Comparing the risk for death with peritoneal dialysis and hemodialysis in a national cohort of patients with chronic kidney disease

Bernard G. Jaar, Josef Coresh, Laura C. Plantinga, Nancy E. Fink, Michael J. Klag, Andrew S. Levey, Nathan W. Levin, John H. Sadler, Alan Kliger, Neil R. Powe

Research output: Contribution to journalArticlepeer-review

254 Scopus citations


Background: The influence of type of dialysis on survival of patients with end-stage renal disease (ESRD) is controversial. Objective: To compare risk for death among patients with ESRD who receive peritoneal dialysis or hemodialysis. Design: Prospective cohort study. Setting: 81 dialysis clinics in 19 U.S. states. Patients: 1041 patients starting dialysis (274 patients receiving peritoneal dialysis and 767 patients receiving hemodialysis) at baseline. Measurements: Patients were followed for up to 7 years and censored at transplantation or loss to follow-up. Cox proportional hazards regression stratified by clinic was used to compare the risk for death with peritoneal dialysis versus hemodialysis. Results: Twenty-five percent of patients undergoing peritoneal dialysis and 5% of hemodialysis patients switched type of dialysis. After adjustment, the risk for death did not differ between patients undergoing peritoneal dialysis and those undergoing hemodialysis during the first year (relative hazard, 1.39 [95% CI, 0.64 to 3.06]), but the risk became significantly higher among those undergoing peritoneal dialysis in the second year (relative hazard, 2.34 [CI, 1.19 to 4.59]). After stratification, the survival rate was no different for patients who had the highest propensity of being initially treated with peritoneal dialysis. Results were consistent with adjustment based on a propensity score model and in sensitivity analyses that used as-treated models and models in which switches in type of dialysis were treated as treatment failures. Results were similar but stronger in analyses that were restricted to patients who were treated only in clinics offering both types of dialysis. Limitations: Patients were not randomly assigned to their initial type of dialysis. Also, more patients undergoing peritoneal dialysis than hemodialysis switched type of dialysis over time, and the reason for switching was often a consequence of the technique. Conclusions: The risk for death in patients with ESRD undergoing dialysis depends on dialysis type. Further studies are needed to evaluate a possible survival benefit of a timely change from peritoneal dialysis to hemodialysis.

Original languageEnglish (US)
Pages (from-to)174-183+I-17
JournalAnnals of internal medicine
Issue number3
StatePublished - Aug 2 2005

ASJC Scopus subject areas

  • Internal Medicine


Dive into the research topics of 'Comparing the risk for death with peritoneal dialysis and hemodialysis in a national cohort of patients with chronic kidney disease'. Together they form a unique fingerprint.

Cite this