307. Evaluation of the Diagnostic Approach to Native Spondylodiskitis

BACKGROUND: Current approaches for diagnosis of native spondylodiskitis are variable, as is the yield of image guided spinal biopsy. METHODS: This is a retrospective cohort study including adults determined to have native spondylodiskitis at our institution from January 2007 through July 2017. Inclusion criteria were imaging suggestive of spondylodiskitis, with either positive blood culture and/or a spinal biopsy culture or histopathology. Those with historical diagnosis or surgical site infections were excluded. Histopathology was the gold standard test for sensitivity/specificity calculation. Univariate logistic regression was used to predict positive biopsy culture. RESULTS: A total of 221 patients met our inclusion criteria, of which 203 (91.8%) had blood cultures done (112/203 positive, 55.2%), and 173 (78.2%) had spinal biopsy done of which 113 (65.0%) had received antibiotics within the preceding 2 weeks and 63 (36.4%) had positive culture. Forty-three bone specimens were cultured, and six (13.9%) were positive, while 136 disk specimens were cultured, and 58 (42.6%) were positive. There were 84 (48.5%) biopsies with histopathology performed on either bone or disk specimens, of which 47 (55.9%) were diagnostic. The sensitivity of bone culture was 27.3%, with a specificity of 91.7%. The sensitivity of disk culture was 52.6%, with a specificity of 75.0%. A single biopsy episode sensitivity was 48.9%, and specificity was 80.8%. A total of 23 (13.4%) patients had repeat biopsies (10 bone, 14 disk), five of which had positive cultures (21.7%). On univariate logistic regression, only a positive blood culture was predictive of a positive biopsy culture (odds ratio (OR) 13.08, 95% confidence interval (CI) 1.97–86.81, P = 0.007). Disk culture had a higher yield than bone culture (OR 2.29; CI 0.91–5.73, P = 0.077) and prior antibiotics decreased the yield (OR 0.17; 95% CI 0.02–1.21, P = 0.078). CONCLUSION: The combination of histopathology and cultures including both bone and disk from spinal biopsies improve the diagnostic yield of native spondylodiskitis. Some patients require repeat biopsy. DISCLOSURES: All authors: No reported disclosures.