174 - Predictors of Failure of Nonoperative Management of Spinal Epidural Ab...

#174 Predictors of Failure of Nonoperative Management of Spinal Epidural Abscesses

General Session: Best Papers Session 2

Presented by: M. Harris


M.B. Harris (1)
S.D. Kim (2)
C.M. Bono (1)

(1) Brigham and Women's Hospital, Harvard Medical School, Orthopaedic Surgery, Boston, MA, USA
(2) Washington University Medical School, Orthopedic Surgery, St Louis, MO, USA


Background: Spinal epidural abscess (SEA) has been traditionally treated with urgent surgical decompression followed by intravenous antibiotics. Cases of successful medical management of select patients have recently challenged that notion. However, independent variables that determine success or failure of medical management of SEA patients have not been previously published.

Methods: A retrospective, case-control study analyzing the management of spinal epidural abscesses was performed. All patients admitted to the authors' healthcare system with a diagnosis of SEA from 1993 to 2011 were identified. Patients eighteen years or older diagnosed with spinal epidural abscess documented by MRI or CT myelogram were included in this study. Patients with postsurgical spinal epidural abscesses or spondylodicitis/osteomyelitis with phlegmon were excluded. Patients with complete spinal cord injury from SEA who presented forty-eight hours after the onset of paralysis or with less than 2 months follow-up were also excluded. For the comparison study, looking at outcomes of medical treatment, both univariate and multivariate analysis were used to identify the significant variables that determine success or failure of medical management of SEA.

Results: 355 patients (214 males to 141 females) with average age of 60 years (range, 18 to 94) met our inclusion criteria. 142 patients were initially intended for nonoperative management but 42 patients failed medical treatment, (one week of antibiotics) and crossed over to the surgical decompression group. 100 patients were treated with long-term intravenous (IV) antibiotics as their definitive treatment and 255 patients were ultimately treated with surgical decompression followed by IV antibiotics. Overall mortality for the entire cohort of SEA was 8.7% during the initial hospitalization and 11.3% within 90 days of admission. The pathogen was staphylococcus in 78% of the cases. Comparing the outcomes of medically managed patients, seventy-three patients were successfully treated with antibiotics as their definitive treatment and 54 patients were considered to have failed initial medical treatment: 12 patients died as a complication of worsening infection and 42 patients crossed over to the surgical treatment group after developing worsening pain, changing neurologic exam, or sepsis despite more than 1 week of intravenous antibiotics. Univariate analysis identified age, neurologic status at the time of presentation, diabetes mellitus, epidural abscess above the level of the conus medullaris and circumferential epidural abscesses as significant risk factors for failure of medical treatment. Multivariate analysis identified incomplete or complete neurologic deficits as the most significant risk factor for unsuccessful medical management, along with age greater than 65 years, diabetes, and methicillin resistant staphylococcus aureus (MRSA) as independent risk factors for failure of medical management. Based on this data, an algorithm for probability of failed antibiotic management of spinal epidural abscess predicted 99% probability of failure in patients older than 65 years with diabetes with MRSA infection presenting with a neurologic impairment.

Conclusion: Spinal epidural abscesses treated with medical management alone have a high risk for failure. In particular, age greater than 65 years, diabetes mellitus, MRSA infection and neurologic compromise were independent risk factors for failure. In the absence of these risk factors, medical management may be considered with probability of failure of 12%.