Survival of Continuous Ambulatory Peritoneal Dialysis Patients Increases in the Early Years of Dialysis: Comparison of Parametric Frailty Survival Models.

BACKGROUND

Previous studies on the survival of continuous ambulatory peritoneal dialysis (CAPD) patients have been done using the Cox model. This model requires establishing the basic assumption of proportional hazard (PH) and not establishing it leads to misleading inferences. Alternatively, if the time-to-event distribution is known, accelerated failure time models can be developed in such a way that no PH assumption is required. The purpose of this study is to compare the survival of CAPD patients in two hypertensive and non-hypertensive groups in the condition of not establishing the PH assumption.

MATERIALS AND METHODS

In this retrospective cohort study, 1,400 patients who were referred to dialysis centers from different parts of Iran between 1995 and 2010 were included. The loglogistic, lognormal, and generalized gamma parametric frailty models were compared using Stata 17 software.

RESULTS

The five-year survival rate was 54%. The median survival time for the hypertensive group was lower than for others. Age, body mass index, fast blood sugar, potassium, and creatinine were higher in hypertensive patients. In the early months of dialysis, the risk of death is high and decreases with time. Univariate results revealed that hypertension leads to acceleration of death (HR = 0.79; 95% CI (0.64,0.98)). Multivariate analysis showed that increasing age (HR = 0.97; 95% CI (0.96,0.98)), comorbidity (HR = 0.71; 95% CI (0.55,0.91)) and high blood sugar (HR = 0.997; 95% CI (0.999)) accelerated the risk of death. An increase in serum albumin (HR = 1.27; 95% CI (1.04,1.57)) leads to a delay in death event.

CONCLUSIONS

Parametric models (especially the loglogistic model) are appropriate in evaluating factors affecting the survival of CAPD patients.