Dexfenfluramine, a serotonin agonist with effects on the central nervous system (CNS), lowers blood glucose in patients with non-insulin-dependent mellitus (NIDDM). Previous studies using the hyperinsulinemic clamp have shown that dexfenfluramine improves insulin action on both stimulation of glucose uptake and inhibition of hepatic glucose production (HGP). Since the central nervous system can influence glucose tolerance in ways that may not be detected using a clamp procedure, we investigated the effects of dexfenfluramine on glucose kinetics during an oral glucose tolerance test (OGTT) in patients with NIDDM. Glucose kinetics were measured basally and during an OGTT using a double isotope technique and the modified one-pool model of the glucose system. After a 4-week run-in period, studies were performed before, after two 15 mg doses, and then after 4 weeks on 15 mg twice daily in 10 subjects with NIDDM. Fasting-plasma glucose was significantly lower after 4 weeks on dexfenfluramine (P < 0.01) as was plasma glucose at both 1 and 2 h during the OGTT (P < 0.05). The lower plasma glucose was associated with a reduction in HGP both basally (P < 0.01) and during the 1st hour of the OGTT (P < 0.05). There was no change in peripheral glucose uptake. Plasma insulin levels were unaltered, but plasma glucagon was lower after 1 month of treatment. We conclude that dexfenfluramine improves fasting-blood glucose and oral glucose tolerance predominantly by reducing hepatic glucose production.