Haemodialysis (HD) and peritoneal dialysis (PD) should be regarded as complementary methods of renal replacement therapy. Approximately 10-20% of patients on PD are transferred annually to HD due to technique failure. Much smaller proportion of patients changes modality from HD to PD, predominantly due to vascular access problems, cardiac disease or patient preference. The effects of these transfers on therapy outcome, patient and technique survival have not been studied, with research focusing on outcome measures within the single modality and comparisons between the two methods. We have analysed retrospectively a cohort of 264 patients treated with PD in a single PD centre during 1994-2006. Patient characteristics, therapy measures and outcome of patients were compared between patients for whom PD was the initial method of renal replacement therapy (group 1, n = 197) and those transferred to PD from haemodialysis because of complications (group 2, n = 67). The Kaplan-Meier method and Cox proportional hazards multiple regression analysis were used to assess patient and technique survival. In patients transferred from HD, significantly less had diabetes (11.9% versus 38.1%, P < 0.0001) and there were also significantly more females (57% versus 42.2%, P < 0.05). Baseline Kt/V was significantly higher in the primary PD therapy group (2.46 +/- 0.57 versus 2.11 +/- 0.48, P < 0.001), due to lower residual renal function in patients transferred from HD. Group 2 had also significantly higher peritonitis rate (0.86 versus 0.62 episode/year, P < 0.05). During the time of observation, 71 patients have died, in 100 patients kidney transplantation was performed, 56 were transferred to HD, renal function recovered in 5 and 32 were still on PD at the end of the study. No significant differences were observed in unadjusted patient survival, but technique survival was significantly lower in group 2 (P < 0.05). In the Cox multiple regression model, diabetes status, age and albumin level significantly influenced survival. Relative risk of death was not increased significantly in patients transferred from HD. Our data suggest that outcome of patients transferred from HD is similar to that achieved in patients in whom PD is the first choice therapy. Thus, this option should be strongly considered in patients experiencing complications on HD, mainly vascular access problems, heart failure or intradialytic hypotension.