Peritonitis remains the leading cause of patient dropout from continuous peritoneal dialysis (CPD) therapy. Few studies have compared patient morbidity, mortality, and outcome for patients undergoing CPD who develop gram-positive and gram-negative peritonitis. We retrospectively reviewed the charts of patients who developed either gram-positive or gram-negative peritonitis between January 1, 1993, and December 31, 1995. Three hundred seventy-five patients who developed 415 episodes of gram-positive and gram-negative peritonitis were maintained on CPD therapy during this time period. There was no difference in age, race, and sex between patients who developed gram-positive or gram-negative peritonitis. More patients with diabetes developed gram-negative peritonitis than gram-positive peritonitis (53% v 40%, respectively; P < 0.05). Coagulase-negative staphylococcal species accounted for 47% of all gram-positive episodes, whereas Klebsiella organisms, Escherichia coli, and Enterobacter organisms accounted for 63% of all gram-negative episodes. Significantly more patients who developed gram-positive peritonitis continued CPD therapy 2 weeks and 6 months after the onset of peritonitis than patients who developed gram-negative peritonitis (97% v 73%; P < 0.05 at 2 weeks and 81% v 58% at 6 months; P < 0.05, respectively). Nine percent of the patients who developed gram-positive peritonitis died within 6 months after the onset of peritonitis, whereas 21% of the patients who developed gram-negative peritonitis died (P < 0.05). Patients who developed gram-negative peritonitis were significantly more likely to require hospitalization than patients who developed gram-positive peritonitis (74% v 24%; P < 0.001). More patients with gram-negative peritonitis required peritoneal catheter removal than patients with gram-positive peritonitis (18% v 4%; P < 0.001). Thirty-two percent of the patients who developed gram-positive peritonitis re-developed an episode of peritonitis with the same organism compared with only 9% of the patients who developed gram-negative peritonitis. Furthermore, peritonitis recurrence with the same organism within 6 months after the initial episode was noted in 60% of the patients with peritonitis caused by Staphylococcus aureus compared with 24% of patients with peritonitis caused by other gram-positive organisms (P < 0.05). We conclude that the outcomes of gram-positive and gram-negative peritonitis are different. When rates of peritonitis are used to predict outcome, it appears that gram-positive and gram-negative peritonitis rates need to be examined separately.