Background. The standard therapy of dialysis catheter-related bacteraemia involves both systemic antibiotics and catheter replacement. We reported recently that instillation of an antibiotic lock (highly concentrated antibiotic solution) into the catheter lumen after dialysis sessions, in conjunction with systemic antibiotics, can successfully treat many episodes of catheter-related bacteraemia without requiring catheter removal. The present study evaluated whether the likelihood of achieving a cure with this protocol depends on the type of pathogen. Methods. This was a historically controlled interventional study of an antibiotic lock protocol for the treatment of catheter-related bacteraemia. We analysed prospectively the likelihood of clinical cure (fever resolution and negative surveillance cultures) with an antibiotic lock protocol among patients with dialysis catheter-related bacteraemia. In addition, infection-free catheter survival was evaluated for up to 150 days, and compared with that observed among patients managed with routine catheter replacement. Results. Overall, the antibiotic lock protocol was successful in 33 of 47 infected patients (70%) with catheter-related bacteraemia. The likelihood of a clinical cure was 87% for Gram-negative infections, 75% for Staphylococcus epidermidis infections, and only 40% for Staphylococcus aureus infections (P = 0.04). The median infection-free catheter survival with the antibiotic lock protocol was longer than that observed among patients with routine catheter replacement (154 vs 71 days, P = 0.02). Conclusions. The clinical success of an antibiotic lock protocol in eradicating catheter-related bacteraemia while salvaging the catheter is highly dependent on the bacterial pathogen. Thus, the overall success rate in an individual dialysis programme will depend on the relative frequencies of different bacterial pathogens. © ERA-EDTA 2004; all rights reserved.