Long-term functional outcome in pyogenic spinal infection

Study Design. Retrospective. Objective. To assess long-term adverse functional outcome following pyogenic spinal infection using standardized outcome measures, Oswestry disability index (ODI), and medical outcomes study short form-36 (SF-36). Summary of Background Data. There is minimal published data regarding the long-term functional outcome in pyogenic spinal infection. Previous studies have used heterogeneous, unreliable, and nonvalidated measure instruments yielding data that is difficult to interpret. Methods. All cases of pyogenic spinal infection presenting to a single institution managed operatively and nonoperatively from 1994 to 2004 were retrospectively identified. Follow-up was by clinical review and standardized questionnaires. Inclusion in each case was on the basis of consistent clinical, imaging, and microbiology criteria. Results. Twenty-nine cases of pyogenic spinal infection were identified. Twenty-eight percent were managed operatively and 72% with antibiotic therapy alone. Nineteen patients (66%) had an adverse outcome at a median follow-up of 61 months, despite only 5 patients (17%)having persistent neurologic deficit. A significant difference in SF-36 physical function scores was observed between patients with adverse outcome and patients who recovered (P = 0.003). SF-36 scores of all patients, regardless of management or outcome, failed to reach those of a normative population. A strong correlation was observed between ODI and SF-36 physical function scores (rho = 0.61, P < 0.05). Seventeen percent (n = 5) of admissions resulted in acute sepsis-related death. Delay in diagnosis of spinal infection (P = 0.025) and neurologic impairment at diagnosis (P < 0.001) were significant predictors of neurologic deficit at follow-up. Previous spinal surgery was associated with adverse outcome in patients requiring readmission within 1 year of hospital discharge following first spinal infection (P = 0.018). No independent predictors of adverse outcome, persistent neurologic impairment, readmission within 1 year, or acute death were identified by logistical regression analysis. Conclusion. High rates of adverse outcome detected using SF-36 and ODI suggest under-reporting of poor outcome when American Spinal Injury Association score alone is used to qualify outcome.