This study examined the effects of sustained high-intensity interval training (HIT) on the athletic performances and fuel utilisation of eight male endurance-trained cyclists. Before HIT, each subject undertook three baseline peak power output Wpeak tests and two simulated 40-km time-trial cycling performance (TT40) tests, of which the variabilities were 1.5 (1.3)% and 1.0 (0.5)%, respectively [mean (SD)]. Over 6 weeks, the cyclists then replaced 15 (2)% of their 300 (66) km.week-1 endurance training with 12 HIT sessions, each consisting of six to nine 5-min rides at 80% of Wpeak, separated by a l-min recovery. HIT increased Wpeak from 404 (40) to 424 (53) W (P < 0.01) and improved TT40 speeds from 42.0 (3.6) to 43.0 (4.2) km.h-1 (P < 0.05). Faster TT40 performances were due to increases in both the absolute work rates from 291 (43) to 327 (51) W (P < 0.05) and the relative work rates from 72.6 (5.3)% of pre-HIT Wpeak to 78.1 (2.8)% of post-HIT Wpeak (P < 0.05). HIT decreased carbohydrate (CHO) oxidation, plasma lactate concentration and ventilation when the cyclists rode at the same absolute work rates of 60, 70 and 80% of pre-HIT Wpeak (P < 0.05), but not when they exercised at the same relative (% post-HIT Wpeak) work rates. Thus, the ability of the cyclists to sustain higher percentages of Wpeak in TT40 performances after HIT was not due to lower rates of CHO oxidation. Higher relative work rates in the TT40 rides following HIT increased the estimated rates of CHO oxidation from approximately 4.3 to approximately 5.1 g.min-1.