Association of early-start automated peritoneal dialysis and remote patient monitoring with clinical outcomes: a 17-year single-center cohort study

Early-start peritoneal dialysis (ESPD), initiation within 14 days of catheter implantation, avoids emergent hemodialysis but raises catheter complication concerns. Remote patient monitoring (RPM) of automated peritoneal dialysis (APD) was introduced at our center in October 2020. We compared outcomes across four groups stratified by early-start timing and RPM status, examining RPM's association with clinical outcomes. This retrospective cohort study included 218 APD patients (2008–2025): ES+RPM (n=54), ES+noRPM (n=81), noES+RPM (n=41), noES+noRPM (n=42). Outcomes included peritonitis, mechanical complications, catheter-related and all-cause hospitalization (per 100 patient-years), and mortality. Logistic and Poisson regression (adjusted for age, sex, follow-up duration), Cox proportional-hazards regression for time-to-event outcomes, and Aalen-Johansen/Kaplan-Meier analyses with log-rank testing were performed. Groups were comparable in age, sex, and BMI at baseline. However, the early start groups (ES) with or without RPM had lower BUN, Cr, and albumin levels. Peritonitis incidence differed significantly (31.5% vs 51.9% vs 29.3% vs 64.3%; p=0.001), driven by lower rates with RPM. Catheter-related hospitalization differed but did not reach statistical significance (3.63 vs. 22.77 vs. 3.49 vs. 25.01/100 PY; p=0.214), but all-cause hospitalization was markedly lower with RPM (14.69 vs. 117.56 vs. 12.93 vs. 135.35/100 PY; p<0.001). After adjustment, ES+RPM had lower catheter-related (IRR=0.221; 95%CI 0.121–0.403; p<0.001) and all-cause (IRR=0.141; 95%CI 0.106–0.187; p<0.001) hospitalization versus noES+noRPM. No differences in mechanical complications were found. Crude mortality was lower during the RPM era, but Cox regression accounting for different follow-up time and adjusting for age, sex, diabetes, and hypertension showed no significant survival difference for any group versus noES+noRPM (HR range 1.51-1.79, all p > 0.10). In this 17-year single-center cohort, ESPD on APD was not associated with significantly higher infectious or mechanical complication rates, consistent with the safety of early initiation in carefully selected patients, though this does not establish equivalence. The RPM era was associated with substantially lower hospitalization and infection rates; however, as RPM and calendar period are confounded, these findings reflect an association and do not establish causation. Prospective controlled studies are needed to isolate the specific contribution of remote monitoring. Early-start peritoneal dialysis did not increase complications. Remote monitoring was associated with lower hospitalization, but not with improved survival once follow-up time was accounted for; peritonitis rates were lower but not significantly after adjustment. RPM and calendar period are confounded, so these associations do not establish causation, and prospective studies are needed before recommending RPM as standard practice.