Comparison of risk of peritoneal dialysis-associated peritonitis between roxadustat and recombinant human erythropoietin in peritoneal dialysis patients: a retrospective comparative cohort study

BACKGROUND: Roxadustat and recombinant human erythropoietin (rhuEPO) have been approved for the treatment of renal anemia in patients undergoing dialysis. The comparison of risk of peritoneal dialysis (PD)-associated peritonitis between roxadustat and rhuEPO in PD patients remains uncertain. We aimed to compare the risk of PD-associated peritonitis between roxadustat and rhuEPO and examine possible modifiers for the comparison in PD patients. METHODS: A total of 437 PD patients with renal anemia (defined as hemoglobin ≤10.0 g/dL) from 4 centers were selected. Participants were scheduled for follow-up every 1–3 months at each center. We compared differences in baseline characteristics by medication group and 1:1 matching group based on propensity scores. PD-associated peritonitis was defined according to the International Society for Peritoneal Dialysis guidelines. Univariable and multivariable Cox proportional hazard analyses were performed to compare the risk of PD-associated peritonitis between roxadustat and rhuEPO in PD patients. Propensity score matching method was used to examine the robustness of results. RESULTS: A total of 437 participants, including 291 in roxadustat group and 146 in rhuEPO group, were included in the current study, respectively. During a median follow-up of 13.0 (25th–75th, 10.0–15.0) months, PD-associated peritonitis occurred in 68 patients, including 26 of 291 (0.10 episodes per patient-year) patients in the roxadustat group and 42 of 146 (0.27 episodes per patient-year) patients in the rhuEPO group. Overall, compared to patients in the rhuEPO group, the roxadustat group (hazard ratio, 0.345; 95% confidence interval: 0.202–0.589) was associated with a lower risk of PD-associated peritonitis with adjustment of use of roxadustat medication, age, sex, hypertension status, diabetes status, dialysis vintage, serum potassium, hemoglobin, and albumin. Furthermore, the results were consistent with the propensity score analysis. None of the variables, including age, sex, body mass index, PD vintage, presence of residual renal function, hemoglobin, albumin, serum potassium, and C-reactive protein levels, significantly modified the associations. CONCLUSIONS: Our study demonstrated that compared with rhuEPO, roxadustat may reduce the risk of PD-associated peritonitis in PD patients, highlighting the importance of roxadustat for the prevention of PD-associated peritonitis in PD patients.