Coffee and incidence of diabetes in Swedish women: a prospective 18-year follow-up study

Objectives. To examine the long-term incidence of diabetes in relation to coffee consumption in Swedish women. Design. Prospective longitudinal cohort study. Setting. City of Goteborg, Sweden. Subjects. A random population sample of 1361 women, aged 39-65 years, without prior diabetes or cardiovascular disease took part in a screening study in 1979-1981 with questionnaires, physical examination and blood sampling. Main outcome measures. The development of diabetes until 1999 was identified by questionnaires in a second screening and the Swedish hospital discharge register. Results. Altogether, there were 74 new cases of diabetes. The risk of developing diabetes was 475 per 100 000 person-years in women who consumed two cups of coffee or less per day, 271 in women who consumed three to four cups per day, 202 with a consumption of five to six cups per day, and 267 in drinkers of seven cups or more per day. Associated hazard ratios, after adjustment for age, smoking, low physical activity, education and body mass index were 0.55 (0.32-0.95), 0.39 (0.20-0.77) and 0.48 (0.22-1.06) for daily consumption of three to four, five to six and seven cups or more, respectively, with a consumption of less than two per day as reference. Additional adjustment for serum cholesterol and triglycerides attenuated the relation between coffee and diabetes slightly, indicating a possible mediating effect on the effect of coffee by serum lipids. Conclusions. The findings of the present study support the hypothesis that coffee consumption protects from the development of diabetes in women.