Surgical Site Infection as a Risk Factor for Long-Term Instrumentation Failure in Patients with Spinal Deformity: A Retrospective Cohort Study

BACKGROUND: Surgical site infection (SSI) remains a complication of spine deformity surgery. Although fusion/instrumentation failure in the setting of SSI has been reported, few studies have investigated the relationship between these entities. We examine the relationship between early SSI and fusion/instrumentation failure after instrumented fusion in patients with thoracolumbar scoliosis. METHODS: A retrospective review of a prospectively maintained case series for patients undergoing spine surgery between January 1, 2006, and October 3, 2017. Inclusion criteria included age >= 18 years and surgery performed for correction of thoracolumbar scoliosis. Data collected included various demographic, clinical, and operative variables. RESULTS: 532 patients met inclusion criteria, with 20 (4%) experiencing SSI. Diabetes mellitus was the only demographic risk factor for increased SSI (P = 0.026). Number of fused levels, blood volume loss, and operative time were similar between groups. Fusion/instrumentation failure occurred in 68 (13%) patients, 10 of whom (15%) had SSI, whereas of the 464 patients with no fusion/instrumentation failure, only 10 (2%) had SSI (P < 0.001). Of the 20 patients with SSI, 10 (50%) had fusion/instrumentation failure, whereas in the 512 patients with no infection, only 58 (11%) had fusion/instrumentation failure (P < 0.001). Patients with infection also experienced significantly shorter time to fusion/instrumentation failure (P = 0.025), higher need for revision surgery (P < 0.001), and shorter time to revision surgery (P = 0.012). CONCLUSIONS: Early SSI significantly increases the risk of fusion/instrumentation failure in patients with thoracolumbar scoliotic deformity, and it significantly shortens the time to failure. Patients with early SSI have a significantly higher likelihood of requiring revision surgery and after a significantly shorter time interval.