The aim of this study was to examine the relationship between mixed and pure mania using both narrow (DSM-III-R) and broad (Cincinnati) operational diagnostic criteria to define mixed mania regarding the degree of associated depression. Hospitalized patients aged > or = 12 years and meeting DSM-III-R criteria for bipolar disorder, manic or mixed, were compared regarding demographics, phenomenology, course of illness, comorbidity, family history, and short-term outcome. Seventy-one patients were recruited during a 1-year period. Twenty-four patients (34%) met DSM-III-R criteria for mixed bipolar disorder; 28 (40%) met the broader definition (which required three associated depressive symptoms rather than full syndromal DSM-III-R depression). Compared with pure manic patients, DSM-III-R mixed patients had significantly more depressive symptoms, were more likely to be female, experienced more prior mixed episodes, displayed higher rates of comorbid obsessive-compulsive disorder, and had longer hospitalizations. However, when mixed mania was defined more broadly, differences in gender and hospitalization length were lost. Also, regardless of the definition used, mixed and pure manic patients were similar on most other variables assessed. We conclude that mixed and pure mania differ in some respects but have many similarities, especially when mixed mania is defined by lesser degrees of depression. The use of dimensional rather than categoric systems to describe the degree of associated depression may be a more meaningful method of classifying mania.