Efficacy of CSD:
Effect on ICD or external shocks: The average number of shocks
received by a patient prior to CSD was compared to shocks received after
the CSD at the end of follow up. It showed a very significant decrease
from a 24±37 {median 16(IQR 11-25)} prior to CSD to 2±4 {median 0(IQR
0-2); p<0.01}. Thus, CSD led to a 92% reduction in
requirement of defibrillation shocks.
Freedom from shock or death : At the end of 2 years, 46
(70.8%) patients were free of any sustained VT requiring ICD or
external shock(Figure 1).However, at the end of follow-up (27±23
months), this number had decreased to 43 (66%).
There were total 11 (16.9%) deaths in the complete cohort of patients
up to the end of follow-up. Four patients died during the index
hospitalization for CSD, 3 related to progressive pump failure and
congestive heart failure and one patient due to electro-mechanical
dissociation soon after the CSD. Another 4 patients died in the first
year of CSD, 2 because of pump failure and 2 because of sepsis. Thus,
freedom from death at the end of 1 year was87.7% and at 2 years was
83.1% (Figure2) Three patients died subsequently during follow-up, 2
because of heart failure and 1 due to VT recurrence. Freedom from a
combined end point of ICD shock or death at the end of two years was
52.3% (Figure 3).
Patient characteristics associated with outcome: We observed from
the Kaplan-Meier curve that the number of shocks declined consistently
till 1 year and after that the slope seems to have stabilized.
Therefore, we tried assessing patient characteristics which were
associated with ICD shock, as well as the combined endpoint of ICD shock
and death were assessed whether certain populations were less likely to
benefit from CSD. Kaplan-Meier analysis did not show any difference in
outcomes between patients with ischemic cardiomyopathy and non-ischemic
etiology [Log Rank (Mantel-Cox) p = 0.435] in relation to ICD shock
recurrence or the combined end-point of ICD shock and death(Figure 4).
The only parameter which was shown to have significant association with
combined end point of ICD shock and death was advanced NYHA class [Log
Rank (Mantel -Cox) p = 0.023], as shown in Figure 5.
There was no significant change in the LVEF at one year after CSD. In 5
patients the frequency of VT remained the same after CSD and thus they
were taken up for catheter ablation. Of these, 2 patients had a
significant decrease in VT episodes. Three other patients continued to
have episodes of VT and requirement of ICD shocks. One of these patients
diedduring follow-up because of VT storm.
Device interrogation data and follow-up: Many of our
patients on subsequent follow-up had shown runs of non-sustained VTs and
VTs terminated by ATP. However, we did not analyse this data in detail
for our current study. All the patients were thoroughly followed up
through clinic visits, telephonic interviews, telemetry with periodic
ICD interrogation. The anti-arrhythmic drugs were later stepped down
according to clinical status and VT burden.
Complications: Several patients developed sinus
bradycardia and fall in blood pressure (BP) intraoperatively. The
average heart rate prior to CSD was 88±19/min, which reduced to 64±8/min
at the time of discharge after the surgery. The systolic and diastolic
BP fell peri-operatively from 122±12/78±8 mm of Hg to 106±9/62±9 mm of
Hg respectively. In 13 (20%) patients there was severe fall in BP
during the postoperative period, requiring inotropes even after 48 hours
of CSD. These include the 4 patients who died in the index
hospitalization. Bradycardia (HR less than 60) in the post-operative
period persisted in 6 (9.2%) patients. One patients developed left
sided Horner’s syndrome and another developed loss of sweating on left
forehead. Another patient developed persistent transudative pleural
effusion of undetermined etiology. None of our patients developed
hemodynamically significant pneumothorax or hemothorax.
Discussion: In the largest data from India we describe here the
long-term outcome of CSD in patients with refractory VT or VT storm,
mostly with structural heart disease and severe LV dysfunction.. There
has been a very significant fall in the frequency of ICD discharges
after the CSD. Also, freedom from shock at 2 year was 70.8%.
ICD shocks have been shown to have a negative impact on morbidity,
mortality and quality of life and every effort should be made to reduce
shocks without affecting its efficacy (3, 15, 16). In our study, there
was a 92% fall in requirement of shock after CSD. In the latest
published systematic review of CSD by Shah et al, in 173 patients from
13 studies, the efficacy of procedure ranged from 58 to 100% (14).
Vaseghi et al showed a 88% reduction in ICD shocks and freedom from
shock, death or heart transplant of 50% at 1 year. In three previous
reports with a total of 26 patients, 17 (65%) were responders (17-19).
Bourke et al in their study had observed response in 4 (45%) out the 9
patients (<45%) who had undergone unilateral left
sympathectomy (20). Another report by Vaseghi et al. has shown 90%
reduction in shock after CSD and shock free survival of 48% for
bilateral CSD over one year (21).
The overall survival of our cohort was 54/65 (83%) over a follow-up
period of 28±24 months. Previous studies on patients with VT storm have
shown a mortality ranging from 33 to 53 % over a period of 5 years
(22-24). The better survival in our data may be attributed to younger
age of patient, significant proportion with non-ischemic cardiomyopathy
and the lesser duration of follow-up compared to these previous studies.
Afinding which is interesting is that only one patient on follow-up died
because of arrhythmic cause. This suggests that CSD by decreasing the
incidence of arrhythmia and shocks may shift the mortality towards non
arrhythmic causes such as progressive heart failure. This is in
accordance to the finding by Hohnloser et al. who demonstrated mortality
shift from arrhythmic caused to heart failure after prophylactic use of
ICD (25).
The autonomic nervous system extensively affects the normal function of
human heart, and has an important role in initiation and maintenance of
cardiac arrhythmia (26). Studies in the past have documented the role of
sympathetic excitation in ventricular tachycardia (27,28). CSD recently
has been a hot topic of discussion for management of ventricular
arrhythmia storm. The largest evidence in favour of this has been the
study by Vaseghi M, et al. where they studied the effect of CSD in 121
patients with structural heart disease and VT storm (13). This was an
international collaborative study from 5 international centres of which
our centre was one. The 2017 AHA/ACC/HRS Guideline for Management of
Patients With Ventricular Arrhythmias and the Prevention of Sudden
Cardiac Death recommended CSD as one of the therapies for refractory VT
storm if other treatment modalities (medication and catheter ablation)
are either ineffective or non-tolerable (1). However, our cohort differs
considerable from the Vaseghi M, et al. study as only 21% of our
patient had a previous VT ablation procedure as compared to 64% in the
previous study. The reasons for this difference were possibly: i) In our
study, 58.5% of patients had pleomorphic VT, where the results of
ablation are suboptimal; ii) More than 50% of our patients did not have
an ischemic etiology, where again ablation results are inferior; iii)
Our centre has experienced surgical teams adept at performing CSD, even
at times during ongoing VT and iv) Cost constraints did not permit us to
undertake expensive electroanatomic ablation procedures where we did not
expect good results. Our study is also a retrospective analysis which
spans over a period 10 years and as the knowledge and experience
evolved,with time we often subjected our patient upfront for CSD prior
to attempting a VT ablation. Given the significant benefit we have
experienced with CSD, we suggests this therapy can be considered early
in the line of management of refractory VT or VT storm.
In the current contemporary practice of cardiology, CSD is used mostly
for secondary prevention in LQTS and CPVT where it has been a class I
recommendation by theguidelines(1). The mechanism of benefit of CSD is
because of complete disruption of sympathetic control on the heart
(11,29,30 ). Isolated stellate ganglion stimulation(unilateral or
bilateral) may cause increase in dispersion of repolarization (31).
Animal experiments in the past suggested that while dispersion of
repolarization decreases with an associated increase in ventricular
fibrillation threshold with left stellate ganglion resection, right
stellate ganglion resection had either no effect or causes increase in
QT dispersion (32). ,CSD decreases the dispersion of repolarization and
increases the threshold of VT inducibility as shown in porcine heart
model (11). This provided the rationale for unilateral left CSD in LQTS.
Though this approach worked well for LQTS and CPVT, it was unknown if
could it be extrapolated to other VT etiologies, especially with
myocardial scars. The right and left sympathetic nerves subtend to
non-overlapping areas of the myocardium or scar (33). Recent study in
humans have also shown that sympathetic stimulation increases the
dispersion of repolarization in patients with prior myocardial
infarction (MI) (34). Hence we postulate that unilateral left CSD may
not be of benefit if the arrhythmogenic region is supplied predominantly
by the right-sided sympathetic nerves. In fact, canine experiments have
demonstrated that right CSD reduces the incidence of ventricular
arrhythmias, particularly those arising from the right border of
anterior infarct or ischemia. (35). Other animal experiments have shown
that myocardial infarction is associated with bilateral sympathetic
ganglionic and neuronal hypertrophy (36). Furthermore, right-sided
sympathetic nerves may hypertrophy and provide ramifications to the area
subtended by the left-sided sympathetic nerves, partly nullifying the
effects of isolated left CSD. These observations support the role of
bilateral over unilateral CSD for ventricular arrhythmias related to
myocardial scars. Concordant to these observations in one study,
recurrent arrhythmias in 3 cases after a unilateral left CSD were
suppressed with a subsequent right CSD (19). Furthermore, the data of
Vaseghi et al also showed that only left side sympathectomy was
associated with poorer ICD shock-free and transplant free survival (13)
as compared to bilateral CSD. Since patients with channelopathies have
usually a structurally normal heart and thus there may not be
significant neural remodelling as a compensatory phenomenon, making
unilateral left CSD effective. But in patients with structural heart
disease as the neural remodelling is significant, bilateral CSD should
always be preferred.
We observed 4 (6 %) deaths and 6 (9 %) patients requiring prolonged
hemodynamic support after the CSD. In the systematic review by Shah et
al the cumulative complication rate was 28%. Whether to attribute the
four deaths in our study to the procedure is a matter of debate as these
patients were very sick, all having severe LV dysfunction and refractory
VT, and thus the CSD was tried as a last desperate measure. The study of
Vaseghi et al showed NYHA class, unilateral (left sided) sympathectomy
and VT with longer cycle length to be the predictor of recurrence of VT
(13). However, in our study we could not relate any of the studied
parameters to have an effect on recurrence of VT or failure of the CSD.
Optimization of anti-arrhythmic drugs, sedation and general anesthesia
wherever feasible, should be the initial steps in management of patients
with VT storms. It is however easier said as such patients with a poor
ventricular reserve, are often in CHF and cardiogenic shock and do not
tolerate beta-blockers and general anesthesia. Thoracic epidural
anesthesia may be used as a bridge to CSD if the surgery is not feasible
immediately. TEA may also help the clinician to predict the candidates
who are likely to benefit from CSD, although sympathetic blockade may
not be as reliable with TEA as it is with surgical CSD. In our cohort
two patient who did not respond to TEA, did better after CSD. Catheter
ablation of VT including endocardial and epicardial approaches should be
tried in suitable cases as still its recommendations are before CSD. We
also postulate that as there could be existing reserves of catecholamine
in the nerve endings even after CSD, the response to CSD could be
delayed for some time and thus we waited for 2 weeks after CSD for
interpretation of our results. We also inferred from our data that in
some patients the effect of therapy may be delayed upto 3 months. Thus,
a wait for the delayed effect may be required if the intervening period
is not causing significant morbidity to the patient. Probably the CSD
modifiesthe sympathetic discharge in such a way that the overall
arrhythmia burden decreases and it also makes the arrhythmia more
responsive to ATP or other non-shock related therapies of ICD.
In our cohort we had only 14 (21.5%) patients who underwent a EP study
and an attempt of VT ablation prior to being considered for CSD. Out of
these only 6 patients were from our centre and 8 had undergone an VT
ablation at other centre. All these patients either had a failed attempt
or had recurrence of VT prior to being referred to our centre for CSD.
And after our initial experience with CSD, we started preferring CSD
over catheter ablation and thus in 51 (78.5%) of our patients CSD was
preferred over catheter ablation as strategy for VT storm or refractory
VT.Thus, we propose that CSD may be used as one of the definitive
therapies that may help stabilize patient with refractory VT or VT
storm, allowing for a subsequent ablation if necessary. In the long term
CSD may also reduce adverse myocardial remodelling in responders by
reducing the VT burden, thereby stabilizing ventricular function. It
also allows the physician to step down the anti-arrhythmic drugs. There
always remain a question of what can be done further to these patient
who do not respond to CSD and are not suitable for catheter ablation.
Bradfield et al. suggested an incremental role of bilateral renal
denervation in such patients (37). Though the study retrospectively
evaluated only 10 such patients, the idea further strengthens the
predominant role of sympathetic nervous system in maintenance of such
arrhythmia. Indeed, improving the soil seems preferable than just
tackling the seed!
Conclusion: Management of patients with refractory ventricular
tachycardia and VT storms with structural heart disease and severe LV
dysfunctionvery challenging. Catheter based ablation is also not
feasible or effective in a significant proportion of these patients.
Bilateral CSD has an important role to play in managing such patients.
With more evidence gathering in favour of CSD, we suggest it should be
considered as a viable therapeutic option in these patients.