Title : Retrospective analysis of ultrasound-guided stellate ganglion blocks for CRPS
Abstract:
The Stellate Ganglion Block (SGB), also known as the cervicothoracic sympathetic block, is a perineural injection of local anesthetic at the C6-C7 level that interrupts sympathetic outflow to the head, neck, and upper extremity. For nearly a century, the SGB has remained an established intervention for many conditions, including Complex Regional Pain Syndrome (CRPS) types I and II. Ultrasound guidance for SGB allows real-time visualization of pertinent surrounding vascular, neural, and visceral structures as well as the direct observation of needle placement and spread of injectate. Despite the expansion of SGB indications and the increasing adoption of ultrasound guidance, contemporary data describing real-world ultrasound-guided SGB practice remain limited. This study aimed to characterize patient demographics, procedural characteristics, patient-reported treatment outcomes, and procedure-related side effects among patients undergoing ultrasound-guided SGB.
This is a single-center retrospective cohort study prepared in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) reporting guidelines. All blocks were performed by the first author (AK) in an office-based suite at his private pain clinic, located in Tampa, Florida and some at an ambulatory surgery center. Patients treated from May 17, 2021, to July 1, 2026, were included in the analysis. All consecutive ultrasound-guided SGBs performed during the study period were identified through the Electronic Medical Record (EMR) and procedure log. For each patient, variables such as demographics (age, gender), block characteristics (right/left/bilateral, anesthetic agent, volume, nitrous oxide use), complications, and follow up communication (successful, unsuccessful, non-responder) were recorded. The primary outcome studied in this cohort was effectiveness. Block success was determined from free-text clinical documentation, as standardized outcome measures were not routinely collected. Safety of the block was studied as a secondary outcome where complications per block were separated into procedure-related adverse events and expected pharmacologic effects.
Within a larger cohort of 125 patients, 22 were identified to have undergone a SGB for CRPS or generalized pain syndromes. Shortly after the procedure, 22.7% patients experienced Horner's Syndrome, which is an expected pharmacological effect. Examination of follow up communication with patients revealed that 18.1% successfully experienced a decrease in their symptoms, 40.9% did not experience a decrease in symptoms, and 41% did not provide a post-procedure follow up.
This variability in response may reflect differences in underlying disease mechanisms. Conditions characterized by sympathetic or amygdala-linked hyperactivity may be better targeted by SGB’s mechanism of action, whereas conditions involving more diffuse or systemic autonomic dysregulation may be less consistently affected by a single, unilateral block. The comparatively lower CRPS response rate observed in this cohort is consistent with prior literature. Systematic reviews have concluded that evidence supporting sympathetic blocks for CRPS remains limited and of low certainty. This may reflect the heterogeneous pathophysiology of CRPS, in which it is not uniformly sympathetically mediated. As a result, not all patients with CRPS would be expected to experience a meaningful response to SGB. Higher response rates reported in smaller studies combining SGB with adjunctive occupational therapy may reflect differences in patient selection, adjunctive rehabilitation, or identification of patients with a greater sympathetic pain component, rather than differences in block technique alone. Future studies should prioritize prospective, multicenter, indication-specific cohorts using validated, condition-appropriate outcome measures collected at standardized follow-up intervals to improve comparability across studies and minimize missing outcome data.
