Introduction

Refractory ventricular arrhythmias (VAs) represent a life-threatening acute-care emergency condition and pose a major therapeutic challenge. Treatment for ventricular tachycardia (VT)/ventricular fibrillation typically involves suppression with antiarrhythmic drugs, catheter ablation, or both (1). Even if catheter ablation has made tremendous strides in terms of technology, recurrence rates after ablation sometimes exceed 50% (1–3), which might be due to the anatomic and physiological constraints.

Because the autonomic nervous system powerfully regulates myocardial electrical propagation, neuromodulation approaches have been used to manage VT in case of unsuccessful radiofrequency (RF) ablation (4–6). The most powerful and long-term strategy to decrease the burden of VA is cardiac sympathetic denervation (CSD). This technique was developed by experienced cardiothoracic surgeons and now is performed with extensive resection of the sympathetic chains and, when possible, with part of the bilateral stellate ganglia (SG) (7–28). A major limitation of surgical CSD is that many centers do not have enough experience in this operation.

This technique can cause various complications, such as those related to the partial resection of SG (Horner’s syndrome) (5,29) or other major complications such as pneumothorax or major bleeding (5). Even minor complications related to neuraxial change have been reported (4).

In the current study, we sought to evaluate a new, less invasive, and faster CSD technique (sympathicotomy T2–T5) to address the neuromodulation in the setting of untreatable VT. Previous experience on a similar sympathicotomy procedure has been described for palmar hyperhidrosis (30,31)

Methods

This is a single-center retrospective analysis of consecutive patients enrolled since the Neuromodulation VT program started in 2019 in our hospital. The study protocol was approved by the local ethics committee of the Fatebenefratelli Hospital and Policlinico Umberto I of Rome. Written informed consent from all patients was obtained before surgery.

Study design

All patients with refractory VT were enrolled in our VT program at Fatebenefratelli Hospital, Isola Tiberina, Rome. In case of ineffective VT ablation or technical problems in the RF standard approach, CSD was proposed to the patient after a collegial evaluation with our thoracic unit (Policlinico Umberto I, Sapienza University). Refractory VT was defined as VT recurrence despite the combined antiarrhythmic therapy and prior catheter VT ablation attempts.

VT ablation procedure before CSD was conducted under general anesthesia. A 3-dimensional map was obtained with a multipolar mapping catheter in sinus rhythm and during VT, when possible. When the tachycardia was not inducible, a sinus rhythm activation map was obtained to target the deceleration zones. Substrate ablation/modification was performed targeting noninducibility of all VTs. RF power was set at 50 W with an irrigated catheter aiming at significant local impedance drop. The VT was considered nonmappable when systolic blood pressure was <85 mm Hg with the infusion of inotropic/vasoactive agents (noradrenaline and adrenaline).

Data collection

Medical records, demographic data, cause of VT, clinical event, number of VT therapies, type of VT, ejection fraction (EF) (%), medications, prior VT ablation, procedural details, and follow-up data were acquired. Data on preoperative arrhythmias were obtained from internal cardioverter defibrillator (ICD) logs and were reviewed carefully and manually by 2 expert electrophysiologists (F.M.C. and P.R.). Due to the small cohort, data on VTs are descriptive for each patient and are schematically summarized in Tables 1 and 2. VT burden quantification was calculated as percentage of the total number of VT 90 days before and after CSD.

CSD surgical technique

All the procedures were conducted by the same operator (M.A.). Under GA with a double-lumen endotracheal tube, the patient was placed on the surgical table in a supine position and half seated with both arms abducted to expose the axillary cave; the head was fixed to the headrest with bandages (Figure 1). General endotracheal anesthesia was induced with midazolam 2 to 3 mg, Propofol 1.5 mg/kg, fentanyl 4 μg/kg, and neuromuscular blockade achieved with rocuronium 0.6 mg/kg (Figure 1). General anesthesia was maintained with the inhalational agent sevoflurane at an end tidal concentration of 1.8% to 2.0%, with supplemental boluses of fentanyl 1 μg/kg as needed. The blood arterial pressure was invasively monitored. The arrhythmia surveillance during the whole procedure was controlled by the electrophysiology team through the patient’s ICD. Due to the predominance of left sympathetic chain in the heart innervation, we first performed the procedure on the left side. With an 11 blade, we made a skin incision in the axillary cave at the level of the second intercostal space and, inserted a 7-mm port after instituting one-lung ventilation. As a first step, with a 5 mm 0° camera, we identified the sympathetic chain under the thin pleura and the SG on the posterior portion of the chest wall close to costovertebral articulation (Figure 2). The half-seated position was extremely helpful because it facilitated the access to the lung and good exposure of the posterior portion of the chest wall. Through the same incision, next to the camera, we inserted an electrified hook with 80 W power (5-mm L-Hook-Electrode Laprosurge, Watford, Herts, United Kingdom). Upon positioning the hook perpendicularly to the sympathetic chain, we made a deep cut in the nervous fibers (until the intercostal muscle layer) from the first to the fifth metamere without affecting the SG and without the removal of the nerve; this method allowed the procedure to be faster and probably safer. Finally, accessory fibers, when visible, were equally burned. Immediately on completion of the left side, we performed the same procedure on the right side. At the end, a small drainage (15 Charriere) was left inside the pleural cavity to allow the correct re-expansion of the lungs. This drainage was removed after a chest X-ray 2 h later. When the sympathetic chain was less visible (i.e., in patients with obesity or in the presence of adhesions), through the same approach, the parietal pleura near the costovertebral articulation was opened and the nerve fibers were identified and cut. At the end of the procedure, all patients were extubated in the operation room and transferred to the ward. Postoperative analgesia was administered through 1 g of paracetamol, only at the patient’s request. Beta-blockers as well as the hypotensive drugs were discontinued 12 to 24 h before the procedure and reintroduced the day after if not contraindicated. Antiarrhythmic drugs were not interrupted for the procedure and were managed by referring electrophysiologists.

Follow-up

Patients were followed up clinically and by ICD interrogation or home monitoring (3 of 5 patients) at regular intervals every month. ICD devices were programmed as follows: Zone 1 (monitor) 130 to 180 beats/min; Zone 2 (VT) 180 to 220 beats/min, antitachycardia pacing (ATP) (3) + shock; and Zone 3 (ventricular fibrillation) >220 beats/min, ATP (1) + shock. Recurrent VT was defined as sustained VT recorded by the device, VT that required hospitalization, or ICD shock. Zone 1 (monitor) was considered as recurrence unless the duration was <30 s. Two expert electrophysiologists (F.M.C. and P.R.) carefully reviewed ICD logs for arrhythmia burden by home monitoring every month after the CSD. Patients were considered free of recurrence if they had not experienced a clinical recurrence, appropriate ICD shocks, or monitored episodes of sustained VA after the CSD hospitalization discharge. Data regarding noncardiac hospitalization were carefully analyzed.

Results

Case details are presented below and summarized in Tables 1 and 2. Four of 5 patients underwent bilateral sympathicotomy. In patient four, only left CSD was performed because of the presence of strong adhesions in the right pleural cavity. The mean systolic blood pressure decrease during CSD was 5 ± 2 mm Hg. No intraoperative or perioperative complications related to the surgery were observed. The mean ± SD operation time for each side was 18 ± 11 min including bed positioning and wound dressing after drainage tube positioning. None of the patients required blood transfusion. None of the patients developed Horner’s syndrome or compensatory hyperhidrosis in this series. All patients were discharged with a median length of stay of 3 days (range 2 to 5 days). One patient reported some alterations in the sweating pattern; in particular, an excessive dryness of the hands that resolved after the second month. No beta-blocker down titration was required for hypotension after the procedure and during the follow-up. None of the patients received shocks at follow-up after the discharge. The numbers of VTs trends before and after CSD are plotted in the Central Illustration.

Case 1 nonischemic dilated cardiomyopathy

A 69-year-old man was transferred to our department for incessant monomorphic VT that led to multiple shocks despite the treatment with amiodarone. Echocardiogram revealed a mild reduction in EF (45%). Initially, the patient underwent a RF endocardial procedure. After biventricular mapping, VTs were targeted on the interventricular septum. Multiple morphologies were interrupted during RF ablation and the patient was not inducible after final programmed electrical stimulation (PES). After the first procedure, the patient continued experiencing monomorphic VT. A second RF bipolar procedure was also ineffective. The patient continued experiencing slow monomorphic incessant VT with left bundle branch block morphology (superior and inferior axis). Subsequently, the patient was treated with alcohol septal ablation with acute success (VT termination during alcohol infusion). After 1 week, the patient developed recurrences with high burden VT. Mexiletine was introduced in the evaluation, and he was thus referred for CSD (1 month after the last procedure). Six hours post-CSD, the patient developed slow monomorphic ventricular tachycardia (MMVT) (155 ppm) that spontaneously ceased after 3 h. No change in drugs was made at discharge (amiodarone–mexiletine). The patient had no recurrences in the follow-up (11 months). He developed transient but excessive dryness in his hands, which resolved after the second month. After the first month of follow-up, mexiletine was suspended. No shocks were detected in ICD logs. VT burden reduction was 98.7%.

Case 2 nonischemic dilated cardiomyopathy/LMNA

A 56-year-old man was referred to our institution for repeated monomorphic VT leading to multiple shocks. Echocardiogram EF was normal (55%). During RF ablation, multiple septal VT morphologies were targeted, and the patient was not inducible at the end of the procedure (third extra stimulation). The arrhythmic storm occurred 2 days after the RF ablation. He was put on amiodarone and developed thyrotoxicosis. Meanwhile, a laminopathy was also identified (Glu161Lys–gene LMNA). He underwent the ablation twice in another center (2 bipolar ablations) with no reduction in VT burden despite antiarrhythmic drugs (sotalol and mexiletine). He was thus referred for bilateral CSD (5 months after the last procedure). No complications occurred after CSD. VA burden was dramatically reduced (100%). He continued to have a low burden (4 per month, maximum of 10 beats) of nonsustained VT but did not require ICD therapies in the follow-up (10 months).

Case 3 ischemic dilated cardiomyopathy

An 82-year-old man with dilated ischemic cardiomyopathy was referred for RF ablation in our center due to slow monomorphic VT under amiodarone leading to hemodynamic decompensation and recurrent hospitalization for heart failure. Admission echocardiogram revealed a severe depression of EF (25%) with severely dilated left ventricle that required levosimendan infusion 24 h prior to the ablation. His PAINESD risk score was 26. During the mapping phase, the patient spontaneously developed multiple VT morphologies that rapidly led to hemodynamic decompensation despite a high dose of noradrenaline and adrenaline. The procedure was interrupted and deferred with extracorporeal membrane oxygenation support. After a collegial meeting, CSD was proposed to the patient after analyzing the risks and benefits of the 2 procedures (RF ablation with extracorporeal membrane oxygenation vs. CSD). The patient was thus referred for CSD (3 months after the procedure). It should be noted that after the right resection of T4, a slow MMVT (140 ppm) occurred that was treated intraoperatively with lidocaine and manual ATP. The patient was discharged with amiodarone therapy. He developed 3 more episodes of slow monomorphic VT in the follow-up period (2 resolved spontaneously during hospitalization for pneumonia 20 days post CSD, lasting 4 to 6 h, and 1 was treated with external Direct Current (DC) shock in the emergency department the second month). No ICD shocks were detected on ICD logs (10 months). VT burden reduction was 80%.

Case 4 nonischemic dilated cardiomyopathy

A 68-year-old man with a history of frequent shocks related to rapid MMVT was referred to our center. His echocardiogram at admission showed a mild left ventricular dysfunction. A coronary angiogram revealed normal coronary arteries. He was thus considered for RF ablation. An endocardial biventricular map was performed that excluded any endocardial scar. No areas of late potentials (LPs) or local abnormal ventricular activities (LAVAs) were recorded. PES was performed. Two different, nontolerated tachycardias were inducted; both were treated with external cardioversion. Due to the hemodynamic properties of the 2 VTs and the absence of sinus rhythm targets, the procedure was aborted. After amiodarone’s failure, hydroquinidine was attempted with no results (2 episodes in 1 month). After 2 months, CSD was performed. This patient underwent only left sympathetic denervation for strong adhesion in the right pleural cavity. The patient was discharged with no complications under hydroquinidine and metoprolol. No ICD shocks were detected on ICD logs at follow-up (9 months). VT burden reduction was 75%.

Case 5 nonischemic dilated cardiomyopathy

A 62-year-old patient with a previous diagnosis of myocarditis was referred to our center for multiple, drugs refractory VT storms. An epicardial sinus rhythm map showed a confined spot of LAVAs/LPs between 2 obtuse marginal branches. Clinical VT was suddenly interrupted at the inferior border of the LAVA/LPs area (area of deceleration in sinus rhythm) in the epicardium. After LP abolition (80% of the LP area abolished due to coronary contiguity <5 mm), the patient was not inducible at PES (S1-S2-S3-S4). The left ventricular endocardial map performed after the ablation revealed no area of LAVA/LPs and a confined unipolar scar. He was discharged with beta-blockers therapy only. After 2 months, the patient was readmitted to his local hospital for a new episode of VT storm treated with amiodarone and lidocaine. After a collegial evaluation, a new endo-epicardial procedure (considering the clinical circuit proximity to coronaries branches) was deferred in case of a CSD failure. Bilateral CSD was performed with no complications (2 months after the procedure) and the patient was discharged. No ICD shocks were recorded on ICD logs at follow-up (2 months). Data about %VT burden reduction was not calculated due to limited follow-up.

Discussion

CSD seems to be an effective adjunctive therapy in patients with structural heart disease and refractory VT. Despite promising results, CSD is associated with a relatively high rate of intraprocedural and post-procedural complications (probably underestimated because they are sometimes minor and temporary), such as pneumothorax, hemothorax, Horner’s syndrome, and severe hypotension, therefore, limiting this technique to specialized centers. In this article, we sought to describe and evaluate a new technique to use in the setting of refractory VT with encouraging results although tested on a small sample.

The surgical impact of this technique is lighter (uniportal access, no pleural dissection, and no resection of the SG) and surgical time is faster compared with other experiences reported in the published data (7,10,32). This allows a fast CSD in all the patients treated, which results in a quick discharge without significant complications, which is of particular importance in patients with poor prognosis.

As strongly demonstrated by the literature, SG plays an important role in arrhythmogenesis, however, a previous study with sparing of the SG documented positive clinical outcomes (33). The technique used in the present article (in which the inferior 1/3 of the ganglia is spared) permits large postganglionic denervation leaving the SG synapses intact, which might ultimately lead to nerve sprouting and reinnervation in the long term (34–36).

It is to be noted that a high percentage of in-hospital VT recurrences (<12 h from the procedure) is not indicative of arrhythmia recurrences in follow-up. Early recurrences were never reported in the existing literature. The in-hospital recurrence of case 1 (< 6 h) and case 4 (<12 h) and the intraoperative VT of case 3 (2 min after the right chain sympathicotomy) highlight the possible massive neuraxial activation during and soon after the resection. Moreover, as stated in a previous article (16), the antiarrhythmic effects of the denervation of CSD, although present to some extent immediately after denervation, become more complete with chronic denervation. This will explain the optimal results in the follow-up despite the in-hospital recurrence. Moreover, in some patients, antiarrhythmic combined therapies were down-titrated and interrupted due to the impressive results in terms of burden reduction.

We are confident that this operation can be considered as a reasonable alternative to the common surgical technique to allow the worldwide spread of neuromodulation through CSD. Further studies of a large sample size are needed to highlight the possible benefits of this technique. Sympathicotomy can be adopted in patients with multiple RF failed ablation attempts or in high-risk patients in whom the RF ablation cannot be definitive.

Study limitations

This preliminary experience faces limitations of a single-center, nonrandomized, retrospective cohort study with a small sample size. Although the sympathetic chain and most of the accessory fibers (duplication of the chain/Kuntz Nerve and others) are usually visible beneath the parietal pleura, some of them can be thin and far away from the main chain and less visible. Those fibers (if they remain intact) can represent alternative pathways enabling the conduction of electrical impulses through different routes that can bypass the sympathetic chain even when surgically interrupted, virtually creating the basis for long-term recurrences. Effectiveness and clinical impact of nerve regeneration and nerve sprouting are debated, even in the case of a massive chain disconnection. The reported follow-up period of this case series may not adequately answer the clinical impact of these issues. In this case series, surrogates of cardiac denervation were not used. Moreover, 2 of the 4 patients did not undergo VT ablation for the previously mentioned reasons. This would lead to a predicted reduction in the efficacy of the CSD.

Conclusions

Our case series suggest that sympathicotomy with this novel modified technique may be considered as an adjunctive promising therapeutic option for refractory VT. Its use can be adopted for the treatment of all patients whose RF ablation fails or when the ablation implies high intraprocedural or periprocedural risks.

Footnotes