Background Although hypokalemia has been viewed as a significant concern among patients with heart failure (HF), recent advances in HF management tend to increase the risk of hyperkalemia. Objective To characterize contemporary data regarding correlates and prognostic values of dyskalemia in patients with HF. Design, setting, and participants In cross-sectional and longitudinal analyses, we studied 142,087 patients with newly diagnosed HF in US nationwide Veterans Administration database from 2005 through 2013. Exposures Demographic characteristics, laboratory variables, comorbidities, and medication use for the analysis of correlates of dyskalemia as well as potassium level in the analysis of mortality. Main Outcomes and Measures Dyskalemia and mortality. Results Hypokalemia (<3.5 mmol/L) at baseline was observed in 3.0% of the population, whereas hyperkalemia (5.5 mmol/L) was seen in 0.9%. An additional 20.4% and 5.7% had mild hypokalemia (3.5–3.9 mmol/L) and mild hyperkalemia (5.0–5.4 mmol/L). Key correlates were black race, higher blood pressure, and use of potassium-wasting diuretics for hypokalemia, and lower kidney function for hyperkalemia. Baseline potassium levels showed a U-shaped association with mortality, with the lowest risk between 4.0–4.5 mmol/L. With respect to potassium levels over a year after HF diagnosis, persistent (>50% of measurements), intermittent (>1 occurrence but 50%), and transient (1 occurrence) hypo- and hyperkalemia were also related to increased mortality in a graded fashion regardless of the aforementioned thresholds for dyskalemia. These dyskalemic patterns were also related to other clinical actions and demands such as emergency room visit. Conclusions Potassium levels below 4 mmol/L and above 5 mmol/L at and after HF diagnosis were associated with poor prognosis and the clinical actions. HF patients (particularly with risk factors for dyskalemia like black race and kidney dysfunction) may require special attention for both hypo- and hyperkalemia.
ASJC Scopus subject areas