The aim of this study was to establish the experimental paradigm of fasting, followed by refeeding, to investigate individual differences in nutrient partitioning. Eight nonobese men were fed a normal meal (25% of daily energy requirements) on two occasions, after an overnight (13-h) fast and after a prolonged (72-h) fast. During the entire fasting period, subjects were resident in a whole room indirect calorimeter, and blood samples were drawn periodically. Because no other food was consumed over the 12 h after either meal, negative energy balance was observed after the overnight and prolonged fast. Postprandial carbohydrate oxidation was significantly reduced after the 72- vs. 13-h fast (P < 0.0001), whereas fat oxidation was significantly increased (P < 0.0001). Interestingly, carbohydrate balance was positive after the prolonged fast but negative after the overnight fast (24 ± 17 vs. -57 ± 16; g/12 h, respectively; P < 0.001), whereas fat balance was negative under both conditions (-78 ± 7 vs. -47 ± 8 g/12 h, respectively; P < 0.002). With 72 h of fasting, the glucose and insulin excursions in response to the mixed meal were significantly greater compared with the 13-h fast (P < 0.001.). In conclusion, prolonged fasting resulted in a significant decrease in carbohydrate oxidation and an increase in fat oxidation, after a normal mixed meal, in healthy men. This was associated with a significant decrease in glucose tolerance. Because circulating free fatty acids were greatly elevated at all times after the prolonged fast, these may be mediating some of the changes in postprandial metabolism.