The relationship between diuretic therapy and possible increased risk of coronary heart disease (CHD), especially sudden death, is controversial. The initial report from the Multiple Risk Factor Intervention Trial (MRFIT) raised the possibility that the increased CHD mortality observed in a subset of special intervention men with hypertension and certain electrocardiographic abnormalities on their baseline examination might be an unexpected adverse effect of diuretic therapy. Subsequent reports from the MRFIT have revealed a stronger association of CHD mortality to hydrochlorothiazide than to chlorthalidone. There was no consistent relationship of CHD mortality to the dose of either drug, to the most recent serum potassium level, or to the presence of ventricular premature beats. Unfavorable trends of the same magnitude were also seen among similar white men in the Hypertension Detection and Follow-up Program and in the Oslo hypertension trial, although the sample sizes in these two studies were too small to yield clearcut conclusions. Clinical studies have shown an increased risk of CHD death among hypertensive men with left ventricular hypertrophy. Such men are also noted to have a higher frequency of ventricular premature beats, even in the absence of diuretic therapy. Other studies have shown that diuretic-induced hypokalemia is accentuated in the presence of epinephrine and that low potassium levels decrease the threshold for ventricular fibrillation. Thus, although the evidence is still incomplete, it is possible that the excess CHD mortality among MRFIT special intervention men with electrocardiographic abnormalities may have been caused by a combination of increased left ventricular mass in the presence of coronary atherosclerosis, and hypokalemia caused by good compliance with diuretic therapy and accentuated by stress-induced increases in circulating catecholamines. Given the very large population of patients receiving diuretic therapy, further evaluation of this possibility is important.