Comparison of operative and nonoperative management of spinal epidural abscess: a retrospective review of clinical and laboratory predictors of neurological outcome

Abstract

Object Spinal epidural abscess (SEA), once considered a rare occurrence, has showed a rapid increase in incidence over the past 20–30 years. Recent reports have advocated for conservative, nonoperative management of this devastating disorder with appropriate risk stratification. Crucial to a successful management strategy are decisive diagnosis, prompt intervention, and consistent follow-up care. The authors present a review of their institutional experience with operative and nonoperative management of SEA to assess morbidity and mortality and the accuracy of microbiological diagnosis. Methods A retrospective analysis of patient charts, microbiology reports, operative records, and radiology reports was performed on all cases involving patients admitted with the diagnosis of SEA between July 1998 and May 2009. Results Seventy-seven cases were reviewed (median patient age 51.4 years, range 17–78 years). Axial pain was the most common presenting symptom (67.5% of cases). Presenting signs included focal weakness (55.8%), radiculopathy (28.6%), and myelopathy (5.2%). Abscesses were localized to the lumbar, thoracic, and cervical spine, respectively, in 39 (50.6%), 20 (26.0%), and 18 (23.4%) of the patients. Peripheral blood cultures were negative in 32 (45.1%) of 71 patients. Surgical site or interventional biopsy cultures were diagnostic in 52 cases (78.8%), with concordant blood culture results in 36 (60.0%). Methicillin-resistant Staphylococcus aureus (MRSA) was the most frequent isolate in 24 cases (31.2%). The mean time from admission to surgery was 5.5 days (range 0–42 days; within 72 hours in 66.7% of cases). Outcome data were available in 72 cases. At discharge, patient condition had improved or resolved in 57 cases (79.2%), improved minimally in 6 (8.3%), and showed no improvement or worsening in 9 (12.5%). Patient age and premorbid weakness were the only factors found to be significantly associated with outcome (p = 0.04 and 0.012, respectively). Conclusions These results strongly support immediate surgical decompression combined with appropriately tailored antibiotic therapy for the treatment of symptomatic SEA presenting with focal neurological deficit. The nonsuperiority discovered in other patient subsets may be due to allocation biases between surgically treated and nonsurgically treated cohorts. The present data demonstrate the accuracy of peripheral blood culture for the prediction of causative organisms and confirm patient age as a predictor of outcomes.