The relation of coffee consumption to the risk of nonfatal first myocardial infarction in men under 55 years of age was assessed in a hospital-based case-control study conducted from 1980 to 1983 in hospitals in Massachusetts, Rhode Island, Connecticut, and New York: 1,873 men with first nonfatal myocardial infarctions were compared with 1,161 controls admitted for conditions unrelated to coffee ingestion. After allowance for major risk factors for myocardial infarction, the relative risk estimate for recent consumption of caffeine-containing coffee increased with increasing level of daily intake, from 1.4 for one to two cups per day to 1.6 for three to four cups, 1.8 for five to nine cups, and 2.9 for greater than or equal to 10 cups, relative to consumption of no coffee (p less than 0.001 for trend). The association was apparent in each age group and in both smokers and nonsmokers. For those who drank decaffeinated coffee only, on the basis of small numbers, there was a suggestion of an increased risk among men who had consumed at least five cups daily for less than five years but not among those who had drunk this amount for at least five years; whether the apparent association among the shorter-term drinkers was due to previous consumption of caffeine-containing coffee could not be determined. The findings suggest that caffeine-containing coffee increases the risk of myocardial infarction and that men who drink at least five cups daily may increase their risk by about twofold or more.