Treatment of essential hypertension with thiazides can normalize blood pressure but also has various electrolyte and metabolic side effects. Control of these untoward effects should maximize the therapeutic benefits derived from these drugs. In his zeal to promote the use of these antihypertensive agents, Dr. Freis [1] conveys a mixed and somewhat confusing message regarding the side effects of thiazides.

Diuretic-induced hypokalemia is perhaps the best studied and most obviously pernicious of these metabolic perturbations. In the first half of his review, Dr. Freis strives to dismiss the importance of this electrolyte abnormality. In the second half of the paper, however, he essentially reverses his position by recommending the use of low-dose thiazides to avoid hypokalemia. He also urges physicians to carefully monitor serum cation levels, presumably to treat any depletion that develops [1]. We reject the first half of the paper and endorse the second half.

Dr. Freis dismisses the loss of 5% of body potassium stores as "biologically insignificant" [1]. This figure translates into a potassium depletion of more than 200 mEq. Loss of 150 to 200 mEq of the cation has been shown to impair renal concentrating capacity [2], and the resulting increase in the transcellular potassium concentration ratio hyperpolarizes skeletal muscle cells; this process leads to paralysis. Hypokalemia increases phase 4 diastolic depolarization, thereby resulting in increased automaticity [3]. Furthermore, by inhomogeneously increasing the rate of repolarization, hypokalemia predisposes patients to such tachyarrhythmias as torsade de pointes [4]. Thus, hypokalemia-induced electrical instability of the heart is hardly a "misconception" [1].

Dr. Freis misquotes the study by Holland and colleagues [5]. The normal baseline 24-hour electrocardiogram became abnormal because of thiazide-induced hypokalemia and returned to normal by the addition of a potassium-sparing diuretic. Thus, three periods were studied, not two.

Given the imperfections inherent in clinical research, we believe that enough evidence exists to indicate that thiazide-induced hypokalemia may cause fatal arrhythmias. Modest salt restriction and use of the lowest effective dose of the diuretic are prudent recommendations for avoiding hypokalemia. We also believe that current evidence justifies normalizing the serum potassium concentration after it is reduced. If thiazides are to be used and hypokalemia corrected, calls for controlled studies with various doses of thiazides and potassium-sparing diuretics seem irrelevant. Thiazides are cheap, effective antihypertensive agents and will be safer and even more effective with sustained normokalemia.