Study Design: A six-year retrospective analysis of all instrumented spinal fusions performed in the Nuffield Orthopaedic Centre and the John Radcliffe Hospital.

Objective: To assess the incidence of infection following instrumented spinal fusion, the nature of the infecting organisms and their subsequent management.

Subjects: All patients who had undergone removal of spinal metalwork were analysed for evidence of infection. The indications for removal of metalwork included proven deep infection, refractory postoperative pain or planned removal after thoraco-lumbar fracture.

Outcome Measures: Successful treatment of infection was documented when the patient was asymptomatic and inflammatory markers remained within normal limits following cessation of antibiotic therapy. Failure was documented when the patient had recurrent sepsis, refractory pain following removal of metalwork or died.

Results: 80 spinal infections following instrumented fusions were found between 1997 and 2003. 34 of the infecting organisms were propionibacteria, 19 were coagulase negative staphylococcus, 10 were staphylococcus aureus, 8 were methicillin resistant staphylococcus aureus, 3 were coliforms, 2 were proteus, 2 were diphtheroids, 1 was alpha haemolytic streptococcus and 1 was anaerobic streptococcus. 29 of these infections were polymicrobial. Of 55 patients who had metalwork removed secondary to pain, 20 patients had proven infection postoperatively (36.3%). Preoperative inflammatory markers failed to accurately predict the presence of infection for trauma patients. Our management of infection is removal of metalwork with six intraoperative samples sent for culture and histology specimens, followed by administration of at least six weeks of intravenous or oral antibiotic, depending on the organism and its antibiotic sensitivity. Prolonged treatment is used where inflammatory markers remain raised.

Conclusions: Infection of spinal implants presents different management problems to those which follow infected total joint replacement. The lack of specific clinical, laboratory and radiological findings in patients who are subsequently diagnosed as having infections associated with spinal instrumentation presents a challenging clinical problem. We found the most predictive sign of infection following instrumented fusion of scoliotic spines was postoperative pain. CRP and ESR were unreliable as predictors of infection.