Renal transplantation is the optimal treatment strategy for patients with end-stage renal disease (ESRD); few are afforded the opportunity due to limited organ supply. Of the alternatives, peritoneal dialysis (PD) and hemodialysis (HD), it is unclear which confers the greater survival advantage, as prior comparisons have demonstrated conflicting results due to lack of case-mix adjustment, limited follow-up, and failure to consider switches in modality over time.
We compared all-cause and cause-specific mortality between PD and HD in national cohort of 263,556 new ESRD patients in the U.S. who began treatment between 5/1995 and12/2000, and followed until 12/2001. A propensity analysis, predicting the probability of assignment to PD, was used to control for baseline differences through regression adjustment and matching based on 23 demographic and comorbid indicators. The C-statistic for this model was 0.75, indicating excellent discrimination between treatments. Time-dependent Cox regression, stratified by age and diabetes, compared PD and HD using an intent-to-treat and as-treated approach and patients were censored at transplantation, loss to follow-up or end of study.
There were 122,672 deaths (46.5%), 24,596 renal transplants (9.3%) and 17,432 (6.6%) patients lost to follow-up within the 6-yr period. The adjusted relative PD/HD hazards ratios [RR] with 95% Confidence Intervals for all-cause and cause-specific mortality are shown (intent-to-treat analysis).
Mortality risks were significantly greater for PD compared with HD among diabetic patients and were principally confined to older patients. The excess mortality could be accounted for, in decreasing order, by increased death risk from infection, cardiac, stroke and the other causes of death category.
In conclusion, haemodialysis should be preferentially considered over PD among older (>50 yrs) diabetics with ESRD in order to improve patient survival.