In a prospective study of the association between self-reported diabetes mellitus and sick leave from work, 21,149 men and women, aged 35-59 years, participated. Prevalence of diabetes was obtained by questionnaire. Sick leave was prospectively obtained from departments of human resources in the year after the baseline survey. Outcome variables related to duration of sick leave, repetitive absences, and long absences of at least seven consecutive days. In gender-specific multivariate analyses the relation between diabetes and sick leave was studied independent of age, body mass index, physical activity, alcohol intake, education, job title, and smoking habit. Prevalence rates of self reported diabetes were 2.8% and 2.2% in men and women, respectively. For each outcome variable a significant association between diabetes and sick leave was observed. Significant associations were found between diabetes and duration of sick leave and repetitive absences in men. In women a similar trend was observed for duration, and a significant association was found for repetitive absences.