There have been improvements in the technique of peritoneal dialysis (PD) over the last ten years. However, peritoneal infections remain the major complication associated with this treatment, and the risk of infection cannot be accurately predicted. Nevertheless, it is widely accepted that simple connections should be replaced by improved systems of connection, and that patient training is important. Peritoneal infection should be suspected when the dialysate is turbid, whether or not associated with peritoneal irritation. None of the various techniques used for the culture of dialysates has been shown to be either more sensitive or more specific than any of the others. Thus, collaboration between the physicians supervising the dialysis and microbiologists is necessary to choose the culture techniques best adapted. The sensitivity should be at least 85 to 90%. If the sensitivity is lower, the techniques used should be reconsidered. There have been several hundred publications assessing treatments of peritoneal infections associated with PD. However, no particular antibiotic treatment has been demonstrated to be systematically superior. The use of associated antibiotics seems to be preferable initially, until the causative agent has been identified. For example, vancomycin with a third generation cephalosporin seems to be the association of choice, because of its efficacy, tolerance and ease of use. The optimal duration of treatment has not been established by randomised study, but 10 days is commonly used for Gram-positive infections, and longer for Gram-negative. Whatever the treatment used, the success rate should be at least 80 to 90%. Randomised trials with sufficiently large numbers of patients are required to determine the indications and delay before withdrawal of the DP catheter in cases of peritonitis which do not respond to antibiotics.