Left Cardiac Sympathetic Denervation for Long QT Syndrome: 50 Years’ Experience Provides Guidance for Management

Dusi V, Pugliese L, De Ferrari GM, Odero A, Crotti L, Dagradi F, Castelletti S, Vicentini A, Rordorf R, Li C, Shkolnikova M, Spazzolini C, Schwartz PJ.

JACC Clin Electrophysiol. 2022 Mar;8(3):281-294. doi: 10.1016/j.jacep.2021.09.002. Epub 2021 Oct 27.

PMID: 35331422


Take Home Points:

  1. Left stellate cardiac sympathetic denervation (LCSD) is an effective and safe adjunct to beta-blocker therapy
  2. LSCD in addition to increasing the ventricular fibrillation threshold can shorten the QTc interval providing another mode of risk reduction.
  3. LCSD can reduce the yearly cardiac event rate by 86% in symptomatic patients.
  4. LCSD has durable long-term efficacy during the follow-up of 12.9 ± 10.3 years
  5. A new algorithmic approach for the use of LCSD in the management of LQTS in those not protected by beta-blocker therapy is proposed by the authors.

Comment from Dr. Akash Patel (Cleveland), section editor of Congenital Electrophysiology Journal Watch.

Management of Long QT syndrome (LQTS) has been relatively unchanged over the last several decades. Beta-blockers have been the mainstay in treatment to prevent cardiac events, but breakthroughs still can occur. Pioneering work in 1973 by Dr. Arthur Moss and Dr. Peter Schwartz showed the efficacy of left stellate cardiac sympathetic denervation (LCSD) as an adjunctive therapy. Due to efficacy of beta-blocker therapy, lack of expertise to perform the LCSD procedure, and limited long term data, LCSD has seen limited use. In 2005, Dr. Ackerman and colleagues introduced a minimally invasive thoracoscopic approach which reinvigorated interest in LCSD for the management of LQTS. There continues to be limited long term data and lack of consensus about timing and patient selection for LCSD and ICD. This single center study aimed to review their 50-year experience of LCSD in LQTS.

This was retrospective single center study of all consecutive patients with LQTS who underwent a LCSD from 1973 to 2020. Follow-up was obtained on all but 1 patient. The study analyzed the population in 4 groups: 1) very high risk with either symptoms in the first year of life or high-risk features including CALM, CACNA1C, JLN, and recurrences on beta-blockers (n=18); 2) patients with previous aborted cardiac arrest (n=31); 3) patients with previous syncope and/or ICD shock on beta-blocker or previous syncope and intolerance/refusal to beta-blocker (n = 45); and 4) patients with LCSD in primary prevention such as patients still asymptomatic or with syncope off treatment but deemed to be at high risk based on a combination of factors such as QT interval corrected for frequency [QTc] >500 ms at resting ECG and/or on Holter ECG recordings, T wave alternans, intolerance to beta-blockers, prolonged sinus pauses, or bizarre repolarization recognized as dangerous pattern by clinical experience (n=31). Clinical characteristics and outcomes were analyzed. Cardiac events were classified as at least 1 arrhythmic syncope, aborted cardiac arrest (ACA), or ICD appropriate shocks. Major cardiac event was classified as sudden death (SD), ACA, and ICD shocks.

Baseline characteristics showed a total of 125 patients that consisted of LQT1 (26%), LQT2 (44%), LQT3 (10%), other mutations (9%), and genotype unknown (21%). The mean age at LCSD was 20 ± 15 years. The median age of first symptom was 7 (IQR: 2- 14) years. The mean time from first major cardiac event to LCSD was 8.0± 8.1 years. 82% were symptomatic before LCSD with aborted cardiac arrest in 34%. The first symptom was aborted cardiac arrest in 13% and syncope in 69%. A QTc ≥ 500 msec was seen in 66%. The majority were on treatment, beta-blockers (90%) and mexiletine (10%). In addition, a pacemaker or ICD was present in 17%. Of importance, recurrence of cardiac events was seen in 81% with a median of 5 (IQR: 1 – 13.5) events. See Table 1.


Table Description automatically generated


Surgical LCSD was conducted by 3 surgeons over 43 years. The surgical approach consisted of resection of the lower half of the left stellate ganglion, together with the thoracic ganglia from T2 to T4. A sub clavicular approach (n=94) was used before 2014 and a thorascopic approach (n=31) after 2014.

After LCSD, there was a cardiac event seen in 44% with mean follow-up of 12.9 ± 10.3 years. The cardiac events included sudden death (4%), aborted cardiac arrest (10%), ICD shock only (11%), and syncope (18%). The most common first cardiac event was syncope seen in 84%. Aborted cardiac arrest was the first symptom in 16%. There was no sudden death seen as the first event after LCSD. See Table 2.


Table Description automatically generated


After LCSD, 56% were totally asymptomatic. Syncope was seen in 18%. Major cardiac event in 26% including 4% with sudden death and 11% with aborted cardiac arrest. The median time from LCSD to first cardiac event was 20 months (IQR: 7-52 months), major cardiac event was 129 months (IQR: 24 – 205), and 103 months (IQR: 36 – 160) for sudden death. Based on the subgroups analysis, 97% of the primary prevention group, 49% of the syncope group, 67% of ICD shock group, and 52% of the aborted cardiac arrest group remained asymptomatic. See Figure 1 which excludes the very high risk.



Overall, there were 5 sudden deaths of which 2 were in the very high-risk group and 3 in the previous syncope/ICD shock. 2 had stopped taking beta-blockers. Of note, the sudden death episode was always preceded by at least 2 cardiac events after LCSD (median 3, IQR 2-9). Right cardiac sympathetic denervation was done in 5 patients because of a major cardiac event after LCSD with 3 patients having improvement.

On-treatment analysis of event free survival at 10 and 15 years for cardiac event was 97% for both in the primary prevention group, 49% and 42% in the syncope/ICD shock group, 60% and 45% in the aborted cardiac arrest group, and 15% and 8 % in the very high-risk group (p<0.0001). On-treatment analysis of event free survival at 10 and 15 years for major cardiac event was 100% for both in the primary prevention group, 87% and 74% in the syncope/ICD shock group, 78% and 69% in the aborted cardiac arrest group, and 35% and 9% in the very high-risk group (p<0.0001). See Figure 2.


Line chart Description automatically generated with medium confidence


Survival from cardiac events after LCSD in those without ICD shock in follow-up or ICD at first major cardiac event was 79% at 10 and 15 years in those with previous syncope on beta-blocker or previous aborted cardiac arrest. See Figure 3.


Graphical user interface, application Description automatically generated


Cardiac events in the secondary prevention group reduced by 86% after LCSD from 1.66 (IQR: 1.57 – 1.75) events/year to 0.26 (IQR: 0.23 – 0.28) (p<0.001). Among those with aborted cardiac arrest the major cardiac event rate decreased by 80% after LCSD from 0.63 to 0.13. (p=0.03) Among those with very high risk the cardiac event rate decreased by 75% after LCSD from 2.1 to 0.48. (p=0.03)

LCSD was done as monotherapy due to beta-blocker intolerance/refusal in 12 patients (4 primary prevention, 4 syncope, and 4 prior aborted cardiac arrest). During a follow-up of 18±12 years, there were no major cardiac events and only 2 with syncope.

LCSD did show a significant reduction in the QTc duration after LCSD which persisted at 6 months in most patients. The mean QTc at baseline was 527 ± 60 msec which reduced by 16 ± 24 msec at discharge (p<0.0001) and 26 ± 35 msec at 6 months (P<0.0001) In those patients with QTc>500 msec, 33% had a reduction to <500 msec QTc. See Figure 4.


Chart Description automatically generated


As expected, a QTc >500 before LCSD (HR 3.17) and at discharge from LCSD (HR 2.54) were significant predictors of cardiac events. In those with prior syncope or ICD shock, a QTc at 6 months after LCSD of <500 msec compared to >500msec was associated with 15-year survival from cardiac events of 80 vs 9%, p=0.0001. LCSD efficacy did not shows any difference across genotypes. See Below.


Diagram Description automatically generated


Finally, LCSD was safe with no major complications. There was permanent ptosis seen in 2.4% and transient neuropathic pain in 35% of VATS procedure.

This study provides a unique review of a single center’s 50-year experience using LCSD to manage LQTS. It showed that LCSD was safe and durable. It resulted in an overall reduction in cardiac event rates by 86%. In addition, this study showed LSCD resulted in QTc shortening and conferred additional benefit when the QTc was shortened to <500 msec. Based on their data, the authors developed a practice algorithm (Central illustration below) which provides a logical but not prospectively validated approach to the management of LQTS. The authors favor the use of LCSD, mexiletine, RSCD, and pacing before ICD consideration in primary prevention and syncope on beta-blocker patients. The authors advocate for LCSD with ICD placement in high and very high-risk patients due to the marked reduction in event rates. This study highlights the importance of LCSD and the need to individualize patient care management in LQTS.


Timeline Description automatically generated