BACKGROUND: Few studies have evaluated the viability and outcomes between peritoneal dialysis (PD) and haemodialysis (HD) in urgent-start renal replacement therapy (RRT). This study aimed to compare infectious and mechanical complications related to urgent-start PD and HD. Secondary outcomes were to identify risk factors for complications and mortality related to urgent-start dialysis. METHODS: A quasi-experimental study with incident patients receiving PD and HD in a Brazilian university hospital, between July 2014 and December 2017. Subjects included individuals with final-stage chronic kidney disease who required immediate RRT, that is, HD through central venous catheter or PD in which the catheter was implanted by a nephrologist and utilized for 72 h, without previous training. Patients with PD were subjected, initially, to high-volume PD for metabolic compensation. After hospital discharge, they remained in intermittent PD in the dialysis unit until training was completed. Mechanical and infectious complications were compared, as well as the recovery of renal function and survival. RESULTS: In total, 93 patients were included in PD and 91 in HD. PD and HD groups were similar regarding age (58 ± 17 vs. 60 ± 15 years; p = 0.49), frequency of diabetes mellitus (37.6% vs. 50.5%; p = 0.10), other comorbidities (74.1% vs. 71.4%; p = 0.67) and biochemical parameters at the beginning of RRT, that is, creatinine (9.1 ± 4.1 vs. 8.0 ± 2.8; p = 0.09), serum albumin (3.1 ± 0.6 vs. 3.3 ± 0.6; p = 0.06) and haemoglobin (9.5 ± 1.8 vs. 9.8 ± 2.0; p = 0.44). After a minimum follow-up period of 180 days and a maximum follow-up period of 2 years, there was no difference regarding mechanical complications (24.7% vs. 37.4%; p = 0.06) or bacteraemia (15.0% vs. 24.0%; p = 0.11); however, there was a difference regarding infection of the exit site (25.8% vs. 39.5%; p = 0.04) and diuresis maintenance [700 (0-1500) vs. 0 (0-500); p < 0.001], with better results in the PD group. There was better phosphorus control at 180 days in the PD group (62.4% vs. 41.8%; p = 0.008), with a lower requirement for phosphate binder usage (28% vs. 55%; p < 0.001), erythropoietin (18.3% vs. 49.5%; p < 0.001) and anti-hypertensives (11.8% vs. 30.8%; p = 0.003). Time to death was similar between groups. In the multivariate analysis, PD was a predictor of renal function recovery [odds ratio: 3.95 (1.01-15.4)]. CONCLUSION: PD is a viable and safe alternative to HD in a scenario of urgent-start RRT with complication rates and outcomes similar to those of HD, highlighting the results regarding renal function recovery.