RESUMO
OBJECTIVE: To study if the utility of acridine orange (AO) staining method on blood extracted through intravenous device (ID) is a reliable method to diagnose catheter-related bloodstream infection (CRB). DESIGN: Prospective and observational study. PATIENTS: Patients with central ID and clinical data consistent with CRB who gave their consent to participate. Patients having another infection site were excluded. INTERVENTION: At the moment of the clinical suspicion of CRB and before removing the ID, blood samples were extracted from peripheral veins and through the ID to be analyzed by AO staining. After extracting the samples, the catheter was removed and sent for microbiological analysis with Liñares et al and Maki et al techniques. CRS was defined as development of the same microorganism in the tip of the catheter (endoluminal surface with > or = 10 (3) UFC/ml and/or extraluminal surface > or = 15 UFC/ml) and in the peripheral blood. VARIABLES OF INTEREST: Sensitivity, specificity, negative and positive and negative predictive values and positive likelihood ratios (LR) were calculated for the diagnosis of CRB. RESULTS: A total of 121 patients were studied and 4 were diagnosed with CRB: 2 infected with Staphylococcus aureus, 1 with Pseudomonas aeruginosa and 1 with Candida albicans. AO sensitivity was 87.5%, specificity 92.7% and the negative predictive value was 99.5%. Positive likelihood ratio was 12.04 and negative LR 0.13. CONCLUSIONS: Although the number of events does not allow for the estimation of the efficacy of AO to diagnose CRB, its high negative predictive value would make it possible to rule out this infectious complication with some degree of safety.