Vertebral Osteomyelitis after Spine Surgery - A Disease with Distinct Characteristics.

STUDY DESIGN Prospective cohort study. OBJECTIVE We aimed to determine the 2-year survival and to identify clinical and microbiological characteristics of patients with native vertebral osteomyelitis (VO) as compared to postoperative VO to find further strategies for improvement of the management of VO. SUMMARY OF BACKGROUND DATA A relevant subgroup (20%-30%) of patients with VO has a history of spine surgery. Infection in these patients might be clinically different from native VO. However, clinical, microbiological, and outcome characteristics of this disease entity have not been well studied as most trials either excluded these patients or are limited by a small cohort and short observation period. METHODS Between 2008 and 2013, patients who presented at a tertiary care center with symptoms and imaging findings suggestive of VO were reviewed by specialists in infectious diseases, clinical microbiology, and orthopedics to confirm the diagnosis and followed prospectively for a period of 2 years. Statistical analysis for group comparisons, survival analysis, and uni- and multivariable Cox regression models were performed. RESULTS Thirty percent of the patients with VO (56/189) reported a history of spine surgery in the same segment. Patients with postoperative infection had a lower ASA score (American Society of Anesthesiologists) (P = 0.01) and were less likely to suffer from comorbidities compared to native cases (P = 0.003). Infections caused by coagulase-negative staphylococci (33.3 vs. 6.5%, P < 0.001) and other bacteria of the skin flora (15.2 vs. 0%, P = 0.002) were more prevalent in postoperative patients. Suffering from native VO increased the 2-year mortality risk 3-fold, also when adjusted for the remaining risk factors ASA score and number of comorbidities (hazard ratio 2.916 [95% confidence interval 1.215 -6.999], P = 0.017). CONCLUSION Beside clear microbiological differences, the significant better 2-year survival supports the concept of postoperative VO presenting a distinct disease entity. The subtle disease presentation of patients with postoperative VO should not attenuate clinical suspicion of physicians. LEVEL OF EVIDENCE 3.