Objective: To examine the cross-sectional and longitudinal associations between benzodiazepine use and daily activity limitations, according to drug indications and duration of action.Design: Prospective cohort study.Setting: Population-based 3-city studyParticipants: 6 600 participants aged 65 years and over included between 1999 and 2001 and followed after 2, 4 and 7 years. Measurements: Benzodiazepine users were separated into hypnotic, short-acting anxiolytic and long-acting anxiolytic users and compared with non users. Three outcomes were examined assessing restrictions in mobility, instrumental activities of daily living (IADL) and social participation. Results: In multivariate simple or mixed logistic models adjusted for socio-demographic variables, impairments and comorbidity and for anxiety, insomnia and depression, hypnotic benzodiazepines were moderately associated with mobility limitation prevalence and IADL limitation incidence. Short-acting and long-acting anxiolytics were associated with IADL limitation prevalence and with mobility limitation prevalence and incidence and long-acting anxiolytics were also associated with IADL limitations incidence. Chronic benzodiazepines users were at a marked risk of developing restrictions for the three outcomes (odds ratio 1.71 (95% confidence interval (CI), 1.23-2.39) for mobility, 1.54 [95%CI, 1.14-2.10] for IADL and 1.74 [95%CI, 1.23-2.47] for participation limitations).Conclusions: Benzodiazepine users are at increased risk of activity limitations regardless of the duration of action or indication. Chronic use of benzodiazepines should be avoided in order to extend disability-free survival.