Mixed bipolar states are associated with more severe symptoms and outcome. Our aim is to review the literature examining their treatment. We conducted a literature search of randomized clinical studies and post-hoc analyses on mixed bipolar states' treatment. Remarkably, there is only one double-blind, placebo-controlled trial, recruiting a mixed episode cohort, and one post-hoc analysis of this trial, while most data come from post-hoc analyses of trials including both manic and mixed patients. Improvement of manic symptoms in mixed episodes is similar to that seen in pure manic episodes and independent of baseline depressive features. The magnitude of response to manic symptoms' treatment probably exceeds that of depressive symptoms, which appear to resolve later. Valproate and carbamazepine are effective in acute mixed episodes, but the efficacy of lithium appears questionable. Atypical antipsychotic monotherapy improves both manic and depressive symptoms. Mood-stabilizer-atypical antipsychotic combination increases this effect. Atypical antipsychotic-antidepressant combination against acute mixed depression does not increase the risk for mania, although its superior efficacy vs. atypical antipsychotic monotherapy cannot be supported by current data. As regards prophylaxis, atypical antipsychotic monotherapy is associated with a lower incidence of and a longer time to relapse of any kind. The augmentation of lithium or divalproex with atypical antipsychotics increases prophylactic efficacy. Lithium or divalproex monotherapy have not been associated with significant prophylactic benefits following mixed mania. New, randomized prospective trials involving homogeneous cohorts of mixed bipolar patients are needed in order to delineate the appropriate pharmacological treatment of mixed states.