Association of hypo- and hyperkalemia with disease progression and mortality in males with chronic kidney disease: the role of race

Abstract

Background/aims: Abnormal serum potassium is associated with higher mortality in dialysis patients, but its impact on outcomes in predialysis chronic kidney disease (CKD) is less clear. Furthermore, blacks with normal kidney function have lower urinary potassium excretion, but it is unclear if such differences have a bearing on race-associated outcomes in CKD.

Methods: We studied predialysis mortality and slopes of estimated glomerular filtration rate, eGFR) associated with serum potassium in 1,227 males with CKD. Mortality was examined in time-dependent Cox models, and slopes of eGFR in linear mixed effects models with adjustments for case mix and laboratory values.

Results: Both hypo- and hyperkalemia were associated with mortality overall and in 933 white patients, but in 294 blacks hypokalemia was a stronger death predictor. Hypokalemia was associated with loss of kidney function independent of race: a 1 mEq/l lower potassium was associated with an adjusted difference in slopes of eGFR of -0.13 ml/min/1.73 m(2)/year (95% CI: -0.20 to -0.07), p < 0.001.

Conclusion: Hypo- and hyperkalemia are associated with higher mortality in CKD patients. Blacks appear to better tolerate higher potassium than whites. Hypokalemia is associated with faster CKD progression independent of race. Hyperkalemia management may warrant race-specific consideration, and hypokalemia correction may slow CKD progression.

Fig. 1

Multivariable adjusted log hazards (solid line) and 95% CI (dashed lines) of all-cause predialysis mortality associated with serum potassium levels in the entire study population (n = 1,227). Associations were examined in time-varying Cox models adjusted for age, race, smoking status, CCI, DM, cardiovascular disease, BMI, the use of ACEI, ARBs, diuretics, potassium supplements, NSAIDs, beta-blockers, eGFR, serum albumin, bicarbonate, calcium, phosphorus, blood hemoglobin, and 24-hour urine protein.

Fig. 2

Multivariable adjusted log hazards (solid lines) and 95% CI (dashed lines) of all-cause predialysis mortality associated with serum potassium levels in white (n = 933; a) and black (n = 294; b) patients. Associations were examined in separate time-varying Cox models adjusted for age, race, smoking status, CCI, DM, cardiovascular disease, BMI, the use of ACEI, ARBs, diuretics, potassium supplements, NSAIDs, beta-blockers, eGFR, serum albumin, bicarbonate, calcium, phosphorus, blood hemoglobin, and 24-hour urine protein.

Fig. 3

Hazard ratios (95% CI) of all-cause mortality associated with hypokalemia in time-dependent Cox models, overall and in white and black patients. Hypokalemia was defined as a serum potassium <3.8 mEq/l in the overall group and in whites, and as a serum potassium <3.7 mEq/l in blacks. The group with serum potassium 3.8–5.5 mEq/l served as referent in the overall group and in whites, and the group with serum potassium of 3.7–5.5 mEq/l served as referent in blacks. Models represent unadjusted associations (model 1) and associations adjusted sequentially for age, race, smoking status, CCI, DM, cardiovascular disease and BMI (model 2), plus the use of ACEI, ARBs, diuretics, potassium supplements, NSAIDs, beta-blockers (model 3), plus eGFR, serum albumin, bicarbonate, calcium, phosphorus, blood hemoglobin, and 24-hour urine protein (model 4). N values represent the number of patients with hypokalemia in the respective groups.

Fig. 4

Hazard ratios (95% CI) of all-cause mortality associated with hyperkalemia in time-dependent Cox models, overall and in white and black patients. Hyperkalemia was defined as a serum potassium >5.5 mEq/l in all three groups. The group with serum potassium 3.8–5.5 mEq/l served as referent in the overall group and in whites, and the group with serum potassium of 3.7–5.5 mEq/l served as referent in blacks. Models represent unadjusted associations (model 1) and associations adjusted sequentially for age, race, smoking status, CCI, DM, cardiovascular disease and BMI (model 2), plus the use of ACEI, ARBs, diuretics, potassium supplements, NSAIDs, beta-blockers (model 3), plus eGFR, serum albumin, bicarbonate, calcium, phosphorus, blood hemoglobin, and 24-hour urine protein (model 4). N values represent the number of patients with hyperkalemia in the respective groups.