Reversible “diabetes” was produced in eight normal male subjects by administering between 400 mg. and 645 mg. of diazoxide and 8 mg. of trichlormethiazide per day for four to seven days. Insulin secretory responses to oral and intravenous glucose loads were determined on the last treatment day and the following morning, respectively. Four to six weeks later both tests were repeated to obtain control data.
The speed of breakdown in glucose tolerance was proportional to daily diazoxide dosage, which in turn was limited by the severity of side effects. Carbohydrate turnover tended to worsen each day, even after correction of severe symptomatic hypokalemia in two subjects. The induced “diabetic” state was associated with sluggish insulin release as blood glucose rose, and with greatly decreased hormonal output per unit of glycemic stimulus—the same two secretory defects which characterize the clinical onset of hereditary diabetes mellitus.
The dala suggest that the potent diabetogenicity of diazoxide given with trichlormethiazide may be almost entirely due io blockade of insulin release, either by direct action or by stimulating alpha adrenergic receptor sites in the beta cell. The synergistic effect of natriuretic analogues like trichlormethiazide probably reflects superimposed potassium deficiency, evidence of which consistently underlies the weaker hyperglycemic property of diuretic benzothiadiazines but is not caused by diazoxide itself. Thus, “diazoxide diabetes” appears to be functionally distinct from ordinary “thiazide diabetes” and, in fact, closely resembles spontaneous diabetes mellitus clinically. Since the condition can be predictably induced and is entirely reversible, it may provide a useful experimental model for further study of the inherited disease in man.