Risk factors for Gram-negative bacterial infections in febrile neutropenia

Abstract

BACKGROUND AND OBJECTIVES: The objective of this study was to evaluate the risk of Gram-negative bacterial infections in febrile neutropenic patients and to develop a specific risk score. DESIGN AND METHODS: This prospective study included 513 consecutive febrile neutropenic, evaluable patients. Forty-five per cent of the patients were receiving prophylactic gut decontamination, and 6% were receiving prophylactic quinolones at the onset of febrile neutropenia. Data were collected from the onset of febrile neutropenia until 30 days later. Risk factors for Gram-negative bacterial infection were identified by comparing baseline characteristics of patients with and without Gram-negative bacterial infection. Independent risk factors in multivariate analysis were used to build a predictive score for Gram-negative bacterial infection. RESULTS: The prevalence of Gram-negative bacterial infection was 55/513 (10.7%). Gram-negative bacterial infections were due to E. coli in 30 patients, other enterobacteriae in 11, Pseudomonas spp. in 13, and Salmonella spp. in one. In multivariate analysis, the occurrence of Gram-negative bacterial infection was significantly associated with age > 45 years (p=0.009), recent administration of betalactams (p=0.04), chills (p=0.0001), urinary symptoms (p=0.01), and absence of gut decontamination with both colimycin and aminoglycosides (p=0.001). The relative risk for Gram-negative bacterial infection was 4.4, 12.6, 25.4 and 100 in the presence of 1, 2, 3 , or at least 4 parameters, respectively. The performances of our scoring system and the post-test probabilities according to different prevalence rates of Gram-negative bacterial infection (0.05, 0.10, 0.20) lead us to propose a Gram-negative risk score of >or=3 as indicating a high probability of Gram-negative bacterial infection. INTERPRETATION AND CONCLUSIONS: Our scoring system identifies patients with a high probability of Gram-negative bacterial infection as those with a score >or=3. If confirmed in a validation set, this score could be considered in the choice of the first-line antibiotics in febrile neutropenic patients.