RESUMEN
Pyogenic vertebral osteomyelitis (VO) is diagnosed according to several lines of evidence: clinical, biological, radiological, and histological. Definitive diagnosis requires the isolation of a causative pathogen or histological confirmation. The aim of our study was to describe the microorganisms isolated by percutaneous needle biopsy (PNB) and to analyze their susceptibility patterns, in order to assess the possibility of empirical combination therapy for the treatment of nonbacteremic patients without resorting to PNB. Based on a French prospective multicenter study of 351 patients with VO, we compiled clinical, biological, and radiological findings for 101 patients with microbiologically confirmed VO. Based on antibiotic susceptibility testing of PNB isolated pathogens, the suitabilities of four antibiotic combinations were analyzed: ofloxacin plus rifampin, levofloxacin plus rifampin, ciprofloxacin plus clindamycin, and ciprofloxacin plus amoxicillin-clavulanate. The main causative pathogens identified were coagulase-negative Staphylococcus spp. (26% of isolates), followed by Staphylococcus aureus (21%), Streptoccocus spp. (13%), and enterobacteria (21%). Empirical antibiotic combination therapy was effective in nearly 75% of cases, and the different combinations gave similar results, except for ofloxacin-rifampin, which was effective in only 58% of cases. A "perfect" empirical antibiotic therapy does not exist. If PNB is not possible, a combination of a fluoroquinolone with clindamycin or rifampin can be used, but the high risk of microbiological failure does not allow the exclusion of PNB. (This study has been registered with EudraCT, number 2006-000951-18, and ClinicalTrials.gov, number NCT00764114.).