Dexfenfluramine has been shown to lower blood glucose concentrations independently of its effects in reducing food intake and body weight, in human and animal syndromes of non-insulin dependent diabetes. This study aimed to determine whether dexfenfluramine could also reduce glycaemia in rats with severe insulin-deficient diabetes induced by the beta-cell toxin, streptozotocin (55 mg kg-1). Three weeks after diabetes induction, nine groups (each n = 10) of diabetic and non-diabetic rats underwent oral glucose tolerance tests (1 g kg-1, by gavage). These tests were preceded by 12-18 h of fasting to remove the confounding effects of hyperphagia in diabetic rats, and to stabilize glycaemia. Dexfenfluramine (1.0 mg kg-1), given 2 h before the glucose challenge, significantly reduced basal glycaemia and decreased the post-challenge glycaemic rise (P < 0.01 vs. untreated diabetics). Dexfenfluramine dosages of 2.5 and 5.0 mg kg-1 both further flattened the post-challenge glycaemic profiles (both P < 0.01 vs. untreated diabetics) and achieved levels that did not differ significantly from those in non-diabetics (both P > 0.05). Subsequently, the studies using dexfenfluramine dosages of 2.5 and 5.0 mg kg-1 were repeated to determine whether the drug affected plasma insulin levels 2 h after dosing. In diabetic rats, plasma insulin concentrations were reduced to 10-20% of non-diabetic values, and were not significantly altered by dexfenfluramine. Acute dexfenfluramine administration therefore improves and (at dosages of 2.5 and 5.0 mg kg-1) essentially normalizes glucose tolerance in rats with severe insulin-deficient diabetes.(ABSTRACT TRUNCATED AT 250 WORDS)