In short-term studies, caffeine has been shown to increase insulin levels, reduce insulin sensitivity, and increase cortisol levels. However, epidemiological studies have indicated that long-term consumption of beverages containing caffeine such as coffee and green tea is associated with a reduced risk of type 2 diabetes mellitus. There is a paucity of randomized studies addressing the metabolic and hormonal effects of consuming caffeine over periods of more than 1 day. We evaluated the effect of oral intake of 200 mg of caffeine taken twice a day for 7 days on glucose metabolism, as well as on serum cortisol, dehydroepiandrosterone (DHEA), and androstenedione, and on nighttime salivary melatonin. A double-blind, randomized, placebo-controlled crossover study with periods of 7 days and washouts of 5 days comparing caffeine with placebo capsules was conducted. Participants were 16 healthy adults aged 18 to 22 years with a history of caffeine consumption. Blood samples from each subject were assayed for glucose, insulin, serum cortisol, DHEA, and androstenedione on the eighth day of each period after an overnight fast. Nighttime salivary melatonin was also measured. Insulin levels were significantly higher (by 1.80 microU/mL; 95% confidence interval, 0.33-3.28) after caffeine intake than after placebo. The homeostasis model assessment index of insulin sensitivity was reduced by 35% (95% confidence interval, 7%-62%) by caffeine. There were no differences in glucose, DHEA, androstenedione, and melatonin between treatment periods. This study provides evidence that daily caffeine intake reduces insulin sensitivity; the effect persists for at least a week and is evident up to 12 hours after administration.