We compared the effects of caffeinated vs non-caffeinated carbohydrate electrolyte (CE) drinks on urine volume (UV), free water clearance (CH2O), fractional excretion of water (FEH2O), and osmolar excretion during 4 h of rest or 1 h rest followed by 3 h of cycling at 60% VO2max in six subjects. We also tested maximal performance at 85% VO2max following the 3-h exercise trials. Throughout the two resting trials and the two rest + exercise trials, subjects ingested CE (total volume = 35 ml/kg) without (PLAC) or with (CAFF) caffeine (25 mg/dl). Blood samples were collected, and body weight and UV were recorded every hour. Urine and blood were analyzed for osmolality and creatinine, and plasma catecholamine concentrations were determined. At rest, mean (+/-SE) UV between 60 min and 240 min was greater for CAFF (1843 +/- 166 ml) vs PLAC (1411 +/- 181 ml) (p < 0.01); during exercise the difference in UV between CAFF (398 +/- 32 ml) and PLAC (490 +/- 57 ml) was not significant. Cycling performance was unaffected by caffeine. Plasma catecholamine concentrations were not different between PLAC and CAFF but were greater during exercise than rest (p < 0.01) and may have counteracted the diuretic effect of caffeine observed at rest. Thus, CAFF consumed in CE during moderate endurance exercise apparently does not compromise bodily hydration status.