Data on the outcomes of patients who are transferred between hemodialysis (HD) and peritoneal dialysis (PD) are scant. This study analyzes data on patients transferred from their initial modality at the University Hospital Gent between 1978 and 1996 for the reasons for, and outcomes from, the transfer. The reasons for the transfer from HD to PD (n = 35) were access problems (25%), cardiovascular (40%) and blood pressure problems (25%), and personal choice (10%). For the transfer from PD to HD (n = 32) peritonitis (50%), social problems (14%), adequacy or ultrafiltration problems (25%), and leakage (11%) were the reasons. There were no differences in survivals between patients who started on HD and who were transferred to PD, and those who remained on HD (P = 0.4). There were improved survivals for the patients who started on PD and who were transferred to HD compared to those remaining on PD (P = 0.01). It is concluded that transferring problematic HD patients to PD does not improve their survivals, whereas the survival of patients with PD-related complications is improved by transferring them to HD. This is due to the differences in the reasons for transfer, which are PD-related complications in the case of PD patients, and cardiovascular problems in the case of HD patients.