In this study, we investigated the impact of a controlled 4-day caffeine withdrawal period on the effect of an acute caffeine dose on endurance exercise performance. Twelve well-trained and familiarized male cyclists, who were caffeine consumers (from coffee and a range of other sources), were recruited for the study. A double-blind placebo-controlled cross-over design was employed, involving four experimental trials. Participants abstained from dietary caffeine sources for 4 days before the trials and ingested capsules (one in the morning and one in the afternoon) containing either placebo or caffeine (1.5 mg · kg(-1) body weight · day(-1)). On day 5, capsules containing placebo or caffeine (3 mg · kg(-1) body weight) were ingested 90 min before completing a time trial, equivalent to one hour of cycling at 75% peak sustainable power output. Hence the study was designed to incorporate placebo-placebo, placebo-caffeine, caffeine-placebo, and caffeine-caffeine conditions. Performance time was significantly improved after acute caffeine ingestion by 1:49 ± 1:41 min (3.0%, P = 0.021) following a withdrawal period (placebo-placebo vs. placebo-caffeine), and by 2:07 ± 1:28 min (3.6%, P = 0.002) following the non-withdrawal period (caffeine-placebo vs. caffeine-caffeine). No significant difference was detected between the two acute caffeine trials (placebo-caffeine vs. caffeine-caffeine). Average heart rate throughout exercise was significantly higher following acute caffeine administration compared with placebo. No differences were observed in ratings of perceived exertion between trials. A 3 mg · kg(-1) dose of caffeine significantly improves exercise performance irrespective of whether a 4-day withdrawal period is imposed on habitual caffeine users.