Several prospective epidemiologic studies over the past 4 y concluded that ingestion of caffeinated and decaffeinated coffee can reduce the risk of diabetes. This finding is at odds with the results of trials in humans showing that glucose tolerance is reduced shortly after ingestion of caffeine or caffeinated coffee and suggesting that coffee consumption could increase the risk of diabetes. This review discusses epidemiologic and laboratory studies of the effects of coffee and its constituents, with a focus on diabetes risk. Weight loss may be an explanatory factor, because one prospective epidemiologic study found that consumption of coffee was followed by lower diabetes risk but only in participants who had lost weight. A second such study found that both caffeine and coffee intakes were modestly and inversely associated with weight gain. It is possible that caffeine and other constituents of coffee, such as chlorogenic acid and quinides, are involved in causing weight loss. Caffeine and caffeinated coffee have been shown to acutely increase blood pressure and thereby to pose a health threat to persons with cardiovascular disease risk. One short-term study found that ground decaffeinated coffee did not increase blood pressure. Decaffeinated coffee, therefore, may be the type of coffee that can safely help persons decrease diabetes risk. However, the ability of decaffeinated coffee to achieve these effects is based on a limited number of studies, and the underlying biological mechanisms have yet to be elucidated.