Strenuous exercise is associated with transient alterations in neutrophil number and function (Smith, 1997). Caffeine is a non-selective adenosine receptor antagonist consumed by athletes for its ergogenic properties (Graham, 2001). Neutrophils express both A₁ and A₂a adenosine receptors with their occupancy producing differing responses (Thibault et al., 2002) yet little is known about the effect of caffeine on neutrophil responses to exercise. Therefore, the aim of this study was to investigate the effect of caffeine ingestion on neutrophil counts and the formyl-methionyl-leucyl-phenylalanine (fMLP)-induced oxidative burst response following prolonged cycling. Following local ethics committee approval, 8 endurance trained males (mean ± SEM: age 24 ± 1 years; height 1.79 ± 0.03 m; body mass 72.8 ± 2.8 kg; VO₂max 65.6 ± 1.9 ml.kg^-1.min^-1) cycled for 90 min on a stationary ergometer at 72.3 ± 0.7% VO₂max. On two occasions, separated by 1 week, participants arrived at the laboratory following an overnight fast and 60 h abstention from caffeine containing products and were randomly assigned to ingest either 6 mg.kg^-1 body mass of caffeine (CAF) or dextrose powder (PLA) 60 min before exercise. Participants consumed water only (2 ml.kg^-1 body mass) every 15 min during exercise. Venous blood samples were collected at rest, pre-exercise, immediately post-exercise and 1 h post-exercise. Neutrophil counts were performed using an automated cell counter. The in vitro neutrophil oxidative burst response to fMLP was assessed using a chemiluminescence (CL) assay (Knight Scientific Limited, Plymouth, UK). Serum caffeine and plasma adrenaline were determined using a spectrophotometric assay (Dade Behring, Milton Keynes, UK) and high-performance liquid chromatography, respectively. Results were analysed using a two-factor repeated measures ANOVA. Post hoc t tests with Holm-Bonferroni adjustment were applied where appropriate. Statistical significance was accepted at P<0.05. Immediately post-exercise, serum caffeine and plasma adrenaline were significantly higher on CAF than PLA (P<0.01). At this time, blood neutrophil counts were ∼125% greater than at rest (main effect of time, P<0.01) however, peak fMLP-induced neutrophil CL had fallen significantly on PLA only (CAF: rest, 11.3 ± 1.3 relative light units (RLU), post-exercise, 6.2 ± 0.7 RLU (P>0.05); PLA: rest, 13.5 ± 2.1 RLU, post-exercise, 4.9 ± 0.7 RLU, P<0.01). These findings suggest that in addition to its known ergogenic effects, caffeine may attenuate the exercise-induced fall in the neutrophil oxidative burst response to fMLP.