RESUMEN
OBJECTIVES: To describe the outcome and risk factors for treatment failure of 76 Gram-negative bacilli (GNB) prosthetic joint infections (PJIs) managed with a curative intent according to a standardized protocol derived from published guidelines. METHODS: We analysed data from all the cases of GNB-PJI treated surgically over an 8 year period. Treatment failure was defined as persistence or recurrence of PJI signs during follow-up, resulting in additional surgery and/or antibiotic administration or death. RESULTS: Treatment failure within the follow-up period (medianâ=â2.6 years) was observed in 16 of 76 (21.1%) patients. The failure rate was similar whether the patients were treated with fluoroquinolones in the whole cohort (22.4% versus 16.7%, Pâ=â0.75) and after stratification according to the surgical procedure. The low failure rate observed in patients not receiving fluoroquinolones might be explained by the standardized attitude of maintaining intravenous ß-lactams throughout treatment duration (medianâ=â90 days). In multivariate analysis, C-reactive protein level ≥175 mg/L was significantly associated with treatment failure (adjusted HRâ=â7.75, 95% CIâ=â2.66-22.59, Pâ<â0.0001). CONCLUSIONS: Management according to standardized procedures may improve the prognosis of GNB-PJI. Intravenous ß-lactams, continued for 3 months, should be considered an effective alternative to fluoroquinolones.