BACKGROUND: Outcomes of a combination of peritoneal dialysis (PD) and once-weekly haemodialysis (PD + HD) have not been extensively studied. METHODS: This prospective cohort study using the Japanese Society for Dialysis Therapy Renal Data Registry included those who transitioned from PD to PD + HD therapy or thrice-weekly HD from 2011 to 2018. Exposure was PD + HD therapy compared with thrice-weekly HD. The outcome was time to all-cause or cause-specific death. Associations between PD + HD therapy and outcomes were examined by Cox regression. Sensitivity analyses were performed by propensity score (PS) matching, PS matching with a shared frailty model in which dialysis facilities were treated as a random effect, inverse probability weighting (IPW), PS adjustment, PS stratification, competing risk regression and on-treatment analyses in which data were censored at the transition to thrice-weekly HD for those on PD + HD therapy. RESULTS: During the study period, 1001 subjects transitioned to PD + HD therapy and 2031 to thrice-weekly HD. During a median follow-up of 3.5 years, 575 subjects died. All-cause, cardiovascular, congestive heart failure-related or infection-related mortality were not significantly different between those on PD + HD and those on thrice-weekly HD [hazard ratio 0.95 (95% confidence interval 0.78-1.16), 1.26 (0.92-1.72), 1.24 (0.77-1.99) and 0.89 (0.57-1.39), respectively]. Sensitivity analyses yielded similar results except that PD + HD therapy was associated with significantly lower all-cause mortality by PS adjustment and PS matching with the shared frailty model and lower infection-related mortality by PS adjustment and IPW. CONCLUSIONS: PD + HD therapy was associated with similar or potentially lower mortality compared with thrice-weekly HD. Considering a flexible lifestyle, PD + HD therapy could be a great option.