Treatment with thiazide diuretics causes impaired glucose tolerance, biochemical diabetes, and insulin resistance. The effect of diuretics on glucose tolerance is clearly dose-related. Spironolactone does not impair glucose tolerance, even at high dosage, but differences among other diuretics could be due to comparisons at doses that are not equal. Diuretic-induced changes in glucose metabolism are not conclusively related to altered potassium homeostasis, and impaired glucose tolerance occurs even when relatively low doses of thiazide are combined with potassium-sparing agents. The effects of diuretics on glucose homeostasis are in large part and probably entirely reversible. These disturbances of glucose metabolism have been detected only by detailed biochemical testing, and their clinical relevance is uncertain. In established diabetes, diuretics have a rapid and substantial adverse effect on metabolic control. In nondiabetic subjects, diuretics rarely cause or trigger a serious hyperosmolar nonketotic diabetic syndrome. Otherwise, it is not known whether the metabolic changes cause clinical diabetes or lead to microvascular complications in the long term. Evidence from large outcome trials suggests that biochemical diabetes, glucose intolerance, and insulin resistance do not increase the risk of coronary heart disease in treated hypertensive patients. Diuretics should be avoided in patients with diabetes unless their use is essential. Otherwise, a low dose of thiazide remains as excellent choice for first-line antihypertensive therapy. Di-hydropyridine calcium antagonists, diltiazem, and verapamil appear to have no important effects on glucose homeostasis. There is very limited evidence that selective a-antagonists increase insulin sensitivity. The importance of metabolic differences between drug classes will be established only by comparative outcome trials with coronary events as the end point. © 1992 Raven Press. Ltd.