Background: Hyperkalemia is common and lethal electrolyte disorder with little known long-term consequences. This was retrospective, observational study of hospitalized patients with initial serum K > 5.3 mEq/L. 143 consecutive episodes of hyperkalemia were analyzed in 133 patients. Survival was analyzed by parameters of renal dysfunction (admit eGFR), CHF, admit K and EKG abnormalities.

Methods: Hazard ratios (HR) for mortality were computed by Cox proportional hazards multivariate regression. Primary end point, all-cause mortality determined by Social Security Death Index and medical record review.

Results: Admit eGFR was the most powerful predictor of mortality. The effect of renal function was nonlinear(figure 1). Highest mortality is eGFR group of 15-59 HR 6.92. More severe renal impairment with eGFR(HD) HR 3.67. ESRD had lower mortality HR 1.33(table 1). Hyperkalemic severity had a modest effect(figure 2). Compared to patients Admit K 5.3-5.9 mEq/L, patients with K 6-7, HR 2.21 (p=0.0210) and K >7.0, HR 2.62 (p=0.0521). History of CHF, increased mortality by univariate analysis (p

Conclusions: Survival in hyperkalemic patients is predicted by lower admit eGFR in a non-linear fashion. ESRD patients exhibited lower mortality perhaps reflecting adaptation to chronic hyperkalemia. CHF has an additive effect on mortality in non HD patients. We emphasize that 86% of the mortality was after discharge. This extraordinary mortality necessitates the need to develop risk stratification strategies in the long-term care of the hyperkalemic patients.