Surgical timing and neurological outcomes in spinal epidural abscess: A systematic review of contemporary evidence (2020–2026)

Background Spinal epidural abscess is a rare but potentially catastrophic neurosurgical emergency, with incidence tripling over two decades. The optimal surgical timing remains actively debated. This systematic review evaluates contemporary evidence (2020–2026) on early versus delayed decompression and neurological outcomes in adults with SEA. Methods Five databases, PubMed/MEDLINE, Embase, Cochrane Library, Scopus, and Web of Science, were searched for studies published January 2020–January 2026. Eligible studies enrolled adults (≥18 years) with confirmed SEA reporting surgical timing data and validated neurological outcomes, such as Frankel or ASIA scale. Two reviewers applied PRISMA 2020 guidelines; risk of bias was assessed with ROBINS-I and evidence certainty with GRADE. PROSPERO: CRD42026132673. Results Ten primary studies were included. Early decompression, within 24–48 h of diagnosis (variously defined as ≤12 h from admission, ≤24 h from diagnosis, or ≤ 48 h from symptom onset across studies) was associated with neurological improvement in 41–91% of patients (73–91% in studies applying a strict ≤24-h threshold), compared with 22–61% following delayed or conservative management. Surgery was associated with lower all-cause mortality (13.6% vs 38.6%; p = 0.03) and was associated with a significantly lower readmission rates in 30-day readmission. Severity of preoperative neurological deficit, advanced age, diabetes mellitus, and MRSA infection were independent predictors of poor outcome. Minimally invasive techniques demonstrated comparable efficacy with reduced morbidity in selected patients. Overall certainty of evidence was low. Conclusions Available observational evidence consistently associates early surgical decompression with improved neurological recovery and lower mortality in neurologically symptomatic SEA; however, all included studies were rated as low certainty of evidence, and causal inference is not supported. Non-operative management may be appropriate in carefully selected neurologically intact patients under close monitoring. Standardized timing definitions and prospective data are needed to confirm these associations.