Sex differences in peritoneal dialysis duration and management among Japanese patients: a nationwide registry analysis

Peritoneal dialysis (PD) duration is limited worldwide by PD-related infections and difficulties in fluid and solute management. Sex-associated differences in PD duration among Japanese patients have not been well characterized. We examined sex-associated differences in clinical characteristics, PD duration, and PD-related outcomes using a nationwide Japanese registry. We retrospectively analyzed annual aggregated data of adult Japanese patients on PD extracted from the Japanese Society for Dialysis Therapy Renal Data Registry (2014–2023; cumulative 74,274 patients). Because the extract contained annual aggregated statistics rather than individual-level data, analyses were descriptive and based on within-year cross-sectional comparisons. We compared male and female patients regarding PD duration, fluid management, solute clearance, and PD-related infections, and performed diabetes-stratified analyses, a descriptive temporal trend analysis, and body-composition–corrected sensitivity analyses for fluid output indices. Female patients consistently had longer mean PD duration than male patients in every year from 2014 to 2023; in 2023, the difference was 4.80 months (95% confidence interval, 3.10 to 6.50), and the 10-year sample size-weighted mean was 28.59 months in male versus 34.73 months in female patients. Female patients also had higher registry-reported total Kt/V (weighted mean, 1.90 vs. 1.76) and a higher total output (urine output + PD ultrafiltration volume) / body weight ratio, calculated as a ratio of annual means (24.0 vs. 21.5 mL/kg/day). The sex difference in PD duration persisted in both diabetic and non-diabetic strata. In this nationwide descriptive analysis, female patients on PD in Japan had longer registry-recorded PD duration and higher registry-reported indices of fluid and solute management than male patients, with differences persisting across diabetes strata. However, because the analyses were based on aggregated registry data with substantial, potentially non-random missingness and without patient-level adjustment, causal inference is not possible, and the findings should be interpreted strictly as registry-level descriptive associations rather than patient-level survival or technique-failure differences.