Significant hyperkalaemia occurred in 406 out of 29 063 patients admitted to a major Scottish teaching hospital in one year (1.4%). Mortality was higher in these patients than in control patients and was strongly correlated with the severity of the hyperkalaemia. Overall seven deaths were directly due to hyperkalaemia (out of 58 deaths among patients with hyperkalaemia). Factors contributing to a poor prognosis were severity and speed of onset of hyperkalaemia and the presence of appreciable renal impairment. Patients with hyperkalaemia were older and more likely to be male; this trend was present in all diagnostic subcategories. Genitourinary disease, gastrointestinal disease, and cancer were significantly more common among the patients with hyperkalaemia than the controls. Hyperkalaemia due to drug treatment was invariably mild and non-fatal, whereas genitourinary disease was often associated with moderate to severe hyperkalaemia, which in two cases proved fatal. Use of electrocardiographic monitoring was rare, and although the treatment of hyperkalaemia was effective, it was often used when not required. Hyperkalaemia is a potential hazard in diabetic ketoacidosis, and use of potassium supplements should be carefully monitored during correction of the acidosis.