SummaryBackground and objectives Unlike hemodialysis (HD), peritoneal dialysis (PD) is a continuous therapy and does not induce myocardial stunning. Yet, the death risk in HD and PD patients is similar. This study tested the hypothesis that serum potassium abnormalities contribute more to the death risk in PD patients than in HD patients.Design, setting, participants, & measurements Data from patients treated in DaVita facilities between July 1, 2001 and June 30, 2006 (n=10,468 PD patients; n=111,651 HD patients) were used to determine association of serum potassium with mortality.Results PD patients were significantly more likely to have serum potassium ,4 mEq/L, with an adjusted odds ratio of 3.30 (95% confidence interval [95% CI], 3.05, 3.56). There was a U-shaped relationship between timeaveraged serum potassium and all-cause and cardiovascular mortality of PD patients, with adjusted hazards ratios of 1.51 for all-cause mortality for potassium ,3.5 mEq/L (95% CI, 1.29, 1.76) and 1.52 for potassium $5.5 mEq/L (95% CI, 1.32, 1.75). The population-attributable risks for all-cause mortality for serum potassium ,4.0 and $5.5 mEq/L were 3.6% and 1.9%, respectively, in PD patients, and 0.8% and 1.5%, respectively, in HD patients.Conclusions Abnormalities in serum potassium contribute disproportionately to the high death risk in PD patients. This may, in part, account for the equivalent cardiac risk seen with the two therapies.