The effects of carbohydrate supplementation on high-intensity exercise performance were examined in 5 moderately-trained subjects (age = 28.4 +/- 1.5 yr; ht = 171.0 +/- 4.3 cm; wt = 66.25 +/- 6.32 kg). High-intensity exercise tests (initiated at the power output (PO) associated with 90% VO2 peak [mean = 201 +/- 21 watts] x 60 min, with drop-off in PO allowed over time) were completed under the following randomized double blind conditions: 1) pre-exercise glucose polymer (G)/placebo during exercise (G/P), 2) G pre-exercise and during exercise (G/G), and 3) placebo pre-exercise and during exercise (P/P). Subjects ingested 300 ml of a sweetened placebo or a similarly flavored 10% G solution, immediately prior to and every 15 min during exercise. No differences were observed in PO among the 3 treatments until min 40-60 where PO was greater with G. This resulted in significantly greater total work (and less drop-off in PO) with G (G/P = 619 +/- 234kJ [14.5% lower than the value associated with 201 watts maintained for 60 min (724kJ)], G/G = 599 +/- 235 kJ [17.3% lower than the value associated with 201 watts maintained for 60 min]) compared with placebo (P/P = 560 +/- 198 kJ [22.7% drop-off in average PO]) (p < 0.05). VO2 followed a similar pattern with no difference in VO2 over min 0-40 and significantly higher VO2 in G/P and a trend for higher VO2 in G/G during min 40-60 compared to placebo. Results of the present study indicate that, compared to placebo, pre-exercise ingestion of G (30 g in 10% solution) results in less drop-off in PO during 1 hour of high-intensity exercise performance, and that no further benefit is observed when the same amount of G is also ingested every 15 min during exercise.