The aim of the present study was to compare the metabolic fate of repeated doses of fructose or glucose ingested every 30 min during long-duration moderate-intensity exercise in men. Healthy volunteers exercised for 3 h on a treadmill at 45% of their maximal oxygen consumption rate. "Naturally labeled" [13C]glucose or [13C]fructose was given orally at 25-g doses every 30 min (total feeding: 150 g; n = 6 in each group). Substrate utilization was evaluated by indirect calorimetry, and exogenous sugar oxidation was measured by isotope ratio mass spectrometry on expired CO2. Results were corrected for baseline drift in 13C/12C ratio in expired air due to exercise alone. Fructose conversion to plasma glucose was measured combining gas chromatography and isotope ratio mass spectrometry. Most of the ingested glucose was oxidized: 81 +/- 4 vs. 57 +/- 2 g/3 h for fructose (2P < 0.005). Exogenous glucose covered 20.8 +/- 1.4% of the total energy need (+/- 6.7 MJ) compared with 14.0 +/- 0.6% for fructose (2P < 0.005). The contribution of total carbohydrates was significantly higher and that of lipids significantly lower with glucose than with fructose. The blood glucose response was similar in both protocols. From 90 to 180 min, 55-60% of circulating glucose was derived from ingested fructose. In conclusion, when ingested repeatedly during moderate-intensity prolonged exercise, fructose is metabolically less available than glucose, despite a high rate of conversion to circulating glucose.