Background In a cross-section approach we investigated prescription practice in acute mania in 63 German, Swiss and Austrian hospitals between 1994 and 2004. Methods Our data were gathered from a large drug safety program (AMSP) within which on two reference days each year all administered drugs are recorded. For the present study, all cases with a primary diagnosis of acute euphoric mania ( n = 1291) or mixed-state mania ( n = 143) were identified. Prescription rates from two periods, 1994 to 1999 and 2000 to 2004, were compared. Results In euphoric mania, prescription of lithium decreased by about one-fifth (43.3% to 34.5%, p < 0.01), while prescription of anticonvulsants increased by one-half (from 40.0% to 60.7%, p < 0.001). Administration of atypical antipsychotics more than doubled (18.5% to 43.9%, p < 0.001), while use of typical antipsychotics decreased significantly (56.9% to 27.8%, p < 0.001). Overall prescription rates of antipsychotics (79.6% vs. 81.6%) and antidepressants (14.0% vs. 15.5%) remained stable, while administration of tranquilizers increased significantly (26.3% to 34.3%, p < 0.01). In mixed-state mania, similar trends over time to those seen in euphoric mania were observed for lithium (43.2% to 33.3%), anticonvulsants (50.0% to 69.7%, p < 0.05) and tranquilizers (22.7% to 40.4%). Prescription rates of antipsychotics slightly increased (63.6% to 72.7%), while prescription of antidepressants slightly decreased (54.5% to 46.5%). Polypharmacy was a common phenomenon: patients with euphoric mania were treated with a mean number of 2.9 ± 1.2 psychotropic agents, and patients with mixed-state mania with 3.3 ± 1.5 psychotropic agents. Both groups showed a significant increase over time. Second-generation atypical antipsychotics were adopted quite rapidly for the treatment of acute mania considering the availability of the scientific evidence at that time. Off-label use was a common phenomenon. Deviations from recommended guidelines were found mainly in the use of antidepressant and antipsychotic drugs both in mixed-state and euphoric mania. Conclusions Naturalistic prescription studies like this may encourage a critical scrutiny of clinical treatment habits and may also drive further research thus moderating potential differences between evidence-based knowledge and everyday clinical practice.