Glucose and fructose were studied in eight healthy volunteers who fasted twice for 4 days. Before and after the fasts each subject received a 4-hr glucose or fructose infusion providing 0.5 g/kg/hr. Glucose infusion during starvation resulted in a mean maximal plasma glucose rise of 401 +/- 21 mg/100 ml (+/- SEM) as compared to 119 +/- 10 mg/100 ml before starvation. Insulin/glucose ratios were lower than normal in fasted subjects. Fructose infusion during fasting produced a maximal plasma glucose rise of 91 +/- 9 mg/100 ml as opposed to 5+/-1 mg/100 ml before starvation. During fructose infusion in the fasted state, plasma fructose levels were higher than control and the rise in blood lactate and pyruvate was delayed, but finally lactate concentrations were above control values. The antiketotic effects of intravenous glucose and fructose were similar during fasting but fructose was significantly less potent in reducing free fatty acid levels. After starvation, urinary carbohydrate losses during glucose infusion were 5 times higher than those observed during fructose infusion. Thus, fructose utillization was less impaired during fasting than was glucose utilization, although fasting induced abnormalities in both glucose and fructose metabolism.