Image-Guided Biopsy in Acute Diskitis-Osteomyelitis: A Systematic Review and Meta-Analysis

BACKGROUND. The reported sensitivity and yield of image-guided biopsies for diskitis-osteomyelitis vary widely. OBJECTIVE. The purpose of this study was to perform a systematic review of the literature and meta-analysis of pooled sensitivity data to elucidate strategies for optimal image-guided biopsies among patients suspected to have diskitis-osteomyelitis. EVIDENCE ACQUISITION. A comprehensive literature search was performed for studies of patient populations with proven or suspected diskitis-osteomyelitis that included percutaneous image-guided biopsy as part of the workup algorithm. Type of pathogens, imaging modality used for biopsy guidance, tissue targeted, antibiotic administration at the time of biopsy, true microbiology positives, true microbiology negatives, false microbiology positives, false microbiology negatives, disease (i.e., diskitis-osteomyelitis) positives as determined by reference standard, true infection positives (i.e., positive microbiology or pathology results), and total number of biopsies performed were extracted from the studies. Microbiology sensitivity, microbiology biopsy yield, and infection sensitivity were calculated from the pooled data. These terms and the data required to calculate them were also defined in detail. EVIDENCE SYNTHESIS. Thirty-six articles satisfied inclusion criteria and were used for analysis. The pooled microbiology sensitivity, infection sensitivity, and microbiology biopsy yields were 46.6%, 70.0%, and 26.7%, respectively. Mycobacterium tuberculosis-only microbiology sensitivity was significantly higher than both pyogenic bacteria and mixed-organism microbiology sensitivity (p < .001). Staphylococcus aureus was the most common causative organism (28.6%). Pooled microbiology sensitivity was not significantly different for CT guidance and fluoroscopy guidance (p = .16). There was a statistically significant difference between pooled microbiology sensitivity of bone/end plate (45.5%) and disk/paravertebral soft-tissue (64.8%) image-guided biopsies (p < .001). There was no statistically significant difference in pooled microbiology sensitivities for patients who received antibiotics before the procedure (46.2%) and those who did not (44.6%) (p = .70). CONCLUSION. Image guidance by CT or fluoroscopy does not affect microbiology yield, disk and paravertebral soft-tissue biopsies should be considered over bone and end plate biopsies, and preprocedural antibiotic administration does not appear to impact biopsy results. CLINICAL IMPACT. Understanding and correctly applying reported statistics contribute to appropriate interpretation of the abundant literature on this topic and optimization of care for patients with diskitis-osteomyelitis.