Drinking ethanol is widely believed to predispose to hypoglycaemia in patients with Type 1 diabetes, the suggested mechanism being suppression of hepatic gluconeogenesis. The hypoglycaemic effect of ethanol was investigated by measuring steady-state glucose infusion rate during a hypoinsulinaemic (mean plasma insulin 14 +/- 1.3 (SEM) mU l-1), euglycaemic (blood glucose 5 mmol l-1) clamp. Nine patients with Type 1 diabetes fasted overnight and then had, in single-blind fashion, ethanol 0.5 g kg-1 by intravenous bolus followed by 0.25 g kg-1 h-1 or matched volumes of saline. After 1 h of ethanol or saline, all infusions were stopped and blood glucose monitored for a further 90 min. A 60-min ethanol infusion leading to a steady-state blood concentration of 26.2 +/- 1.4 mmol l-1 (120.7 mg %) did not alter the glucose infusion rate needed to maintain euglycaemia (1.22 +/- 0.12 mg kg-1 min-1 before and 1.23 +/- 0.12 during ethanol infusion), the initial rate of fall of blood glucose (ethanol 0.039 mmol l-1 min-1 vs control (0.033), the lowest blood glucose (4.43 mmol l-1 vs 4.31), or the rate of blood glucose recovery (ethanol 0.050 mmol l-1 min-1 vs control 0.054). We conclude that a moderate amount of ethanol, administered intravenously under controlled conditions, does not lead to hypoglycaemia in patients with Type 1 diabetes who have fasted overnight.