Predictors of positive bone cultures from CT-guided bone biopsies performed for suspected osteomyelitis.

Abstract

Bone biopsies are often used to direct antibiotic choice in patients with suspected osteomyelitis. The aim of this study was to identify the best predictors of positive bone biopsy cultures. A retrospective review of 845 patients who underwent computed tomography (CT)-guided non-spine bone biopsies at a tertiary academic healthcare institution. Thirty-seven patients (4.4%) had biopsies performed for suspected osteomyelitis. Laboratory markers, as well as imaging features, were measured. t-Tests and Fisher's exact tests were used to compare clinical and demographic variables between patients with positive bone cultures and patients with negative bone cultures. Multivariable logistic regression was used to identify the best predictors of bone culture positivity. All patients had negative blood cultures; however, only eight patients (21.6%) had positive bone cultures, with Staphyloccocus the most common organism. Multivariable logistic regression analysis showed that an open wound (OR=14.00, 95% CI (1.74, 112.4), p=0.013) and any fluid aspirated at the time of biopsy (OR=10.50, 95% CI (1.21, 91.01), p=0.033) were the best predictors of bone culture positivity. The area under the curve (AUC) for this multivariable model was 0.784 with sensitivity and specificity of 0.778 and 0.778, respectively. Interestingly, and contrary to popular belief, open wounds with exposed bone did not always yield positive bone cultures, and when cultures were positive, were not polymicrobial. Aspiration of fluid at the time of biopsy and the presence of an open wound are the best predictors of positive bone cultures.