Peritoneal dialysis and kidney transplant. A two-way ticket in an integrated renal replacement therapy model

Abstract

Background: Peritoneal dialysis (PD) seems to be a good option to initiate renal replacement therapy (RRT), but patients with graft failure choose PD less frequently than incident patients (de novo).

Objective: To describe patient movements between PD and kidney transplantation (TX) and risk factors for failure of the PD technique.

Method: Multicentre observational study of patients starting PD between 2003 and 2009 with follow-up up until January 2010. Survival analysis based on switching from PD to HD as an event using Kaplan-Meier (KM) and forward, stepwise Cox proportional hazards models. Hazard ratio and 95% confidence intervals (HR [CI]) are shown. MAIN VARIABLE: Switch from PD to HD. Two-group comparison: PD post transplant (post-TX) patients (76) compared to pure incident PD (de novo-PD) patients (830).

Patients: 906 PD patients from 19 public hospitals with a mean age of 54.8 years (64.9% male); main ESRD aetiology: glomerulonephritis (25.4%), diabetes (16.7%), vascular-ischaemic (10.7%), interstitial (13.6%) and polycystic (11.2%). Comorbidity conditions: Charlson Index 5.1 (SD 2.4); 21.6% diabetes mellitus (DM), 24.0% cardiovascular (CV) events.

Results: Mean follow-up period on PD: 1.85 years (95% CI [1.68-2.02 years]). KM estimation for switching to HD due to PD failure was 5.46 years [4.42-6.50 years]. At the end of follow-up, 88 patients had died, 154 had been transferred to HD and 306 had received a graft (annual rate for patients on waiting list: 0.49 TX per year on PD). The best Cox multivariate model for switching from PD to HD includes: post-TX (HR: 1.63 [1.01-2.63]), DM (HR: 1.69 [1.19-2.40]) and age (1.01 [1.00-1.02]) per year. Post-TX patients were younger (43.8 years vs 55.3 years) and with less comorbidity conditions than de novo-PD patients (DM 18.4% vs 21.9%; CV 15.8% vs 24.7%). However post-TX patients had worse clinical evolution with a rapid decline of renal function (∆-3.88 vs -1.8 ml/min per year); a higher admission rate (0.9 vs 0.62 per year) but similar peritonitis rate (0.45 vs 0.53 episodes per year). They also needed to be transferred to HD more frequently (28.9% vs 15.8%; P<.006) and needed more time to TX (4.8 years vs 1.7 years, Kaplan-Meier). Consequently, time spent on PD was higher in the post-TX group (2.8 vs 1.8 year).

Limitations: Observational study with absence of a standard protocol to switch PD-HD.

Conclusion: PD seems to be a good first choice technique due to low mortality and high TX ratio in our area. A previous graft failure is associated with a higher rate of PD-failure but time spent on PD is enough to consider this technique as a good option.