Depression is associated with work absenteeism, reduced productivity, and significant personal and societal economic burden. We describe patterns and determinants of sick leave among working Danish antidepressant users.Persons starting antidepressant treatment (January 1, 2004 through December 31, 2005) were identified from a representative 25% sample of the Danish population by linking Danish national registries. Inclusion criteria were age 18–64 years, being in the workforce the week prior to the first antidepressant prescription (index prescription, IP), and no antidepressant prescription in the year prior to the IP. Only sick leaves >2 weeks are centrally registered in Denmark and could be assessed. Cox regression analyses identified predictors of sick leave during the year following the IP, based on previous history of sick leave and clinical and socio-demographic baseline characteristics.In the cohort of 25,908 (59.7% women), sick leave prevalence increased from 37.5% (year prior to IP) to 45.3% (year after the IP); 30.7% were on sick leave for >8 weeks. Incidence peaked (35.5% of individuals) the week after the IP. Of persons with sick leave in the year before the IP, 62.7% were on sick leave the first week after the IP, vs 5.7% of those without previous sick leave. Predictors associated with increased risk of sick leave among those without previous sick leave were unemployment, female gender, age 25–54 years, couples with children, and vocational and higher intermediate education (including e.g. teachers and nurses).Reasons for sick leave, sick leaves of less than 14 days and the indications for antidepressant treatment were unknown.Sick leave was prevalent in persons starting new antidepressant use, often lasting >8 weeks. Previous sick leave was the strongest predictor of subsequent sick leave.