Most of the evidence presented by Drs. Shivkumar and Narins concerns potassium depletion produced by dietary restriction or by low potassium perfusates of tissues rather than depletion produced by administration of diuretics. Results obtained by such methods may not be relevant to hypokalemia produced by thiazide diuretics. For example, I am not aware of any definite evidence that either short- or long-term administration of diuretics leads to damaged renal function or structure or that either leads to muscle weakness or paralysis.

The reference to the possibility that diuretic-induced hypokalemia may cause a reduction in phase-4 depolarization is inappropriate. They refer to a study in which action potentials were recorded from Purkinje fibers of the pig while the area was perfused with Krebs solution containing less than 1 mmol/L of potassium. This process resulted in reduction of phase 4 depolarization. This experiment is not relevant because the potassium level was far lower than would occur during administration of thiazides.

Holland and colleagues [1] did find a reduction in ventricular ectopic activity when they added spironolactone; this addition increased serum potassium levels. The treatment was given, however, to only the 7 of 21 patients who showed the greatest increase in arrhythmic activity. In view of the great day-to-day spontaneous variation in ectopic activity, these patients were most likely to show improvement because of regression toward the mean. The study by Holland and colleagues was flawed because of biased case selection. However, seven other better controlled studies that used 24- to 48-hour electrocardiographic monitoring found no significant changes in ventricular arrhythmias after administration of large doses of thiazides [2]. One investigator also found that normalization of severe hypokalemia did not affect arrhythmic activity [3].

The main purpose of my review was to present the evidence that the toxic effects of thiazide diuretics, even in large doses, have been greatly exaggerated. However, I agree that in the light of present findings, small doses of diuretics should be given preference over large doses in the treatment of hypertension.