All of the patients in our study met the clinical case definition of catheter-related bloodstream infection suggested by the Centers for Disease Control and Prevention: isolation of the same microorganism from blood cultures obtained by venipuncture and drawn from the catheter in a patient with signs of sepsis and no source of bloodstream infection other than the catheter [1]. The Centers for Disease Control and Prevention guidelines further state that defervescence after removal of the catheter is indirect evidence of catheter-related infection in the absence of laboratory confirmation. Laboratory confirmation (which can be done by quantitative culture of blood or culture of a catheter segment) would have provided an even more stringent experimental definition of catheter-related bacteremia.

Our study was designed to assess the outcome of attempted catheter salvage in patients undergoing hemodialysis who have bacteremia, using systemic antibiotics alone; this practice is currently standard at our institution. We are aware of Capdevila and colleagues' data suggesting that for bacteremia caused by certain organisms, catheter salvage may be more successful when antibiotics are "locked" into the catheter in addition to being administered systemically [2-4]. However, this study was limited by its small size and lack of a control group. Moreover, of the 11 patients treated with the antibiotic-lock method, 2 (18%) had recurrent bacteremia due to the same organism (and yet were considered to have had a successful outcome); in 3 patients, the catheter was still in place at the end of the study period, with no follow-up cultures documenting that the bacteremia had cleared. Finally, only 2 of the patients had bacteremia due to Staphylococcus aureus, the most common etiologic organism in our study and the one most likely to be associated with complications. The antibiotic-lock technique is appealing in concept, but we believe that available data are not sufficiently conclusive to justify its adoption as standard practice. The work of Capdevila and others [3, 4] should generate further study of this technique, preferably in a randomized study that compares the efficacy and safety of the antibiotic-lock technique with those of early catheter removal.