Twelve young healthy male volunteers participated in a randomized blind study in which they consumed 150 ml decaffeinated coffee with or without 200 mg added caffeine. Resting metabolic rate by indirect calorimetry (ventilated hood), skin temperature on chest, arm, thigh, calf and abdomen, and internal temperature (rectal) were measured in the fasting state and up to 3 h after coffee consumption. All treatments were done in duplicate. After caffeine consumption the metabolic rate increased immediately with 0.2 +/- 0.2 kJ/min (p < 0.05) and remained elevated for the 3 h during which measurements were taken. Mean total caffeine-induced thermogenesis was 0.30 +/- 0.20 kJ/min, which means a mean increase in the metabolic rate of 7 +/- 4% during 3 h. The internal temperature started to increase after both treatments, but became slightly different 2 h after caffeine consumption. Mean skin temperature after caffeine consumption increased, and became significantly different compared to the control treatment after 90 min. Thus, the increase in both internal temperature and skin temperature after caffeine consumption was largely delayed compared to the direct detectable increased energy expenditure as measured by indirect calorimetry. The correlation between the change in metabolic rate during the 3 h after the caffeine consumption and the mean change in skin temperature was low (r = 0.38; p = 0.06), and does not seem to justify the use of skin temperature to measure energy expenditure or changes in energy expenditure in short-term studies.