Depressive symptoms during mania have prognostic value in bipolar disorder. For depressive symptoms, it has been proposed that shorter scales should be cost-effective and practical. To determine the usefulness of 5-item Hamilton Depression Rating Scale (HAMD-5) in manic and mixed bipolar disorder, we used a four-week follow-up prospective, observational study. Convergent and discriminant validity, internal consistency, and reliability were analyzed and compared with HAMD-21, HAMD-5, and HAMD-21 cut-off points were calculated versus CGI-BP. A total of 173 manic and mixed patients were evaluated. HAMD-5 showed appropriate convergent validity, discriminant validity, internal consistency, and test-retest reliability. Discriminant validity was higher for HAMD-5 than HAMD-21. Best cut-off point of remission was: HAMD-21 < or =5 and HAMD-5 < or =1. HAMD-5 presents appropriate validity and reliability estimates. It is comparable to HAMD-21 and focuses more specifically on depressive symptoms.