OBJECTIVE To analyze pathogenetic associations, clinical features, management, and outcome of ascites following discontinuation of continuous peritoneal dialysis (CPD). METHODS Retrospective analysis of symptomatic ascites, defined as ascites requiring at least one therapeutic paracentesis, developing in patients who discontinued CPD. METHODS Dialysis unit of one tertiary care center. METHODS Twelve patients with 13 episodes of symptomatic ascites diagnosed soon after (a few days to 2 months) discontinuation of CPD. METHODS Diagnostic tests to characterize the pathogenesis of ascites; management of ascites by hemodialysis or CPD. METHODS Evolution of clinical features and nutritional parameters, survival. RESULTS Ascites was infectious in 3 episodes (non-tuberculous mycobacterial peritonitis) and noninfectious in the remaining 10 episodes. Serum-to-ascites albumin concentration gradient (AG) was 6.3 +/- 1.5 g/L in infectious ascites and 17.3 +/- 2.7 g/L ( >11 g/L in every episode) in noninfectious ascites. Infectious ascites was managed with hemodialysis, prolonged courses of antimicrobial agents, and repeated paracentesis. Paracentesis ceased after 3-9 months. The patients were alive after 52 +/- 19 months. Seven episodes of noninfectious ascites were managed by hemodialysis and repeated paracentesis. Five patients died within 6 months from cardiac causes or sepsis. The remaining 2 patients died after 14 and 16 months from cardiac causes. Three episodes of noninfectious ascites in 2 patients were treated by restarting CPD within 2-5 months. Patients were alive at 16.9 +/- 13.2 months. They were asymptomatic and achieved fluid control. On the same CPD schedule, peritoneal clearances of urea and creatinine and normalized protein nitrogen appearance were unchanged between the initial and restarted CPD. Serum albumin was 33.3 +/- 2.5 g/L at the end of the first CPD period, 23.6 +/- 2.5 g/L soon after restarting CPD, and 31.3 +/- 5.5 g/L 4 months after restarting CPD. CONCLUSIONS Noninfectious ascites after discontinuation of CPD is often characterized by an AG > 11 g/L, suggesting portal hypertension. Restarting CPD in noninfectious ascites may be associated with improvement in ascites symptomatology and nutritional parameters and with satisfactory survival.