124. Risk factors, biomarkers and microorganism data for surgical site infections after posterior spinal fusion: a modern series of 3,965 patients

Abstract

<h3>BACKGROUND CONTEXT</h3> Surgical site infection (SSI) after posterior spinal fusion (PSF) is an uncommon but morbid complication. Microorganism and biomarker data are often excluded from large-scale studies on SSI after PSF. <h3>PURPOSE</h3> Identify independent risk factors for SSI after PSF and characterize the role of biomarker data on modern management. <h3>STUDY DESIGN/SETTING</h3> Retrospective chart review at an urban, academic tertiary referral center. <h3>PATIENT SAMPLE</h3> We identified all patients who underwent PSF at our institution between 2000 and 2015, for a total of 3,965 consecutive patients. <h3>OUTCOME MEASURES</h3> Surgical site infection, 30-day readmission, reoperation, antimicrobial culture results. <h3>METHODS</h3> We identified all patients who underwent PSF at our institution between 2000 and 2015, and collected demographic, clinical and biomarker data. We performed multivariable regression, time-to-event analysis, and receiver operating characteristic analyses to identify factors associated with SSI and SSI management decisions. <h3>RESULTS</h3> Among 3,965 patients, 191 patients (4.8%) developed SSI. A significantly greater proportion of infections were gram positive (61.8%, p <0.001). Time-to-infection was shorter for gram-negative infections. Comorbid disease burden, staged procedures, and more levels fused are consistent predictors of adverse outcomes, including readmission, reoperation, infection and prolonged antibiotic therapy. Leukocytosis was absent in 41% of SSI patients. A C-reactive protein (CRP) threshold of 3.9 mg/L was 68.2% sensitive and 64.7% specific for identifying patients who underwent wound washout (AUROC=0.72, p <0.001). An erythrocyte sedimentation rate (ESR) threshold of 42 mm/hour was 74.4% sensitive and 64.7% specific (AUROC=0.72, p=0.001). For predicting prolonged antibiotic use, a threshold CRP peak of 15.7 mg/L was 66.7% sensitive and 74.7% specific (AUROC=0.69, p=0.011). An ESR threshold of 70 mm/hour was 86.7% sensitive and 65.4% specific (AUROC=0.70, p=0.002). <h3>CONCLUSIONS</h3> Comorbid disease burden, staged procedures and more levels fused are risk factors for adverse outcomes, including readmission, reoperation, infection and prolonged antibiotic therapy. The timing and risk factors for infection vary by bacterial peptidoglycan class. Leukocytosis can be mild with SSI, and inflammatory biomarkers are only moderately accurate in predicting clinical decisions. <h3>FDA DEVICE/DRUG STATUS</h3> This abstract does not discuss or include any applicable devices or drugs.