PREFABL: predictors of failure of antibiotic locks for the treatment of catheter-related bacteraemia.

BACKGROUND Antibiotic lock (ABL) solutions can effectively treat catheter-related bacteraemia (CRB) without the need for catheter exchange. This approach does not increase secondary infectious complications. We evaluated the risk factors that contribute to failure when CRB is treated with ABLs and systemic antibiotics in paediatric haemodialysis patients. METHODS A retrospective chart review of 72 children on haemodialysis between January 2004 and June 2006 was performed. We evaluated risk factors for ABL treatment using patients' characteristics, CRB/catheter characteristics and patients' biochemical profiles. The first CRB of each catheter was included in the statistical analysis. Our end points were outcome at 2 weeks of treatment and at 6 weeks following treatment. Compound symmetry covariance structure was employed for statistical analysis. RESULTS We treated 149 CRB in 50 patients. The incidence was 3.4 CRB/1000 catheter days. Thirty CRB failed to be cleared with the use of ABL and systemic antibiotics at 2 weeks of treatment (30/149, 20 vs 80%, P < 0.001). Twenty-four of these catheters required exchange. Thirty-nine of the treated catheters got re-infected within the next 6 weeks (39/125, 31 vs 69%, P < 0.001). CRB aetiology was the only statistically significant independent variable for 2-week outcome (P = 0.033). Coagulase-negative Staphylococcus CRB had higher odds of being cleared at 2 weeks compared with other CRB aetiologies. For the 6-week outcome, the statistically significant independent variables in the final model included age (P = 0.048) and serum phosphorous level (P < 0.001). Younger age and higher serum phosphorous levels were independent risk factors for failure at 6 weeks with re-infection. Area under the receiver operating characteristic (ROC) curve for the model of the 2-week outcome was 0.736 with the percentage of correct predictions at 81.2%. Area under the ROC curve for the model of the 6-week outcome was 0.689 with the percentage of correct predictions at 75.5%. CONCLUSIONS CRB can effectively be treated with ABLs and systemic antibiotics. CRB aetiology is the only independent variable of early treatment failure. Younger age and higher serum phosphorous levels are independent risk factors for re-infection at 6 weeks.