Long-Term Modality-Related Mortality Analysis in Incident Dialysis Patients

Abstract

Background The effects of the various dialysis modalities on patient survival are different, especially for diabetic patients. Hemodialysis (HD) and peritoneal dialysis (PD) are the predominant renal replacement modalities. This study analyzes modality-related mortality in long-term dialysis patients. Methods This prospective cohort study was conducted between May 1991 and October 2005. Incident patients that had initiated dialysis and had been on dialysis for more than 3 months were enrolled. All cause, infection related, and cardiovascular disease-related mortalities were used as end points. Patient survival was analyzed by the Cox proportional hazards model after adjusting for age, sex, diabetes, comorbidity, and time-averaged values of laboratory data to control influential covariates. Results In total, 1347 patients (258 on PD and 1089 on HD) were enrolled. Adjusted all cause, infection related, and cardiovascular disease-related mortality did not differ significantly between HD and PD patients. In diabetic patients, adjusted all-cause [HD vs PD: hazard ratio (HR) 0.717, 95% confidence interval (CI) 0.400 – 1.282] and infection-related mortality (HD vs PD: HR 1.341, 95% CI 0.453 – 3.969) did not differ significantly between patients on HD and patients on PD. However, adjusted cardiovascular disease-related mortality increased significantly in diabetic PD patients (HD vs PD: HR 0.375, 95% CI 0.154 – 0.913). For nondiabetic patients, adjusted all cause, infection related, and cardiovascular disease-related mortality did not differ significantly between HD and PD patients. Conclusions Dialysis modality had no significant impact on all-cause or infection-related mortality. More studies are needed to clarify the putative difference in cardiovascular mortality risk between diabetic patients on PD and diabetic patients on HD.