Data concerning outcomes of children on hemodialysis (HD) and peritoneal dialysis (PD) are scarce and frequently derived from single-center experiences. We sought to compare survival and transplantation rates in a large cohort of PD and HD patients. We extracted all patients initiating dialysis under 16 years of age between 2004 and 2013 from the Italian Registry of Pediatric Chronic Dialysis. Patients on PD were propensity-matched to those on HD based on gender, age, primary cause of ESRD, and the number of co-morbidities. Stratified Cox proportional hazard models were used to compare outcomes by dialysis modality. Three hundred ten patients were matched from 452 incident patients. In the unmatched cohort, PD patients were younger, more likely to be diagnosed with CAKUT, and had a higher urine output than HD patients. In the propensity-matched cohort, covariates were balanced between the two groups. At 2 years, the cumulative hazard ratio for death was similar (CHR 0.95, 95% CI 0.17-5.20) for HD relative to PD patients; and at 5 years, the CHR was lower for HD patients (0.22 95% CI 0.16-0.29). The cumulative incidence of transplantation at 3 years after dialysis initiation was 60.9% in HD patients and 59.7% in PD patients, with a CHR of 1.03 (95% CI 0.73-1.45).
CONCLUSIONS: Pediatric PD and HD patients have distinct characteristics. After controlling for treatment-selection biases, children selected to start on PD or HD exhibit a similar mortality risk during the first 2 years on treatment, after which this risk increases in PD children. What is Known: • Few studies have compared hard outcomes in children on maintenance dialysis. • Children started on different dialysis modalities have distinct characteristics that impact on survival. What is New: • After controlling for treatment-selection biases, children selected to start dialysis on PD or HD exhibit a similar mortality risk during the first 2 years on treatment, after which this risk appears to be increased in PD children. • An "integrative care" approach should be used in children on PD, switching them to HD when PD-related morbidity tends to increase.