Association between hyperkalemia and outcomes in hospitalized patients with acute kidney injury

To evaluate the association of hyperkalemia with outcomes of hospitalized patients with acute kidney injury (AKI) and explore the potential non-linear relationship between potassium and the outcomes in a publicly accessible dataset.

AKI patients with measured serum potassium were included. Hyperkalemia was defined as serum potassium level higher than 5.5 mmol/L. The primary outcome was a composite of AKI progression, dialysis, or death within 14 days of randomization. Other outcomes examined included components of the primary outcome, AKI duration, and a series of process outcomes. The association between hyperkalemia and examined outcomes was investigated using odds ratio (OR) and 95 % confidence intervals (CIs) in logistic regressions. Restricted cubic spline regression and inflection point analyses was used to assess the dose-response relationship between potassium and the outcomes.

A total of 5969 patients were included. Of these, 1258 (21.1 %) developed the primary outcome. Hyperkalemia was independently associated with the outcomes of interests. After full adjustment, hyperkalemia was independently associated with the composite outcome (OR 2.41, 95 % CI 1.83---3.18), AKI progression (OR 1.76, 95 % CI 1.29---2.38), death (OR 2.98, 95 % CI 2.12---4.21), dialysis (OR 2.02, 95 % CI 1.24---3.29), kidney consultation (OR 2.06, 95 % CI 1.53---2.78), and discharge to home (OR 0.48, 95 % CI 0.36---0.64). A U-shaped association between serum potassium and the risk of each outcome of interest was observed, with the nadir of risk at the range of 4.1 to 4.2 mmol/L in the inflection points analysis. The findings generally remained stable in the different subgroups.

Hyperkalemia was significantly associated with AKI progression, death, dialysis, and a number of process outcomes in hospitalized patients with AKI. The nadir of risk for the examined outcomes was at potassium that is much lower than 5.0 mmol/L.