Percutaneous left stellate ganglion block for refractory ventricular tachycardia in structural heart disease: our single-centre experience

Ventricular arrhythmia (VA) can develop in apparently stable patients with structural heart disease due to subclinical neurohumoral hyperactivation and ventricular remodelling [1]. The most severe form of VA, electrical storm (ES), constitutes an important cause of morbidity and mortality, carrying a relative mortality risk of 3.15 [2]. Guideline-recommended first-line treatment aims to eliminate VA triggers and modulators by suppressing the cardiac sympathetic tone using a combination of beta-blockers, amiodarone and/or sedation [3]. Urgent catheter ablation (CA) and/or deep sedation are recommended when the former fail to control VA. However, CA might not be feasible for a variety of reasons, such as haemodynamic instability, substrate not amenable to ablation, ventricular thrombus or left-sided mechanical valvular prosthesis. Logistical considerations may also be a limiting factor in some centres. The temporary nature of deep sedation implies that it is primarily a bridge to more definitive treatment. The 2022 ESC guideline now classifies modulation of the autonomic nervous system (ANS) as a class IIB recommendation in selected patients with refractory VA ([3]; Fig. 1).

The pro-arrhythmogenic effects of the sympathetic branch of the ANS in structurally abnormal hearts have been described extensively [4]. Neural remodelling results in areas of sympathetic hyperinnervation and relative denervation predisposing to autonomically mediated heterogeneity of refractory periods. Cellular electrical remodelling predisposes to neurotransmitter-mediated intracellular Ca²⁺ overload and delayed afterdepolarisations. The sympathetic fibres innervating the ventricular myocardium derive from the ganglia of the cervicothoracic paravertebral chain, primarily from the left and right stellate ganglia (C7‑8 to Th1-2) [5, 6]. Surgical excision of the left stellate ganglion by means of video-assisted thoracoscopic left cardiac sympathetic denervation (LCSD) has been shown to reduce the VA burden in catecholaminergic polymorphic ventricular tachycardia and congenital long QT syndromes [7, 8]. More recently, a reduction in VA burden has also been described following percutaneous stellate ganglion block (PSGB) in patients with structural heart disease [9‐15]. This percutaneous approach has the advantage of being a bedside intervention using ultrasound guidance, while major complications are rare [16]. Although its effects are temporary, expected to last no more than 2 weeks, PGSB provides a time window for scheduling definitive therapy. Since 2018, PSGB and LCSD procedures have been performed at our centre in selected patients with refractory VA.

Twelve patients were identified who underwent PSGB for refractory VA and underlying structural heart disease during the study period. Median follow-up was 601 days (± 529). All patients were men, aged between 51 and 85 years of age, with monomorphic VT (Tab. 1). Gender did not play a role in patient selection for PSGB, and autonomic modulation has been performed for various indications in both genders at our centre. Time between either initial diagnosis of a cardiomyopathy or first myocardial infarction and PSGB was a mean of 19 years (± 8.5 years). The majority of patients (75%) had ischaemic cardiomyopathy with the remainder having a non-ischaemic dilated cardiomyopathy. Median left ventricular ejection fraction was 32% (± 12%). Endocardial VT ablation had previously been performed in 9 of 12 patients (75%). In the remaining 3 patients prior VT ablation had been deferred either due to comorbidity or because important technical limitations were expected due to patient-related factors. At baseline, 8 of 12 (67%) patients were already being treated with the combination of both beta-blocker and amiodarone. Three were on sotalol, and one patient was treated with bisoprolol (Tab. 2).

After PSGB a reduction in the absolute number of ATP episodes and ICD shocks was seen in 10 of 12 patients (83%) (Fig. 2), which was significant (p = 0.04) In the entire cohort the decrease in ATP, although showing a clear trend, did not reach statistical significance (p = 0.066), whereas ICD shocks did decrease significantly (p = 0.028). One patient ultimately developed refractory incessant VT after an initial reduction in VA (patient 12). In the presence of advanced heart failure and severe comorbidities treatment was discontinued. In total, 4 of 12 (33%) patients died during long-term follow-up at a median of 555 days (54–1327). The causes of death were refractory VA (one case), worsening heart failure/malignancy (one case) and complicated pneumonia (two cases) (Tab. 3).

In three patients a second or even third PSGB was performed as a bridge to LCSD after VA recurred more than a week after the initial block. Two PSGBs resulted in a VT-free interval of 12 and over 14 days, allowing LCSD to be scheduled. In the third patient VT recurred at day 3 in the context of a COVID-19 infection.

Following PSGB one patient complained of temporary hoarseness. Anisocoria was reported in 2 of 12 (17%) blocks. In one patient temporary hypotension during pre-existing therapy-refractory incessant VT was seen immediately following PSGB but resolved after ‘spontaneous’ conversion to sinus rhythm 2 h later. No vascular complications were observed.

LCSD was performed in five patients following PSGB. Antiarrhythmic drugs had been unchanged in the weeks before and following surgery. One patient with cessation of VA following PSGB had a recurrence of ES 8 days after LCSD. We are unable to explain this difference in response, as his treatment had otherwise been unchanged before and following both PSGB and LCSD. In one patient with a mid-septal VT circuit not amenable to ablation LCSD did not reduce the VA burden, but higher amiodarone levels resulted in adequate VA suppression. The other three patients had substantial reduction of their VA burden with one remaining completely free of VA for 15 months, but he passed away due to a combination of disseminated bladder cancer and worsening heart failure. A second patient with 2281 instances of ATP in the 6 weeks preceding LCSD had only 1 h of recurrent VA terminated by ATP 13 days after LSCD. Only non-sustained VT was seen 6 months after LCSD. After these 6 months VT occurred only infrequently, but could be terminated by ATP. The third patient had only two episodes of VT terminated by ATP in the year following LCSD and a recurrence of ES a year later.

We found PSGB to be efficacious in stabilising recurrent drug-refractory VA in the majority of patients with severe structural heart disease in whom urgent CA was deemed to be either technically infeasible or to carry an unacceptably high risk. We consider the comparative simplicity of the procedure and low risk of major complications to be advantages of PGSB. Another potential benefit of PGSB is identifying patients who could benefit from LCSD for long-term suppression of VA.

There are several limitations to the data presented here. With antiarrhythmic drugs and CA being the preferred treatment for VA, there is an implicit selection bias in opting for autonomic modulation. It was not feasible to create a matched historical cohort. Thus, in the absence of a control population the reduction in VA burden might in part be attributed to a late effect of the antiarrhythmic drugs administered. However, with two-thirds of patients already receiving long-term treatment with amiodarone/beta-blocker, an antiarrhythmic effect of additional amiodarone would not be expected. The absence of females in this series is also notable. Gender was not a factor when considering patients for autonomic modulation.

We speculate that some non-responders might have benefitted from a subsequent right-sided PSGB, as animal research has demonstrated that in canines the left and right stellate ganglion innervate different segments of the myocardium. It is plausible that the sympathetic innervation pattern is similar in humans [6, 21‐23]. Secondly, in non-responders, e.g. those with scar-related arrhythmias, the ANS might be only a minor modulator. Finally, ineffective blockade of the stellate ganglion might be a factor in non-responders. However, this is unlikely, as extensive spread of agents injected into the paravertebral space has been previously demonstrated [26‐29]. Methods to quantify cardiac autonomic tone would certainly be helpful in differentiating between the above hypotheses. We found traditional ANS markers, such as QTc dispersion and heart rate variability, to be impracticable in our patients due to the presence of frequent ventricular ectopia and/or permanent (bi-)ventricular pacing or antiarrhythmic drugs. MIBG (metaiodobenzylguanidine) scans were also logistically infeasible in unstable patients. Further research into novel methods, such as sensory skin nerve activity, are of particular interest for evaluating the response to PSGB and potentially selecting patients who could benefit from autonomic modulation.

In this retrospective series, left-sided PSGB showed significant promise as a treatment to temporarily suppress VA and ES in patients with extensive structural heart disease considered not amenable to urgent CA. Cessation of VA was seen in several patients for whom no other evidence-based acute treatment options were available, allowing definitive treatment to be scheduled. Further randomised studies are required to determine the role PSGB might play in the treatment of VA. Moreover, methods for quantifying the change in cardiac sympathetic tone, which has been achieved following a percutaneous block, are of particular interest.