Caffeine is considered to have a diuretic effect in doses that are commonly ingested in the diet of normal individuals living in the UK (Neuhèuser-Berthold et al. 1997). In the present study, which was approved by the local ethics committee, seven individuals, who all consumed daily at least the equivalent of two cups of instant coffee (4 males, 3 females) were dehydrated by a combination of exercise in the heat followed by overnight fluid restriction. Subjects were tested on four separate occasions in the morning at least 12 h after exercising. On two occasions subjects were rehydrated following the dehydrating exercise by ingesting, with their evening meal, a water volume equivalent to twice their exercise-induced sweat loss plus 500 ml before going to bed and 500 ml in the morning 2 h before coming to the laboratory. On both hypohydrated trials, after the dehydrating exercise subjects drank only 100 ml of water with their evening meal and consumed no other liquids before being tested in the morning. The same meals were given to the subjects on all four trials. Each morning, subjects drank, in 30 min, 1000 ml of either decaffeinated coffee or the same coffee to which had been added 250 mg of caffeine; each trial was performed in an euhydrated and hypohydrated state. Urine samples were collected before ingestion and at hourly intervals for 4 h after drinking finished. All subjects were familiarised with the study procedures before undertaking their first experimental trial; the treatment order was carried out using an incomplete block design. Data were found to be normally distributed and statistical significance (P < 0.05) was determined using repeated-measures and one-way ANOVA with Tukey’s post-hoc analysis as appropriated.

Initial mean (± S.D.) body mass (72.5 ± 16.7 kg) was similar (P = 0.97) at the beginning of all trials. Immediately before ingesting the test solutions, subjects on both hypohydrated trials were dehydrated to a similar extent (1.9 ± 0.4% P = 0.63) of initial body mass (mean ± S.D.). Peak urine output was produced 1 h after finishing drinking on all trials. Similar mean (± S.D.) urine volumes were produced at that time on the hypohydrated trials whether subjects drank decaffeinated (HD, 266 ± 158 ml) or caffeine-containing coffee (HC, 148 ± 117 ml); this was less (P = 0.001) than on the euhydrated trials following ingestion of the decaffeinated (ED, 766 ± 209 ml) and caffeine-containing coffee (EC, 774 ± 161 ml), which were essentially the same (P > 0.05). Cumulative urine output over the 4 h measurement period was similar (P > 0.05) on trials HD (518 ± 177 ml) and HC (292 ± 179 ml), but both were less (P = 0.001) than that on trial ED (1289 ± 266 ml) and EC (1368 ± 325 ml), which were essentially the same (P = 0.66).

The lack of a caffeine-induced increase in urinary output in both well hydrated and hypohydrated subjects in this study clearly demonstrates that moderate intakes of caffeine do not have a significant diuretic effect in individuals who normally consume caffeine in the diet.