recorded for 20 min. BP was measured with an automated sphygmomanometer in the brachial artery (SureSigns VS3, Philips Medical Systems) six times over the 20-min period.

Data Analysis

For protocol 1, data were examined during the following three time periods: 1) 10 min of baseline before the procedure, 2) 10 min of data before the end of the RFCA procedure, and 3) 10 min of immediate postablation data after the procedure. We also obtained 3 min of data after the infusion of propofol or dexmedetomidine but before the ablation began. Because the number of the burning bouts, the interval time between the burning bouts, and the order for the treated areas varied from case to case, it was difficult to examine the responses during earlier stages of the procedure. Thus, the 10 min of data before the end of the RFCA procedure were examined, when the procedure was close to being completed in all cases. For protocol 2, 20-min periods of resting recordings were analyzed during the preablation and 1 day postablation procedure visits.

As described in previous reports (13, 15), MSNA bursts were first identified in real time by visual inspection of the data coupled with the burst sound from the audioamplifier. These bursts were further evaluated by a computer program that identified bursts based on fixed criteria, including an appropriate latency following the R wave of the electrocardiogram and having a signal-to-noise ratio of at least 2:1. The bursts were also manually checked on a beat-to-beat base. Integrated MSNA was normalized by assigning a value of 100 to the mean amplitude of the largest 10% of the bursts during the baseline (13). Total MSNA activity for each cardiac cycle was calculated from burst area of the integrated neurogram (13). If a burst was not detected for a given cardiac cycle, a 0 was assigned to that cycle. Mean values of MSNA data are reported as burst rate (i.e., bursts/min), burst incidence (i.e., bursts/100 heart beats), and total activity (i.e., total burst area/min). The data segments with AF were excluded for MSNA analysis. In protocol 1, delivering radiofrequency energy (~1–2 min for each bout) induced strong electronic noise in the neurogram traces. MSNA data were obtained in the intervals between the bouts of delivering radiofrequency energy. If the MSNA signal was lost and later reobtained immediately after the procedure, the MSNA trace was normalized again by assigning a value of 100 to the mean amplitude of the largest 10% of the bursts during the recording under the resting condition.

The beat-to-beat time series of RR intervals (RRI) were analyzed for HRV using Hemolab software (Harald Stauss Scientific). For HRV analysis during the 10-min periods, we used only high-quality 5-min continuous recording segments of sinus rhythm. Data segments with AF were excluded. Premature ventricular contractions, atrial premature contractions, and electrical artifacts were excluded from the analysis. HRV parameters were obtained and used as indicators of autonomic activity according to previously reported guidelines (63a). Time-domain HRV parameters in this report included 1) the SD of normal to normal intervals (SDNN), 2) the root mean square of differences between successive normal to normal intervals (RMSSD), and 3) the proportion (in %) of adjacent normal to normal intervals differing by >50 ms (pNN50) of 5 min. Frequency-domain parameters included 1) low-frequency components (LF; 0.04–0.15 Hz), 2) high-frequency components (HF; 0.15–0.45 Hz), and 3) the LF-to-HF ratio (LF/HF). HF and RMSSD served as indicators of cardiac parasympathetic activity, and LF/HF reflected cardiac sympathetic activity or cardiac sympathovagal balance (63a). SDNN and pNN50 may reflect both cardiac sympathetic and parasympathetic activities (63a).

Statistical Analyses

Statistical analyses were performed with the use of SigmaPlot software (version 13, Systat Software). Effects of RFCA on HRV indexes and hemodynamic variables during the RFCA procedure (protocol 1) were examined using one-way repeated-measures ANOVA. When appropriate, Tukey’s post hoc method to adjust for multiple comparisons was used. If the normality test (Shapiro-Wilk) was failed, Friedman repeated-measures ANOVA on ranks was performed. The effects of RFCA on MSNA, HRV indexes, and hemodynamic variables 1 day postablation (protocol 2) were examined via paired t-tests. When the normality test (Shapiro-Wilk) failed, the Wilcoxon signed-rank test was performed. The level of significance was set at P < 0.05. Values are reported as means ± SE.

RESULTS

Patient Characteristics

Thirteen of eighteen patients had at least one risk factor for heart disease based on guidelines set forth by the American Heart Association (Table 1). No patient had a history of myocardial infarction. All patients (n = 18) included in this report had successful electrical isolation of the pulmonary veins with RFCA.

Table 1.

Clinical characteristics of patients

Comorbidities	Count	Pharmacotherapy	Count
Asthma	2 (11%)	Angiotensin-converting  enzyme inhibitor	4 (22%)
Chronic obstructive pulmonary disease	2 (11%)	Antiarrhythmic	10 (56%)
Depression	1 (6%)	Anticoagulant	17 (94%)
Diabetes	1 (6%)	Aspirin	5 (28%)
Gastresophageal reflux disease	4 (22%)	β-Adrenergic  receptor antagonist	11 (61%)
Hyperlipidemia	7 (39%)	Ca2+ channel antagonists	6 (33%)
Hypertension	8 (44%)	Proton pump inhibitor	8 (44%)
Obstructive sleep apnea	3 (17%)	Statin	5 (28%)
Thoracic outlet syndrome	1 (6%)	Thiazide diuretic	1 (6%)

n = 18 total patients.

Autonomic Response During the Left Atrial RFCA Procedure

After the infusion of propofol or dexmedetomidine but before the ablation procedure began, MSNA values were not significantly different from baseline (26.6 ± 3.2 to 24.4 ± 2.1 bursts/min, P = 0.18, and 636 ± 97 to 601 ± 130 U/min, P = 0.69, n = 8).

MSNA was obtained before and immediately after the procedure in seven patients. MSNA burst rate and total activity immediately after the procedure were lower than baseline values in each of these seven patients studied (Fig. 1). Moreover, MSNA burst incidence also decreased (37.4 ± 5.8 to 21.2 ± 5.5 bursts/100 beats, n = 7, P = 0.01). As expected, the RFCA procedure led to substantial body movement. Despite this, MSNA was successfully monitored in three patients during the procedure. MSNA decreased during and immediate postablation in each of these three patients (Table 2). The representative recordings are shown in Fig. 2.

Fig. 1.

Mean and individual muscle sympathetic nerve activity (MSNA) values for baseline and immediately postablation (n = 7). In 4 of 7 patients, MSNA was lost during the procedure and reobtained immediately after the procedure. Symbols represent individual data.

Table 2.

Individual muscle sympathetic nerve activity values during ablation procedure

	Baseline	Ablation	Immediate Postablation
Burst rate, bursts/min
Patient 1	21.3	14.8	16.2
Patient 2	30.3	12.3	10.9
Patient 3	33.7	20.9	12.2
Total activity, units/min
Patient 1	259	187	210
Patient 2	885	468	345
Patient 3	619	287	168

Data from these three patients are also shown in Fig. 1.

Fig. 2.

Representative tracings of electrocardiogram (ECG) and muscle sympathetic nerve activity (MSNA) during the radiofrequency catheter ablation in one patient. The solid circles above the MSNA tracings indicate the identified MSNA bursts. The solid circles below the MSNA tracings (i.e., zeros) indicate no burst for the cardiac cycles.

BP and HR (n = 10) did not significantly change with the RFCA procedure (Table 3). During the ablation procedure, sinus bradycardia was observed in 7 of 10 patients. The representative recordings of bradycardia are shown in Fig. 3. The time-domain indexes RMSSD and SDNN were significantly higher during the RFCA procedure than they were at baseline (Table 3). Likewise, the frequency-domain HRV index HF was higher during the RFCA procedure than the baseline value (Table 3). Compared with baseline value, LF/HF was not altered during the procedure and was significantly lower after the procedure.

Table 3.

Hemodynamic variables and heart rate variability indexes during the left atrial radiofrequency catheter ablation procedure

	Baseline	Ablation	Immediate Postablation	P Value
Systolic blood pressure, mmHg	119.9 ± 3.7	123.2 ± 4.2	119.3 ± 4.5	0.434
Mean arterial blood pressure, mmHg	94.9 ± 3.5	98.1 ± 3.4	92.3 ± 3.9	0.171
Diastolic blood pressure, mmHg	77.3 ± 3.6	79.1 ± 3.7	74.9 ± 3.9	0.269
Heart rate, beats/min	77.8 ± 6.4	83.6 ± 5.9	86.4 ± 5.7	0.575
RMSSD, ms	64 ± 18	125 ± 29*	104 ± 32	0.049
SDNN, ms	59 ± 16	94 ± 18†	77 ± 19	0.049
pNN50, %	22.0 ± 8.3	26.9 ± 8.0	25.0 ± 8.8	0.860
HF, ms2	70 ± 38	160 ± 62‡	102 ± 58	0.045
LF/HF	1.28 ± 0.30	0.98 ± 0.23	0.32 ± 0.08§	<0.001

Values are means ± SE; n = 10 for variables. Systolic blood pressure, mean arterial pressure, diastolic blood pressure, and heart rate did not significantly change during the procedure. RMSSD, root mean square of differences between successive normal to normal intervals; SDNN, SD of normal to normal intervals; pNN50, proportion (in %) of adjacent normal to normal intervals differing by >50 ms; HF, high frequency; LF, low frequency. P values were determined by one-way-repeated measures ANOVA.

^*P = 0.043 vs. baseline;

^†P = 0.039 vs. baseline.

^‡Difference of ranks = 11, q = 3.479, P < 0.05 vs. baseline with a rank test;

^§difference of ranks = 16, q = 5.06, P < 0.05 vs. baseline with a rank test.

Fig. 3.

Representative tracings of sinus bradycardia during the radiofrequency catheter ablation (Abl) in 1 patient. RA, right atrium; CS, coronary sinus; p, proximal; d, distal.

Autonomic Function Changes 1 Day After the Left Atrial RFCA Procedure

Preablation and 1 day postablation MSNA recordings were successfully obtained in six patients (Fig. 4). Compared with baseline, MSNA burst rate and total activity increased in each individual 1 day postablation (Fig. 4). Averaged MSNA burst incidence also was higher than before the procedure (36.3 ± 8.4 to 51.1 ± 4.7 bursts/100 beats, n = 6, P = 0.13).

Fig. 4.

Mean and individual muscle sympathetic nerve activity (MSNA) before and 1 day postablation (n = 6). Symbols represent individual data. These patients were different from the patients shown in Fig. 1.

One day postablation, BPs were not significantly different from baseline values noted before RFCA (visit 1). HR rose 1 day postablation (Table 4). One day postablation, the time-domain indexes RMSSD, SDNN, and pNN50 and frequency-domain index HF were significantly lower than values noted before the procedure (Table 4). LF/HF was not different from the value before the procedure (P = 0.20).

Table 4.

Hemodynamic variables and heart rate variability indexes before and 1 day after the left atrial radiofrequency catheter ablation procedure

	Before	1 Day Postablation	P Value
Systolic blood pressure, mmHg	127.3 ± 9.6	120.1 ± 5.5	0.496
Mean arterial blood pressure, mmHg	92.0 ± 4.4	84.8 ± 3.4	0.196
Diastolic blood pressure, mmHg	75.4 ± 2.9	68.1 ± 3.4	0.127
Heart rate, beats/min	63.3 ± 3.2	73.2 ± 3.4	0.024
RMSSD, ms	66 ± 13	31 ± 9	0.008
SDNN, ms	54 ± 6	24 ± 5	0.008
pNN50, %	17.6 ± 3.2	9.4 ± 3.5	0.010
HF, ms2	154 ± 62	64 ± 38	0.024
LF/HF	1.30 ± 0.52	0.72 ± 0.19	0.269

Values are means ± SE. These patients were different from the patients shown in Table 2 (n = 8). RMSSD, root mean square of differences between successive normal to normal intervals; SDNN, SD of normal to normal intervals; pNN50, proportion (in %) of adjacent normal to normal intervals differing by >50 ms; HF, high frequency; LF, low frequency. P values were determined by a paired t-test.

DISCUSSION

The main findings of the present study were that 1) efferent sympathetic tone directed to skeletal muscles (i.e., MSNA) decreased and cardiac parasympathetic tone increased during the RFCA procedure and 2) 1 day postablation, MSNA rose, whereas cardiac parasympathetic tone fell. In the following sections, we will discuss potential mechanisms for these differential responses during and after the procedure.

Autonomic Response During the Left Atrial RFCA Procedure

During the ablation period, the cardiac parasympathetic activity indexes RMSSD and HF were significantly higher than baseline. We suspect that radiofrequency energy directly stimulated cardiac parasympathetic nerve terminals in areas of pulmonary veins during the procedure (33, 39). It is known that bradycardia and hypotension occur in some patients during pulmonary vein ablation (25, 50, 52). The incidence of the ablation-induced bradycardia-hypotension response is higher when radiofrequency energy is delivered close to the pulmonary vein compared with other areas of the left atrium (64). Prior work has suggested that radiofrequency current directly stimulates cardiac parasympathetic fibers traveling from the site of radiofrequency application to the sinus node and can result in a profound sinus bradycardia response (9, 17). Of note, the sinus bradycardia response was observed in 7 of 10 patients during the ablation procedure in our study. The increases in RMSSD and HF we noted during the ablation support this argument. Because an increase in HR after the bradycardia response was also observed in some cases (see Fig. 3), BP and HR (i.e., mean values over 10 min) responses were not affected by the ablation. Prior work that examined HRV showed different results than we did (25, 53). However, our results were obtained during the procedure, whereas the prior work (25, 53) examined HRV after the procedure (days or months). LF/HF during the ablation was not significantly different from baseline. It could indicate a sympathovagal balance (49, 63a) while the cardiac parasympathetic activity rose.

Our data show that MSNA decreased during and immediately after the procedure. To our knowledge, ours is the first study where MSNA was measured directly during left atrial RFCA. In protocol 1, the immediate postablation data were collected within 15 min after the procedure completion. Since the MSNA changes during the ablation and immediately after the procedure showed similar directional changes, we speculate that both ablation and immediate postablation reflect acute autonomic responses during the procedure. Our finding of MSNA in protocol 1 was based on a comparison between baseline and immediate postablation data (n = 7).

A prior study (20) reported MSNA during an acute episode of AF and after spontaneous restoration of the sinus rhythm. The restoration of sinus rhythm from AF induced a decrease in HR, an increase in MSNA, and no changes in BP (20). Thus, the MSNA decrease during and immediately after the procedure in the present study was different from the MSNA response after the spontaneous restoration of the sinus rhythm.

Previous studies have shown that other catheter procedures induced different autonomic responses from those noted in our study. For example, MSNA increases during coronary angioplasty (11). We noted that MSNA fell during the RFCA procedure. Another study (43) showed that coronary angiography lowered LF/HF immediately before and after the procedure only in patients with coronary diseases and did not alter LF/HF in patients without coronary diseases. Based on these prior reports, we speculate that the MSNA and HRV responses observed in our study are induced by the catheter ablation procedure itself.

We speculate that ablation stimulated autonomic afferents in the pulmonary veins and that this led to the changes in MSNA. Prior work (57) has suggested that atrial stimulation evokes systemic sympathetic withdrawal in humans. Moreover, activation of left atrial receptors by distension of balloons at the pulmonary vein-atrial junctions decreases renal sympathetic nerve activity in dogs (28).

A prior study by Hamadan et al. (23) reported MSNA changes during and after atrioventricular junction ablation in patients with AF. They found that after atrioventricular junction ablation, pacing at 60 beats/min evoked an increase in MSNA, whereas pacing at 90 beats/min lowered MSNA values. The difference in results between their study and ours are likely due to the location of the ablation (right atrium radiofrequency energy versus left atrium radiofrequency energy) and the fact that they paced the heart, whereas we did not.

Autonomic Function Changes 1 Day After the Left Atrial RFCA Procedure

One day postablation, the cardiac parasympathetic activity indexes RMSSD and HF significantly decreased from values obtained before the procedure. We speculate that the effects observed were due to partial vagal afferent denervation. This would lead to efferent vagal withdrawal. The observations regarding vagal withdrawal are consistent with prior reports (25, 27, 42, 53, 54, 56, 60). Of note, these prior reports examined HRV weeks to months after the procedure.

One day postablation, when we observed increased MSNA, we did not note a statistically significant reduction in BP; however, HR was significantly increased. This suggests that the changes in afferent inputs from the pulmonary vein tonically contribute to MSNA activation. This led to a speculation that the partial pulmonary vein afferent denervation seen with RFCA evokes both vagal cardiac withdrawal and systemic sympathetic activation, as discussed below.

Prior studies have shown that sympathetic and parasympathetic nerves act reciprocally. In healthy individuals, when cardiac vagal activity (indicated with HF power of RRI) decreases, MSNA increases during tilt (18) or during insulin infusion (66). Compared with healthy individuals, the HF power of RRI was lower and MSNA was higher in patients with heart failure (3). On the other hand, a meta-anlysis study (12) showed that when cardiac vagal activity was raised (increases in HF power of RRI and RMSSD) with body weight loss, MSNA was decreased. Moreover, another meta-analysis study (55) showed that exercise training raised vagal activity (increases in HF power of RRI and RMSSD) and decreased MSNA in patients with heart failure. Based on these prior observations, we speculate that the cardiac vagal denervation seen with RFCA evokes both vagal cardiac withdrawal and systemic sympathetic activation.

Although LF power was used as an cardiac sympathetic index in some prior reports (21, 37), other studies have shown that LF power was not correlated with cardiac sympathetic activity in humans (30, 44) or sheep (41). Thus, only LF/HF is reported in the present study. LF/HF has been considered as an indicator of cardiac sympathovagal balance and/or to reflect cardiac sympathetic tone (49, 63a). Some prior work has shown that LF/HF changes correlate with plasma norepinephrine concentrations (63a). One day postablation, LF/HF was not significantly different than the value noted before the procedure. This observation is consistent with some prior reports (42, 54, 59). However, other studies have shown that RFCA either increased (25, 53) or decreased (27, 50, 60) this index. SDNN and pNN50 are the indexes for both cardiac sympathetic and parasympathetic activities (63a). Our data demonstrate that cardiac sympathetic and parasympathetic indexes (i.e., SDNN and pNN50) decreased, whereas sympathovagal balance (LF/HF) did not change. These data may thus suggest a decrease in cardiac parasympathetic activity without an increase in cardiac sympathetic activity. A simultaneous reduction in both cardiac sympathetic and vagal activity after RFCA has been previously reported (27). Why RFCA would lead to the different responses in cardiac sympathetic activity and muscle sympathetic activity is not clear and is speculated below.

The neurons of the intrinsic cardiac autonomic system are located in the epicardial fat pads near the pulmonary veins and at the junction with the left atrium (1, 53, 63). Catheter ablation to treat AF includes isolating the pulmonary vein from the left atrium. Some previous reports have demonstrated that cardiac sympathetic nerves form an important afferent pathway and that the receptor endings of the fibers consist of free terminals scattered diffusely in the heart. The sympathetic fibers are distributed throughout the atria (31). Moreover, ganglionated plexi of cardiac autonomic nerves located along the epicardial aspects of the pulmonary vein antrum can be a target of catheter ablation of AF (46). The cardiac sympathetic and parasympathetic nerve structures are highly colocalized in the heart (atrium); thus, it is not possible to determine if RFCA eliminated only parasympathetic afferent input (8). We suspect that RFCA also ablated some sympathetic nerves in the heart. Thus, cardiac sympathetic activity did not rise as MSNA did.

RFCA is an effective strategy for the treatment of drug refractory AF. The present data suggest that RFCA evokes a differentiated sympathetic response. Thus, it is necessary to assess the sympathetic tone to other vascular beds. Future work will be needed to see if the increase in MSNA as noted here is associated with increases in hypertension (2) and heart failure (34).

Study Limitations

The RFCA procedure was performed under moderate sedation with propofol or with dexmedetomidine. Small doses of midazolam or fentanyl were added as needed. General anesthesia with propofol (16) decreased MSNA and BP, whereas general anesthesia with etomidate did not alter MSNA (16). A prior report using propofol demonstrated that when BP was restored to the preanesthetic level via adjusting the propofol dose during the surgery, MSNA was also restored to the preanesthetic level (61). In the present study, MSNA before the RFCA procedure was not different from baseline. BP during the RFCA procedure was not different from the baseline value. Importantly, patients were fully awake during the data recording period immediately after the RFCA procedure (i.e., data shown in Fig. 1). Thus, we believe that the observed MSNA decrease in protocol 1 was not an effect of anesthesia.

Antiarrhythmic drugs were discontinued before the procedure and were restarted on the day after the procedure. Prior studies have shown that oral flecainide (47) and intravenous quinidine (38) raised MSNA in healthy individuals, whereas intravenous procainamide (58) decreased MSNA in healthy individuals. To our knowledge, there are no reports regarding the effects of stopping or restarting antiarrhythmic agents on MSNA in patients such as the those we studied. Thus, we cannot exclude the possibility that stopping and restarting the antiarrhythmic agent(s) might contribute to the observations made in this study.

Conclusions

The RFCA procedure evokes cardiac parasympathetic activation and suppression in sympathetic activity directed to muscles during the procedure. We speculate that RFCA energy stimulates pulmonary vein afferents, which induce the observed effects in autonomic activity during the procedure. One day postablation, cardiac parasympathetic activity is reduced, whereas the sympathetic outflow to muscles increases. We speculate that partial autonomic afferent denervation induces these effects on autonomic activity 1 day postablation.

GRANTS

This work was supported by American Heart Association Grant 15GRNT24480051 (to J. Cui), a Penn State Heart and Vascular Institute Penny A. Garban Endowment (to J. Cui), and National Center for Advancing Translational Sciences Grant UL1-TR-002014 (to L. I. Sinoway).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.C., M.D.G., and L.I.S. conceived and designed research; J.C., M.D.G., and C.B. performed experiments; J.C., C.B., and A.H. analyzed data; J.C., M.D.G., and L.I.S. interpreted results of experiments; J.C. and M.D.G. prepared figures; J.C. and M.D.G. drafted manuscript; J.C., M.D.G., A.H., and L.I.S. edited and revised manuscript; J.C., M.D.G., C.B., A.H., and L.I.S. approved final version of manuscript.