Background. Failed kidney transplant is becoming a frequent cause of dialysis initiation. Although studies have shown no difference between peritoneal dialysis (PD) and haemodialysis (HD) in terms of patients and technique survival, PD remains quite rarely used in this condition. Studies in larger multicentre matched cohorts are missing. Methods. We conducted a retrospective study about 328 patients registered in the French Language Peritoneal Dialysis Registry (RDPLF) who started PD after kidney transplant failure (Tx group) between January 2002 and December 2012 who were compared with 656 matched never-transplanted patients having started PD during the same period (control group). Patients and PD technique survival as well as peritonitis episodes were analysed. Results. Over the study period, patients' survival was similar between the two groups (P ¼ 0.34). The mean time on PD was significantly shorter for patients in the Tx group [17 months (range 14-20)] compared with the control group [21 months (range 19-23)] (P ¼ 0.003). The main cause of transfer to HD was for both group adequacy and/or ultrafiltration failure. Peritonitis rates were similar in the two groups: 43.6% (n ¼ 143) versus 40.1% (n ¼ 263) in the Tx and control group, respectively (P ¼ 0.3). In multivariate Cox analysis, kidney transplant failure (P < 0.0001), younger age (P ¼ 0.02) and male gender (P ¼ 0.01) were associated with a higher risk of transfer to HD. Using multivariate competing risk analysis, kidney transplant failure was again observed as a predictive factor (P < 0.0001), but not age and gender. The only other significant predictive factor observed was peritonitis episodes experienced during PD treatment (P ¼ 0.002). Conclusions. Comparing the Tx and control groups, we report similar patient survival and peritonitis rates but a higher PD technique failure in the Tx group.