283 Incidence of and Patient Characteristics Associated With Potential Delays in Diagnosis of Spinal Epidural Abscess

      Study Objective

      Spinal epidural abscess (SEA) is a rare condition with increasing incidence over the past decade, that requires prompt diagnosis and treatment to avoid permanent and devastating neurologic disability. The diagnostic gold standard, magnetic resonance imaging (MRI), may be difficult to obtain emergently. Though many risk factors have been described in case series, diagnostic delays are common and a source of malpractice claims. We sought to describe the incidence of and patient characteristics associated with potential delays in diagnosis of spinal epidural abscess (SEA).


      We conducted a retrospective study of patients diagnosed from January 1, 2016 to December 31, 2019 with cervical, thoracic and/or lumbar spine SEA based on ICD-10 diagnosis codes who had MRI within 3 days of diagnosis. Patients were members of Kaiser Permanente Northern California, an integrated health care system at the time of diagnosis and for nine of the twelve prior months, for data completeness. We performed electronic data extraction for patient characteristics (demographics, co-morbidities) and potentially related ambulatory and ED visits in the 30 days prior to the index visit. We performed manual chart review for history and exam characteristics, prior 90 days’ antibiotic use, spine instrumentation and indwelling vascular catheter presence. Potential delay in diagnosis was defined as an ambulatory care or ED visit with history or exam documentation of fever, chills, back or neck pain, radicular pain, numbness, weakness or cauda equina characteristics (post-void residual > 200 ml, new urinary or bowel incontinence, decreased rectal tone). We then conducted bivariate analysis to compare these characteristics in the groups of patients with and without potential delay in diagnosis and examined which factors were associated with a delay in diagnosis.


      After applying exclusion criteria and manually confirming SEA diagnosis, 457 patients with SEA were included in the analysis (41 without health plan membership and 180 without new SEA were excluded). Median patient age was 63 years (interquartile range 45-81 years) and 61% of patients with SEA were male. A majority of patients (323 of 457, 71%) experienced a potential delay in diagnosis. Diabetes mellitus (116 vs 42 patients or 36% vs 31%, p=0.35) and recent infection (144 vs 48 patients or 45% vs 36%, p=0.14) were similar between patients with and without potential delay in diagnosis. Injection drug use (13 vs 9 patients of 4% vs 7%, p=0.21) and chronic steroid use cohort (29 vs 3 patients, 9% vs 2.2%, p=0.01) were infrequent in the overall cohort.


      In this retrospective review of patients diagnosed with SEA in an integrated health care delivery system, we found that 323 patients (71%) had at least one visit for potentially related symptoms or findings in the 30 days prior to SEA diagnosis. Previously described risk factors of diabetes mellitus and recent infection were common but injection drug use and chronic steroid use were uncommon.