1 2 3Q3 4 5Q4 6 7Q5 8 9Q1 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

Subcutaneous nerve activity and spontaneous ventricular arrhythmias in ambulatory dogs Anisiia Doytchinova, MD,* Jheel Patel,* Shengmei Zhou, MD,† Lan S. Chen, MD,‡ Hongbo Lin,§¶ Changyu Shen, PhD,§¶ Thomas H. Everett IV, PhD, FHRS,* Shien-Fong Lin, PhD,* Peng-Sheng Chen, MD, FHRS* From the *Krannert Institute of Cardiology and Division of Cardiology, Department of Medicine, Indiana University School of Medicine, Indianapolis, Indiana, †Department of Pathology and Laboratory Medicine, Children’s Hospital Los Angeles, Keck School of Medicine of University of Southern California, Los Angeles, California, ‡Department of Neurology, Indiana University School of Medicine, Indianapolis, Indiana, § Department of Biostatistics, Indiana University School of Medicine, Indianapolis, Indiana, and ¶Fairbanks School of Public Health, Indiana University, Indianapolis, Indiana. BACKGROUND Stellate ganglion nerve activity (SGNA) is important in ventricular arrhythmogenesis. However, because thoracotomy is needed to access the stellate ganglion, it is difficult to use SGNA for risk stratification.

demonstrated 60 seconds, 40 seconds, and 20 seconds before VT/VF (P o.05). The Pearson correlation coefficient for integrated SCNA and SGNA was 0.73 ⫾ 0.18 (P o.0001 for all dogs, n ¼ 5). Both SCNA and SGNA exhibited circadian variation.

OBJECTIVE The purpose of this study was to test the hypothesis that subcutaneous nerve activity (SCNA) in canines can be used to estimate SGNA and predict ventricular arrhythmia.

CONCLUSION SCNA can be used as an estimate of SGNA to predict susceptibility to VT and VF in a canine model of ventricular arrhythmia and sudden cardiac death.

METHODS We implanted radiotransmitters to continuously monitor left stellate ganglion and subcutaneous electrical activities in 7 ambulatory dogs with myocardial infarction, complete heart block, and nerve growth factor infusion to the left stellate ganglion.

KEYWORDS Atrioventricular block; Autonomic nervous system; Myocardial infarction; Sudden cardiac death; Ventricular arrhythmia

RESULTS Spontaneous ventricular tachycardia (VT) or ventricular fibrillation (VF) was documented in each dog. SCNA preceded a combined 61 episodes of VT and VF, 61 frequent bigeminy or couplets, and 61 premature ventricular contractions within 15 seconds in 70%, 59%, and 61% of arrhythmias, respectively. Similar incidence of 75%, 69%, and 62% was noted for SGNA. Progressive increase in SCNA [48.9 (95% confidence interval [CI] 39.3–58.5) vs 61.8 (95% CI 45.9–77.6) vs 75.1 (95% CI 57.5–92.7) mV-s] and SGNA [48.6 (95% CI 40.9–56.3) vs 58.5 (95% CI 47.5–69.4) vs 69.0 (95% CI 53.8–84.2) mV-s] integrated over 20-second intervals was

Introduction Sympathetic activation is associated with increased risk of ventricular arrhythmias and sudden cardiac death (SCD).1 We have shown that stellate ganglion nerve activity (SGNA) This study was supported in part by National Institutes of Health Grants P01HL78931, R0171140, and R41HL124741; a Medtronic–Zipes Endowment; and the Indiana University Health-Indiana University School of Medicine Strategic Research Initiative. Drs. Lin and Chen have equity interest in Arrhythmotech, LLC. Cyberonics, Medtronic, and St. Jude Medical Inc donated research equipment to Dr. Chen’s research laboratory. Address reprint requests and correspondence: Dr. Peng-Sheng Chen, 1800 N Capitol Ave, E475, Indianapolis, IN 46202. E-mail address: [email protected].

1547-5271/$-see front matter B 2014 Heart Rhythm Society. All rights reserved.

ABBREVIATIONS AIVR ¼ accelerated idioventricular rhythm; CI ¼ confidence interval; ECG ¼ electrocardiogram; FBG/C ¼ frequent bigeminy or couplets; HASDA ¼ high-amplitude spike discharge activity; iSCNA ¼ integrated subcutaneous nerve activity; iSGNA ¼ integrated stellate ganglion nerve activity; LABDA ¼ lowamplitude burst discharge activity; PVC ¼ premature ventricular contraction; SCD ¼ sudden cardiac death; SCNA ¼ subcutaneous nerve activity; SGNA ¼ stellate ganglion nerve activity; VF ¼ ventricular fibrillation; VT ¼ ventricular tachycardia (Heart Rhythm 2014;0:0–9) rights reserved.

I

2014 Heart Rhythm Society. All

precedes spontaneous ventricular tachycardia (VT) and ventricular fibrillation (VF) in an ambulatory canine model of SCD.2 Because sympathetic nerve activity is important in arrhythmia initiation, it is highly desirable to develop a reliable and less invasive method to measure sympathetic outflow for arrhythmia prediction and risk stratification. Heart rate variability and microneurography have been used to assess sympathetic tone in patients. However, because of technical problems, those methods are not widely used for arrhythmia prediction. The skin is well innervated by sympathetic nerve fibers.3,4 Studies in dogs and rats demonstrated that the somata of the cutaneous sympathetic nerve fibers of the upper body are http://dx.doi.org/10.1016/j.hrthm.2014.11.007

55 56 57 58 59 60 61 62 63 64 65 66 67 68

2 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 Q6 113 114 115 116 117 118 119 120 121 122 123 124 125

located in the middle cervical and stellate ganglia.5,6 Innervation of the pectoralis muscle, which is located underneath the hypodermis in the upper chest wall, stems from the brachial plexus,7 which communicates with the stellate ganglion.5 In addition, light microscopy has revealed that large nerve trunks are present in the subcutaneous tissues.5 These observations suggest that recording skin and/or muscle directed postganglionic sympathetic nerve activity, which we will collectively term as “subcutaneous nerve activity” (SCNA), from the hypodermis of the upper trunk may be used as a less invasive surrogate for SGNA to measure sympathetic outflow. We have recently shown that SCNA correlates well with SGNA and heart rate in normal ambulatory dogs.8 However, it is not clear whether SCNA can be used to predict the onset of ventricular arrhythmias and SCD in diseased animals. The aim of the present study was to use an established canine model of SCD and investigate the relationship between SCNA, SGNA, ventricular arrhythmias, and SCD.

Methods Surgical preparation We reanalyzed data from 7 ambulatory dogs with complete heart block, myocardial infarction, and nerve growth factor infusion to the left stellate ganglion from a previous study.2 The protocol was approved by the Institutional Animal Care and Use Committee of Cedars-Sinai Medical Center, Los Angeles, California. One pair of electrodes was used to record nerve activity from the left stellate ganglion; another pair of bipolar electrodes was implanted in the subcutaneous tissues of the right upper and left lower quadrant of the chest. In that study, the subcutaneous electrodes were placed for the purposes of electrocardiogram (ECG) recording.2 In the current investigation, signals from those electrodes were high-pass filtered and inspected for nerve signals. The recording was made in a bipolar mode, with 2 widely spaced bipoles. The electrodes were the stainless steel wires that came with the Data Sciences International D70-EEE radiotransmitter (DSI, St. Paul, MN). The terminal 5 mm of the wires was stripped of its insulation and used for electrical recording. Subcutaneous interelectrode distance was not measured at the time of the study, but in similar size dogs it is estimated at 28 cm. A detailed description of study methods is available in the Online Supplementary Material.

Results The simultaneous SGNA and SCNA recording lasted 43 ⫾ 26 days Presence and characteristics of subcutaneous nerve discharges Similar to a previous report in normal canines, all 7 dogs demonstrated subcutaneous nerve discharges with similar morphology to the signals recorded from the left stellate ganglion.8 In addition, SCNA morphology was similar to filtered skin and muscle sympathetic nerve activity obtained in microneurography studies.9–11 We previously described 2 SGNA patterns in this canine model: low-amplitude burst

Heart Rhythm, Vol 0, No 0, Month 2014 discharge activity (LABDA) with amplitudes between 0.05 and 0.8 mV and high-amplitude spike discharge activity (HASDA) with amplitudes of 0.9 ⫾ 0.16 mV.2 Of 366 randomly selected 15-second frames, 214 contained SGNA and 186 displayed SCNA. In 88% of frames, the presence or absence of SGNA correlated directly with the presence or absence of SCNA. There were frames containing subcutaneous but not stellate ganglion discharges, suggesting that the origin of SCNA is not inadvertent recording of SGNA with the subcutaneous electrodes. All subcutaneous discharges demonstrated LABDA pattern with amplitude of 0.07 ⫾ 0.08 mV. In all but 1 frame, which contained HASDA, SGNA also displayed LABDA pattern with amplitudes of 0.10 ⫾ 0.11 mV. Unlike the SGNA channel, the SCNA channel was more prone to display incompletely filtered ECG signals and pacing artifacts. Figure 1 shows 2 VT episodes from 2 different animals. In the first animal (1A), the raw signal is high-pass filtered to obtain noncontaminated SCNA. In the second animal (1B), the ECG signals from the raw signals could not be filtered well and contaminate the SCNA channel with ECG artifacts (downward arrows) despite high-pass filtering. The onset of the SCNA discharge is still visible (upward arrow); however, integration of SCNA (in mV-s) cannot be accurately accomplished because of unfiltered ECG signals. SCNA and SCD due to VF Two dogs died of SCD due to VF on postoperative days 3 and 52. Figures 2 and 3 show the recordings immediately before and after the onset of VF in these 2 dogs. In both dogs, VF was preceded by both SGNA and SCNA. The tracings in 2A and 2B as well as 3A and 3B are continuous with the 2A and 3A panels preceding the 3A and 3B panels by 40 seconds. LABDA discharges before VF are marked by downward arrows (Figures 2 and 3). Massive SGNA and SCNA (asterisks) occurred after VF, likely responses to acute reduction of blood pressure. Similar to Figure 1B, the dog portrayed in Figure 2 was not included in the quantitative nerve integration analyses because the SCNA channel was contaminated with ECG and pacing artifacts. SCNA and VT Two episodes of VF and 59 episodes of VT from 6 dogs (7– 12 per dog), occurring 23 ⫾ 17 days after surgery with an average heart rate of 156 ⫾ 44 bpm and duration of 20 ⫾ 89 seconds, were analyzed. A total of 75% of VT/VF episodes were preceded by SGNA and 70% by SCNA within 15 seconds of initiation. Of 61 15-second episodes of AIVR selected within 26 ⫾ 26 minutes, 59% contained SGNA and 43% had SCNA. By using a generalized linear mixed-effects model, the odds ratio of observing SGNA and SCNA 15 seconds before VT/VF compared to observing discharges during episodes of AIVR for a specific dog was 2.32 (95% confidence interval [CI] 1.01–5.31, P ¼ .0466) and 3.13 (95% CI 1.45–6.76, P ¼ .004), respectively. Figure 4A shows another representative episode of VT, along with a

126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 F1141 142 143 144 145 146 147 148 149 150 151 152 153 154 F2 QF3 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 F4181 182

Doytchinova et al 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 F5 226 227 228 229 230 231 232 233 234 235 236 237 238 239

Subcutaneous Nerve Activity and Arrhythmias

3

0.3 - 0.3 0.3 - 0.3 3

VT

-3 3 40 s

-3

0.3 - 0.3 0.3 - 0.3 8

VT

-8 8 -8

40 s

Figure 1 Stellate ganglion nerve activity (SGNA) and subcutaneous nerve activity (SCNA) in association with ventricular tachycardia (VT) from 2 different dogs. A: In the first dog, unfiltered (raw) signal (bottom) shows both the ECG and the high-frequency SCNA. The ECG was nearly completely removed by highpass filtering as shown in the SCNA channel. B: In a second dog, unfiltered (raw) signal (bottom) has large contamination by the ECG. High-pass filtering did not completely remove the ECG (downward arrows). However, increase in nerve activity over the baseline still is evident (upward arrow points at onset of SCNA) preceding VT. The SGNA and SCNA channels are filtered at 150 Hz high pass; the electrocardiogram (ECG) is filtered at 30 Hz low pass. The raw signal is unfiltered. Units for SGNA, SCNA, and ECG are displayed in millivolts.

control AIVR episode in the same animal. The average latency of onset from the beginning of the stellate and the subcutaneous discharge to the onset of VT/VF was 17.1 ⫾ 15.2 seconds and 17.8 ⫾ 18.4 seconds, respectively. Because the average VT/VF occurrence was 2.1 ⫾ 1.4 episodes per day, while intermittent LABDA from the subcutaneous tissues occurred throughout the day, o1% of SCNA LABDA episodes were followed by VT/VF. This is similar to data previously reported for SGNA discharges.2 Integrated SGNA (iSGNA) and integrated SCNA (iSCNA) were calculated in 3 20-second intervals for 60 seconds before 49 episodes of VT, 1 VF episode, and 50 20-second frames of AIVR in 5 dogs. By using a linear mixed-effects model, progressive increase in iSGNA and iSCNA was observed before VT/VF (Figure 5).

(95% CI 1.08–5.51, P ¼ .032) for SGNA and 2.47 (95% CI 1.09–5.61, P ¼ .031) for SCNA. The latency of onset from the beginning of the stellate ganglion and the subcutaneous discharges to the beginning of the FBG/C episode was 16.8 ⫾ 14.3 seconds and 17.8 ⫾ 18.4 seconds, respectively. Figure 6A shows an example of FBG/C preceded by SCNA and SGNA. Nerve activity was integrated 20 seconds before 50 episodes of FBG/C from 5 dogs and during 50 AIVR control episodes lasting 20 seconds. Using a linear mixed-effects model, both iSGNA [53.0 (95% CI 42.5–63.6) vs 35.9 (95% CI 29.1–42.7) mV-s, P ¼ .0054] and iSCNA [45.5 (95% CI 37.5–53.5) vs 34.8 (95% CI 28.1–41.4) mV-s, P ¼ .0176] were higher in the 20 seconds before FBG/C than during the 20-second frames of AIVR.

SCNA and frequent bigeminy or couplets Sixty-one episodes of frequent bigeminy or couplets (FBG/ C) from 6 dogs (9–12 per dog), occurring 21 ⫾ 15 days after surgery with an average heart rate of 77 ⫾ 14 bpm, were analyzed. Sixty-nine percent of FBG/C were associated with SGNA and 59% with SCNA within 15 seconds before initiation. In contrast, in 61 15-second frames of AIVR, occurring within 28 ⫾ 26 minutes, SGNA was recorded 51% of the time and SCNA was present 41% of the time. By using a generalized linear mixed-effects model, the odds ratio of detecting nerve discharges 15 seconds before FBG/C compared to during the AIVR period for a specific dog was 2.44

SCNA and premature ventricular contractions Sixty-one premature ventricular contractions (PVCs) from 6 dogs (9–12 per dog), occurring 22 ⫾ 15 days after surgery, were analyzed. Sixty-two percent of isolated PVCs were preceded by SGNA within 15 seconds and 61% by SCNA. Figure 6B shows an example of isolated PVC associated with SCNA and SGNA. In comparison, during 61 15-second frames of AIVR selected within 22 ⫾ 26 minutes, SGNA was recorded 36% of the time and SCNA was noted 30% of the time. By using a generalized linear mixed-effects model, the odds ratio of having SGNA and SCNA discharges 15 seconds before isolated PVCs vs during the 15-second

240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 F6273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296

4 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 T1 350 351 352 353

Heart Rhythm, Vol 0, No 0, Month 2014 0.3

11:22:24

- 0.3 0.3 - 0.3 8 40 s

-4 0.3

11:23:04

- 0.3 0.3 - 0.3 8 40 s

-4 8

11:23:14

p pp

pp

-4

10 s

Figure 2 Subcutaneous nerve activity (SCNA) and stellate ganglion nerve activity (SGNA) precede ventricular fibrillation (VF) and sudden cardiac death. A: Simultaneous firing of SGNA and SCNA is observed 40 seconds before VF. Downward arrows point to low-amplitude burst discharge activity (LABDA). Similar to Figure 1B, the subcutaneous channel in this animal is heavily contaminated by ECG artifacts; however, the presence of nerve discharges still is apparent. B: Continuous tracing from panel A 40 seconds later shows initiation of VF (upward arrow) and large discharges in the stellate and subcutaneous channels (asterisks) following VF. C: Ten-second tracing from the boxed area in B demonstrating the onset of VF in greater detail. Units for SGNA, SCNA and ECG are displayed in millivolts. p ¼ p waves.

periods of AIVR for a specific dog was 4.06 (95 % CI 1.69– 9.74, P ¼ .0019) for SGNA and 7.37 (95% CI 2.75–19.76, P ¼ .0001) for SCNA. Integrated SGNA and iSCNA 20 seconds before 50 isolated PVCs and during 50 20-second frames of AIVR were calculated in 5 dogs. Using a linear mixed-effects model, both iSGNA [50.5 (95% CI 39.7–61.3) vs 32.2 (95% CI 27.0–37.5) mV-s, P ¼ .0019] and iSCNA [45.2 (95% CI 36.0–54.3) vs 31.6 (95% CI 26.1–37.1) mV-s, P ¼ .0058] were higher in the 20 seconds before isolated PVCs compared to the 20-second control frames of AIVR. Correlation between SGNA and SCNA To investigate the correlation between iSGNA and iSCNA, the 150 20-second segments of integrated nerve activity 60 seconds before VT/VF, the 50 20-second segments of integrated nerve activity before FBG/C, and the 50 20second segments of integrated nerve activity before isolated PVCs along with their respective 20 seconds AIVR control periods described in the quantitative analysis earlier were combined for each individual dog. As shown in Table 1, the average Pearson correlation coefficient for all dogs was 0.65 ⫾ 0.16 (P o.0001 for each dog). When selecting the first 10 30-second frames during each hour of the day on the first available complete recording day from the same 5 dogs, a

similar average Pearson correlation coefficient of 0.73 ⫾ 0.18 (P o.0001 for all dogs) was obtained (Table 1). Circadian variation in SCNA To determine if circadian variation is observed in iSGNA and iSCNA, the first 10 30-second frames during each hour of the day on the first available complete recording day were selected for analysis and combined for all 5 dogs. As demonstrated in Figure 7, both iSGNA and iSCNA exhibited circadian variation (po2e-16 for both through generalized additive mixed-effects models).

Discussion The main finding of this study is that subcutaneous nerve discharges were recorded in all dogs studied and that, similar to SGNA, SCNA preceded the onset of ventricular arrhythmias and SCD.

Presence of subcutaneous nerve discharges We recently reported that in ambulatory and anesthetized normal dogs, SCNA recorded by bipolar subcutaneous electrodes correlates with the SGNA and can be used to estimate the sympathetic tone.8 In the present study, the presence of SCNA was demonstrated in all 7 dogs using

354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 F7394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410

Doytchinova et al 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467

Subcutaneous Nerve Activity and Arrhythmias 0.8

5

14:58:52

- 0.8 0.8 - 0.8 3 -3 0.8

40 s 14:59:32

- 0.8 0.8 - 0.8 3 -3 3

40 s 14:59:32

pp p pp p p p p

-3

10 s

Figure 3 Increased stellate ganglion nerve activity (SGNA) and subcutaneous nerve activity (SCNA) precede ventricular fibrillation (VF) and sudden cardiac death in another dog. A: Low-amplitude burst discharge activity in the stellate ganglion is evident 40 seconds before VF. At the same time, SCNA is intermittently quiescent and discharges resume 30 seconds before VF (downward arrows). B: Continuous tracing 40 seconds after panel A showing initiation of VF and massive stellate and subcutaneous discharges after VF onset (asterisks). C: Magnification of the boxed area in panel B showing the beginning of VF (upward arrow). The SGNA and SCNA channels are filtered at 150 Hz high pass; the ECG is filtered at 30 Hz low pass. Units for SGNA, SCNA and ECG are displayed in millivolts. ECG ¼ electrocardiogram; p ¼ p waves.

widely spaced bipolar electrodes placed in the upper trunk. Similar to that in normal dogs, SCNA in these diseased dogs had similar morphology, onset, and duration compared to SGNA, but the recordings were sufficiently different to eliminate significant cross-talk between the channels.8 In addition, SCNA signals were similar to filtered human microneurography signals obtained directly from a peripheral nerve.9–11 Therefore, we postulate that the origin of these discharges is mainly postganglionic sympathetic nerve activity directed to the skin and possibly the skeletal muscles below the hypodermis. The study by Robinson et al8 demonstrates that apamin injection in the stellate ganglion results in increased signal in the subcutaneous nerves. These findings suggest that SCNA represents peripheral output of increased sympathetic outflow. However, additional studies using central pharmacologic blockade with drugs such as clonidine or stellate ganglion ablation and measuring SGNA and SCNA should be undertaken to further elucidate the specific interplay between SCNA and SGNA.

SCNA, ventricular arrhythmias, and SCD Similar to SGNA, SCNA preceded SCD and the majority of ventricular arrhythmias. The association between SGNA and ventricular arrhythmias demonstrated in this study is similar to reports published previously.2,12 Both iSCNA and iSGNA

showed progressive increase before VT/VF consistent with our previous reports.2 Similarly, higher values for iSGNA and iSCNA were noted in the 20 seconds before PVCs and FBG/C compared to control periods of AIVR. These findings further demonstrate that, similar to iSGNA, iSCNA also is increased before ventricular arrhythmias and SCD. The latency from the onset of SGNA and SCNA to the development of VT and VF shown in the present study is about 17 seconds, slightly longer than the latency of 10 to 15 seconds reported for atrial arrhythmias.13 However, the latency is dependent in part on the magnitude of nerve discharge. Large and synchronized sympathetic discharges from the stellate ganglion (the high-amplitude spike discharge activity) may induce VT with very short (less than a few seconds) latency.2 In our study, episodes of paced rhythm contained pacing artifact contamination in the SCNA channel, and episodes of AIVR were used as control periods. Some reports demonstrate that AIVR is associated with increased sympathetic tone,14,15 whereas others propose that it is associated with increased vagal tone.16 Our study suggests that anywhere from 36% to 59% of AIVR episodes contained SGNA; however, we did not measure vagal nerve activity and lack recordings with periods of normal sinus rhythm to compare these results to. As such, conclusions about the autonomic nervous system influences on AIVR cannot be made based on our study.

468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524

6 0.3

SGNA

- 0.3 0.3

SCNA

VT

- 0.3 3

ECG

-3

Raw 3 signal- 3

40 s

0.3

SGNA

- 0.3 0.3

SCNA

- 0.3 3

ECG

-3 3

AIVR

Raw signal- 3

40 s

Figure 4 Examples of stellate ganglion nerve activity (SGNA) and subcutaneous nerve activity (SCNA) before the onset of ventricular tachycardia (VT) and in association with accelerated idioventricular rhythm (AIVR). A: Prolonged low-amplitude burst discharge activity (LABDA) recorded from the stellate ganglion and the subcutaneous tissues are present before the onset of several short runs of VT. B: LABDA associated with episodes of AIVR not preceding VT in the same canine. The maximum amplitudes of SGNA and SCNA are lower in this frame compared to A. Units for SGNA, SCNA, and ECG are displayed in millivolts. The SGNA and SCNA channels are filtered at 150 Hz high pass; the electrocardiogram (ECG) is filtered at 30 Hz low pass. The raw signal is unfiltered.

400

300

INA (mV-s)

525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581

Heart Rhythm, Vol 0, No 0, Month 2014

200

100

W E B 4 C / F P O

0

SGNA SCNA

AIVR 40 (35-44)

43 (35-50)

SGNA SCNA

- 60 s 49 (41-56)

49 (39-59)

SGNA SCNA

- 40 s 59 * (48-69)

62 * (46-78)

SGNA SCNA

- 20 s 69 † ‡ (54-84)

75 † ‡ (58-93)

Figure 5 Integrated stellate ganglion nerve activity (SGNA) and integrated subcutaneous nerve activity (SCNA) in mV-s before the onset of ventricular tachycardia (VT), ventricular fibrillation (VF), and accelerated idioventricular rhythm (AIVR). Progressive increase in both integrated SGNA and integrated SCNA is noted from 20-second periods of AIVR (n ¼ 50) to 20-second intervals measured 60 seconds, 40 seconds, and 20 seconds before initiation of VT (n ¼ 49) and VF (n ¼ 1) in 5 dogs. *P o.05 vs AIVR; †P o.05 vs –60 s; ‡P o.05 vs AIVR. Red bars signify the mean values and the upper and lower 95% confidence intervals. Numerical values underneath the x-axis represent the mean and upper and lower 95% confidence intervals of integrated SGNA and SCNA in mV-s. INA ¼ integrated nerve activity.

582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638

Doytchinova et al 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695

Subcutaneous Nerve Activity and Arrhythmias

7

0.1

- 0.1 0.1

- 0.1

FBG/C

3

-3

30 s

1.5 - 1.5 1.5 - 1.5 3

PVC

-3

30 s

Figure 6 Examples of stellate ganglion nerve activity (SGNA) and subcutaneous nerve activity (SCNA) before the onset of frequent couplets and a premature ventricular contraction in two different canines. A: Low-amplitude burst discharge activity (LABDA) is noted in the stellate and the subcutaneous channel in association with frequent couplets. B: Two LABDA episodes and 1 episode of high-amplitude spike discharge activity (HASDA, arrow) with amplitude of 1.1 mV recorded from the stellate ganglion and 2 episodes of LABDA from the subcutaneous tissues with lower amplitude are associated with the occurrence of a premature ventricular contraction in another canine. The SGNA and SCNA channels are filtered at 150 Hz high pass; the electrocardiogram (ECG) is filtered at 30 Hz low pass. Units for SGNA, SCNA, and ECG are displayed in millivolts. FBG/C ¼ frequent bigeminy or couplets.

Relationship between iSGNA and iSCNA A good correlation was found between iSGNA and iSCNA with an average Pearson correlation of 0.73 (range 0.50– 0.92). This is similar to values obtained in normal dogs.8 It is unclear why some dogs displayed better correlation than

others. Because this is a retrospective analysis of a study, which used the subcutaneous electrodes for recording of cardiac rhythm, it is possible that some of the subcutaneous electrodes were located more caudally within the boundaries of the left lower thorax than others. Because the density of

Table 1 Pearson correlation coefficients between integrated stellate ganglion nerve activity (SGNA) and integrated subcutaneous nerve activity (SCNA) Dog no.

Pearson correlation*

1 2 3 4 5 Average

0.48 0.64 0.80 0.83 0.52 0.65

P value o.0001 o.0001 o.0001 o.0001 o.0001

Pearson correlation† 0.67 0.50 0.91 0.92 0.66 0.73

P value

147.9

SGNA

o.0001 o.0001 o.0001 o.0001 o.0001

* Values for 20-second intervals of integrated SGNA and SCNA 60 seconds before ventricular tachycardia (n ¼ 49; 7–12 per dog) and ventricular fibrillation (n ¼ 1), 20 seconds before frequent bigeminy or couplets lasting for Z10 consecutive beats (n ¼ 50; 9–12 per dog), 20 seconds before premature ventricular contractions (n ¼ 50; 8–12 per dog), and during their respective 20-second frames of accelerated idioventricular rhythm used as control (n ¼ 150) combined for each individual dog in 5 dogs. Total number of integrated nerve activity frames ¼ 400. Values for dog no. 1 were adjusted in 4 episodes of VT for the presence of unfiltered ECG artifacts. † Values for 30-second intervals of integrated SGNA and SCNA during the first 10 frames of each hour over a 24-hour period combined for each dog in the same 5 dogs (total number of integrated nerve activity frames ¼ 1200; 240 per dog).

27.9 0

Hour

24

0

Hour

24

132 3 132.3

SCNA 32.3

Figure 7 Circadian variation in integrated stellate ganglion nerve activity (SGNA) and integrated subcutaneous nerve activity (SCNA) in mV-s. Both SGNA (A) and SCNA (B) exhibit circadian variation (po2e-16 for both). Solid lines represent mean integrated SGNA (A) and SCNA (B) averaged for the first 10 30-second segments of each hour over 24 hours and combined in 5 dogs. Dotted lines represent 95% upper and lower confidence intervals.

696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752

8 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809

sympathetic fibers originating from the stellate ganglion in canines decreases in a cranial to caudal direction with maximum concentration in the upper extremities, more caudal placement of the left thorax electrode could have resulted in worse correlation between iSGNA and iSCNA.5 Another possibility is that some dogs had a larger percentage of ECG contamination in the subcutaneous channel than others. Alternatively, there may be some variation in the density of synapses between the stellate ganglion and the subcutaneous tissues between animals, and for any specific point in the lower thorax, a larger percentage of sympathetic fibers could derive from the thoracic sympathetic chain rather than the stellate ganglion in some canines compared to others. Further studies are needed to determine the optimal location and distance between the electrodes for subcutaneous nerve recordings.

Sources of signals in the subcutaneous space Sources of recordings from the subcutaneous space could include signals from autonomic nerves, motor and sensory nerves, unfiltered ECG or pacing artifacts, and movement artifacts. High-pass (150-Hz) filtering eliminates a significant amount of muscle contractions,17 motion noise,18 and residual ECG signals. When excluding periods during which signal contamination with ECG and/or pacing artifact is pronounced, there is a significant correlation between SCNA, ventricular arrhythmias, and SCD and between SCNA and SGNA. In addition, consistent with previous reports observing circadian variation of sympathetic nerve activity in dogs, both iSGNA and iSCNA showed circadian variation.8,12,19 These collective findings suggest that in the absence of significant ECG and/or pacing contamination, the majority of signals recorded from the subcutaneous space after high-pass signal filtering are sympathetic in origin.

Study limitations The most significant limitation of this study is that incomplete filtering of ECG and pacing artifact was observed in the subcutaneous channel. The degree of signal contamination appeared to vary between dogs, possibly because of variation in the location of the recording device and its proximity to the recording electrodes. Another limitation is that we only recorded signals from the left stellate ganglion in this study, and that the DSI D70-EEE radiotransmitter is designed to record low-frequency signals, such as an ECG, and signals with frequencies higher than 250 Hz are not detected. However, our recent study recorded simultaneously from the right stellate ganglion and right thoracic subcutaneous tissues using equipment with wide bandwidth and high sampling rate. The results showed that both right SGNA and right SCNA correlated well with the sinus rate.8 It is possible that future improvement of recording equipment of the implanted devices will allow more effective use of SCNA for ventricular arrhythmia detection and risk stratification. Although high-pass filtering at 150 Hz eliminates the majority of muscle contractions and motion artifacts, it also

Heart Rhythm, Vol 0, No 0, Month 2014 eliminates nerve signals below 150 Hz.20 In addition, we do not have vagal nerve activity recording. Further studies should include vagal nerve measurements and attempt to determine the relationship between vagal nerve activity, SCNA, and ventricular arrhythmias.

Conclusion In the absence of significant ECG/pacing artifact contamination of the nerve recording channels, SCNA measured from the thorax could be used to estimate the left SGNA and predict the susceptibility to ventricular arrhythmias in a canine model of SCD.

Appendix Supplementary data Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.hrthm. 2014.11.007.

References 1. Zipes DP, Rubart M. Neural modulation of cardiac arrhythmias and sudden cardiac death. Heart Rhythm 2006;3:108–113. 2. Zhou S, Jung BC, Tan AY, Trang VQ, Gholmieh G, Han SW, Lin SF, Fishbein MC, Chen PS, Chen LS. Spontaneous stellate ganglion nerve activity and ventricular arrhythmia in a canine model of sudden death. Heart Rhythm 2008;5: 131–139. 3. Chakravarthy Marx S, Kumar P, Dhalapathy S, Anitha Marx C. Distribution of sympathetic fiber areas in the sensory nerves of forearm: an immunohistochemical study in cadavers. Rom J Morphol Embryol 2011;52:605–611. 4. Donadio V, Nolano M, Provitera V, Stancanelli A, Lullo F, Liguori R, Santoro L. Skin sympathetic adrenergic innervation: an immunofluorescence confocal study. Ann Neurol 2006;59:376–381. 5. Taniguchi T, Morimoto M, Taniguchi Y, Takasaki M, Totoki T. Cutaneous distribution of sympathetic postganglionic fibers from stellate ganglion: a retrograde axonal tracing study using wheat germ agglutinin conjugated with horseradish peroxidase. J Anesth 1994;8:441–449. 6. Baron R, Janig W, With H. Sympathetic and afferent neurones projecting into forelimb and trunk nerves and the anatomical organization of the thoracic sympathetic outflow of the rat. J Auton Nerv Syst 1995;53:205–214. 7. Porzionato A, Macchi V, Stecco C, Loukas M, Tubbs RS, De Caro R. Surgical anatomy of the pectoral nerves and the pectoral musculature. Clin Anat 2012;25: 559–575. 8. Robinson EA, Rhee KS, Doytchinova A, et al. Estimating sympathetic tone by recording subcutaneous nerve activity in ambulatory dogs. J Cardiovasc Electrophysiol 2014; [Epub ahead of print]. 9. Diedrich A, Charoensuk W, Brychta RJ, Ertl AC, Shiavi R. Analysis of raw microneurographic recordings based on wavelet de-noising technique and classification algorithm: wavelet analysis in microneurography. IEEE Trans Biomed Eng 2003;50:41–50. 10. Kamijo Y, Okada Y, Ikegawa S, Okazaki K, Goto M, Nose H. Skin sympathetic nerve activity component synchronizing with cardiac cycle is involved in hypovolaemic suppression of cutaneous vasodilatation in hyperthermia. J Physiol 2011;589(Pt 24):6231–6242. 11. Salmanpour A, Brown LJ, Shoemaker JK. Spike detection in human muscle sympathetic nerve activity using a matched wavelet approach. J Neurosci Methods 2010;193:343–355. 12. Ogawa M, Zhou S, Tan AY, Song J, Gholmieh G, Fishbein MCLH, Siegel RJ, Karagueuzian HS, Chen LS, Lin SF, Chen PS. Left stellate ganglion and vagal nerve activity and cardiac arrhythmias in ambulatory dogs with pacing-induced congestive heart failure. J Am Coll Cardiol 2007;50:335–343. 13. Tan AY, Zhou S, Ogawa M, Song J, Chu M, Li H, Fishbein MC, Lin SF, Chen LS, Chen PS. Neural mechanisms of paroxysmal atrial fibrillation and paroxysmal atrial tachycardia in ambulatory canines. Circulation 2008;118:916–925. 14. Chiladakis JA, Pashalis A, Patsouras N, Manolis AS. Autonomic patterns preceding and following accelerated idioventricular rhythm in acute myocardial infarction. Cardiology 2001;96:24–31.

810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 Q7 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866

Doytchinova et al 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895

Subcutaneous Nerve Activity and Arrhythmias

15. Waxman MB, Cupps CL, Cameron DA. Modulation of an idioventricular rhythm by vagal tone. J Am Coll Cardiol 1988;11:1052–1060. 16. Bonnemeier H, Ortak J, Wiegand UK, Eberhardt F, Bode F, Schunkert H, Katus HA, Richardt G. Accelerated idioventricular rhythm in the post-thrombolytic era: incidence, prognostic implications, and modulating mechanisms after direct percutaneous coronary intervention. Ann Noninvasive Electrocardiol 2005;10:179–187. 17. Tuler SM, Febles D, Bowen JM. Neuromuscular effects of chronic exposure to fenthion in dogs and predictive value of electromyography. Fundam Appl Toxicol 1988;11:155–168.

9

18. Chowdhury RH, Reaz MB, Ali MA, Bakar AA, Chellappan K, Chang TG. Surface electromyography signal processing and classification techniques. Sensors (Basel) 2013;13:12431–12466. 19. Jung BC, Dave AS, Tan AY, Gholmieh G, Zhou S, wang DC, Akingba G, Fishbein GA, Montemagno C, Lin SF, Chen LS, Chen PS. Circadian variations of stellate ganglion nerve activity in ambulatory dogs. Heart Rhythm 2006;3:78–85. 20. Malpas SC. Sympathetic nervous system overactivity and its role in the development of cardiovascular disease. Physiol Rev 2010;90:513–557.

CLINICAL PERSPECTIVES Sympathetic tone is important in cardiac arrhythmogenesis. Because the sympathetic nerve structures, such as stellate (cervicothoracic) ganglion are located deep in the thoracic cavity, it is difficult to directly record from those structures to determine the sympathetic tone. Heart rate variability and microneurography techniques have been used to assess sympathetic tone in patients. However, technical difficulties have prevented routine use of those methods for arrhythmia prediction and risk stratification. Subcutaneous tissues contain sympathetic nerve fibers that originate from the stellate ganglion. We found that it is possible to record the subcutaneous nerve activities with implanted electrodes and radiotransmitters in ambulatory dogs. In addition, we found that subcutaneous nerve activities correlate well with stellate ganglion nerve activity, and that subcutaneous nerve activities precede the onset of ventricular tachycardia and fibrillation. Because subcutaneous tissues are much more accessible than the thoracic cavity, the ability to directly record sympathetic nerve activities from the subcutaneous tissues may provide a new approach to estimate sympathetic tone. These methods can be translated into clinical practice by incorporating them in the implantable loop monitors, pacemakers, or implantable cardioverter-defibrillators with high sampling rate and wide bandwidth. The nerve signals and the electrocardiographic signals can be separated by different filtering settings. The data can be analyzed to determine whether subcutaneous nerve activities precede the onset of spontaneous ventricular arrhythmias and sudden death. These clinical studies may lead to improved arrhythmia prediction in patients with implanted devices.

896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923