In the present study, conducted in northern Italy between 1986 and 1989, the authors investigated the possible association between coffee consumption and serum cholesterol levels in 8,983 subjects, 7,432 men and 1,551 women, managers and employees aged 18-65 years, who were examined during a program of preventive medicine upon an agreement between various companies and the Centro Diagnostico Italiano. Analysis of covariance was used to compare the serum cholesterol levels of the subjects subdivided according to coffee consumption, along with age, body mass index, alcohol consumption, cigarette smoking, and physical activity. An important relation was demonstrated between coffee intake and cholesterol, particularly in the men, the differences in serum cholesterol in the coffee users compared with the nonusers being 6.1 +/- 1.4 (standard error) mg/dl for consumers of 1-3 cups/day (3.4 +/- 1.4 mg/dl after adjustment for age, body mass index, alcohol and cigarette consumption, and physical activity), 9.9 +/- 1.6 mg/dl for those drinking 4-5 cups/day (5.8 +/- 1.6 mg/dl after adjustment), and 14.8 +/- 2.0 mg/dl for those drinking over 5 cups/day (9.6 +/- 2.0 mg/dl after adjustment). This relation remained substantially unvaried when nonsmokers and smokers were analyzed separately. It has been suggested that it is coffee prepared by boiling rather than other methods that has a hypercholesterolemic effect. Our observations demonstrate an interesting relation between coffee and cholesterol, even though the coffee drunk in Italy is mainly filtered and nonboiled. However, our finding is not necessarily in disagreement with the above hypothesis since, when coffee is prepared in the Italian way (with the mocha method), ground coffee is preheated by steam and more importantly, the water passes through the ground coffee at a higher temperature than with the other brewing methods.