Mortality studies comparing peritoneal dialysis and hemodialysis: what do they tell us?

Several recent large-scale epidemiological studies comparing mortality among end-stage renal disease (ESRD) patients receiving hemodialysis (HD) versus peritoneal dialysis (PD) show conflicting results. In this paper, we undertake a critical review of these studies. Our goal is to determine if there are any consistent trends in outcomes between HD and PD within select subgroups of patients once methodological differences have been accounted for. A total of six large-scale registry studies and three prospective cohort studies conducted in the United States (US), Canada, Denmark, and the Netherlands were reviewed. Summary findings from these studies are presented for comparative purposes. Additional summary analyses based on previously reported data on 398 940 incident US Medicare patients are included for the purpose of comparing results from this population of patients to those of the other select studies when similar methods of analysis are applied. Results are summarized in terms of the relative risk of death for PD versus HD (RR[PD:HD]). Differences in results between the nine studies can be attributed to the degree of case-mix adjustment carried out and to the use of different subgroups when comparing mortality between HD and PD. When these differences are accounted for, we found a remarkable degree of synergism in results between the registry studies and, to a lesser degree, the prospective cohort studies. PD was generally found to be associated with equal or better survival among non-diabetic patients and younger diabetic patients in all four countries. However, among older diabetic patients, results varied by country. The Canadian and Danish registries showed no difference in survival between PD and HD among older diabetics while in the US, HD was associated with better survival for diabetics aged 45 and older. All studies show a time-dependent trend in the RR of death with PD generally associated with equivalent or better survival during the first year or two of dialysis. However, results on longer-term survival varied according to study and to different subgroups within studies. Subgroup analyses in the prospective cohort studies were limited by small numbers of patients resulting in highly varied and somewhat controversial results when compared to the larger registry-based studies. Based on our review of recent publications and additional analyses of US Medicare data, we conclude that overall patient survival is similar for PD and HD but that important differences do exist within select subgroups of patients, particularly those subgroups defined by age and the presence or absence of diabetes.