Randomised controlled trials may have generalisability limitations when applied to the complex treatment of patients with bipolar disorder. Observational study designs can inform us about the diversity of bipolar disorder treatment in naturalistic settings. The aim of this paper was to describe the treatments prescribed for acute mania in a large prospective observational study of bipolar disorder. Patients with a manic/mixed episode were enrolled in EMBLEM (European Mania in Bipolar Longitudinal Evaluation of Medication) if they initiated or changed oral medication with antipsychotics, lithium and/or anticonvulsants. The use of monotherapy or combination therapy for treatment of acute mania, concomitant medications and rate of treatment switching during the 12-week acute treatment phase were assessed. Of the 3459 patients, 36% were treated with one drug and 64% with combination therapy. 55% of patients initiating combination therapy started on an atypical antipsychotic plus lithium or an anticonvulsant. Patients prescribed combination therapy at baseline were more clinically severe, were more often treated as inpatients and had more manic episodes in the previous year compared with the monotherapy group. Treatment switching occurred in 54.4% of patients over the 12-week acute phase. Many patients were taking at least one concomitant medication at baseline (69.4%) and week 12 (50.5%). The results of this observational study show that treatment for mania is complex with multiple combinations of treatment and frequent switching during an acute episode.