General Session: Tumor, Trauma, Infection - Hall F

Presented by: E. Koh


N. Ackshota(1), I. Bussey(1), A. Nash(1), K. Banagan(1), E. Koh(1), S. Ludwig(1), D. Gelb(1)

(1) University of Maryland, Orthopedics, Baltimore, MD, United States


Purpose: The incidence of pyogenic vertebral osteomyelitis (PVO) continues to increase in the United States, highlighting the need to recognize unique challenges presented by these cases and develop effectual methods of surgical management. To date, no prior research has focused on the outcomes of PVO requiring two or more contiguous corpectomies.

Methods: After IRB approval, retrospective review was performed on all adults with PVO refractory to standard non-operative treatment who underwent complete corpectomy of two or more contiguous vertebrae at a single institution between January 2002 and December 2015 (n=56). Patients without at least 1-year of radiographic follow-up were excluded. Diagnosis was made using history, standard lab markers, and imaging with radiographs, magnetic resonance, and computed topography. Anterior cervical, thoracotomy, thoracoabdominal, or lumbosacral retroperitoneal approach was used with anterior debridement and stabilization using an interbody fusion cage (n=54), followed by posterior stabilization. Autogenous tricortical iliac bone graft was used alternatively used in the remaining cases (n=2). Spinal canal decompression was performed when appropriate. Primary outcomes were rate of readmission and reoperation. Readmissiond unrelated to PVO were excluded. Secondary outcomes were 30-day and 1-year mortality, perioperative complications, and hospital length of stay (LOS). Mortality was determined by the chart review and query of the Social Security Death Master File (SSDMF). Data from patients who expired intraoperatively (n=2) was limited to analysis of 30-day and 1-year mortality. Patients who expired before discharge during the index hospitalization (n=8) were excluded from length of follow-up and readmission rate analysis.

Results: 56 patients were reviewed (63% male; mean age 56.8 years; mean radiographic follow-up 2.8 years) with nearly half (45%) diabetic, 25% IV drug users, 41% immune compromised, and 59% with prior surgery at the site of spondylodiscitis. Half of the operations were performed in a staged fashion with no difference in outcomes when compared to single surgery cases. The median LOS was 13 days with nearly half readmitted (43.8%) after a median of 233.5 days after surgery. Complications varied widely by type and occurred in more than half (57.1%) of the cohort. Twelve (22%) posterior revisions were required after a median 54 days for infection, painful or failed hardware, proximal junction kyphosis, adjacent level disease, or extension of the fusion. No revision of anterior instrumentation was required. The 30-day and 1-year mortality rate was 10.7% and 19.6%, respectively, with an infectious etiology as the most common cause of death. Patient age was strongly correlated with 1-year mortality after controlling for comorbidities (p=0.0357) and smoking was independently associated with 30-day mortality (p=0.0184).

Conclusion: Multi-level vertebral corpectomy for the treatment of refractory vertebral osteomyelitis is associated with increased rates of complications and mortality, however when complete extirpation of the involved vertebra is achieved, long-term control of the disease can be expected in more than 80% of patients.