Hyperkalemia in diabetes mellitus. Effect of a triamterene-hydrochlorothiazide combination

Hyperkalemia is known to occur with increased frequency in the patient with diabetes mellitus and in the elderly when agents that interfere with renal potassium excretion are employed, but the precise frequency has not been established. We employed data from a post-marketing surveillance trial following the introduction of a triamterene-hydrochlorothiazide (Maxzide) combination to estimate the frequency. In patients normokalemic at baseline, hyperkalemia developed with a frequency of 0.59% in 20,809 nondiabetics and in 1.08% of 922 diabetics. Hyperkalemia was threefold to fivefold more likely in those more than 60 years of age, and all of the excess hyperkalemia in diabetics occurred in the elderly. The severity of hyperkalemia was not influenced by the diabetes mellitus. Hypokalemia occurred with a frequency of about 5% and was not influenced by either age or diabetes. In patients who were hypokalemic prior to treatment, hypokalemia was corrected in more than two thirds and hyperkalemia occurred less frequently. Although hyperkalemia indeed occurs with increased frequency in the elderly diabetic when a potassium-sparing combination is employed, the frequency is not so great that such agents should be avoided routinely when their use could be beneficial. Renal function and serum potassium concentration should be assessed prior to instituting treatment and repeated within a few days and a few weeks thereafter in the patient at risk, especially when renal function is suspected, and in the elderly.