BACKGROUND: Mixed states are a complex entity in the field of mood disorders. Dysphoria has been advocated as an important clinical dimension of mixed states. The objective of this work is to study the frequency of dysphoria within a population of patients with DSM-IV major depressive and/or manic episodes and to determine if it may help establish diagnostic criteria for subthreshold cases of depressive or manic mixed states. SAMPLING AND METHODS: A total of 165 patients were assessed using the Mini International Neuropsychiatric Interview complemented by a section defining dysphoria as a constellation of 3 among 4 symptoms (inner tension, irritability, aggressive behavior and hostility). RESULTS: When classifying patients according to the number of symptoms of the opposite polarity, changes in the frequency of dysphoria revealed a clear contrast between the 2 opposite manic and depressive poles and the full mixed state (DSM-IV definition). The frequency of dysphoria was 17.5% in pure depression, 22.7% in pure mania and 73.3% in full mixed state. Two threshold effects were identified: (1) the frequency of dysphoria increased from 17.5 to 61.1% (p = 0.002) when the number of manic symptoms in DSM-IV depressed patients increased from 0 to 1, and (2) dysphoria increased from 14.3 to 69.2% (p = 0.057) when the number of depressive symptoms increased from 2 to 3 in DSM-IV manic patients. CONCLUSION: Dysphoria is strongly but not necessarily associated with mixed states. When used as a clinical marker for mixed states, dysphoria confirms the modern delimitations of sub-threshold mixed states by specifying the required number of symptoms of the opposite polarity (which could be lower for depressive mixed states than for manic mixed states). The study has limitations related to the inclusion of patients who are not drug-free, to the definition of dysphoria and to the sample size.