Severe Neurological and Systemic Adverse Events Following Cervical Epidural Steroid Injection: A Structured Review of Reported Clinical Cases, Anatomical Considerations, Pathophysiology, and Clinical Implications

Objective: The objective of this study was to map the clinical spectrum, timing, mechanisms, management, and outcomes of reported severe neurological and systemic complications following cervical epidural steroid injections (CESIs), and to correlate the reported events with transforaminal (TF) and interlaminar (IL) technical approaches.Methods: This study was designed as a case-based systematic review and descriptive synthesis. Thirty-two published reported cases were included. Extracted variables included study name and year, injection technique, drug or steroid category, complication type, onset timing, injection level, and management or outcome. Cases were classified according to injection technique, complication category, onset timing, steroid category, and outcome severity. The synthesis was conducted in accordance with PRISMA principles and the Synthesis Without Meta-analysis (SWiM) approach.Results: The analysis showed that the most frequent complication category was non-ischemic neurological, spinal cord, or nerve injury (34.4%). This was followed by vascular or ischemic neurological injury (18.8%), air-related complications (15.6%), hemorrhagic complications (12.5%), autonomic, cranial nerve, or systemic reactions (12.5%), and infectious or systemic complications (6.2%). Immediate or minutes-after-procedure onset was frequently documented. Persistent sequelae were reported in 21.9% of cases, death in 6.2%, and complete recovery in 25.0%.Conclusion: These findings suggest that reported TF complications are commonly associated with vascular or ischemic mechanisms, whereas reported IL complications are more often related to neuraxial access, including hematoma, pneumocephalus, and direct spinal cord or nerve injury. These results should be interpreted as patterns of reported harm rather than incidence estimates.