Objectives: Increasing evidence points to the harmful effects of long-term benzodiazepine treatment. Our objective was to study the incidence of, and predictors for, long-term use of benzodiazepines and Z-drugs in bipolar disorder. Methods: We conducted a population-based cohort study, using data from Swedish national registers. Swedish residents aged 18-75 years with a recorded diagnosis of bipolar disorder or mania between July 2006 and December 2012, and no history of benzodiazepine/Z-drug use in the past year, were included. Patients were followed for 1 year with regard to prescription fills of benzodiazepines/Z-drugs. Initiators were followed for another year during which continuous use for >6 months was defined as “long-term”. Patient and prescription characteristics were investigated as potential predictors for long-term use in multivariate logistic regression models. Results: Out of the 21 883 patients included, 29% started benzodiazepine/Z-drug treatment, of whom one in five became long-term users. Patients who were prescribed clonazepam or alprazolam had high odds for subsequent long-term use (adjusted odds ratios [aORs] 3.78 [95% confidence interval (CI) 2.24-6.38] and 2.03 [95% CI 1.30-3.18], respectively), compared to those prescribed diazepam. Polytherapy with benzodiazepines/Z-drugs also predicted long-term use (aOR 2.46, 95% CI 1.79-3.38), as did age ≥60 years (aOR 1.93, 95% CI 1.46-2.53, compared to age <30 years), and concomitant treatment with psychostimulants (aOR 1.78, 95% CI 1.33-2.39). Conclusions: The incidence of subsequent long-term use among bipolar benzodiazepine initiators is high. Patients on clonazepam, alprazolam or benzodiazepine/Z-drug polytherapy have the highest risk of becoming long-term users, suggesting that these treatments should be used restrictively.