BackgroundAffective temperaments are presumed to underlie bipolar psychopathology. The TEMPS-A has been widely used to assess affective temperaments in clinical and non-clinical samples. Cross-sectional research supports the association of affective temperaments and mood psychopathology; however, longitudinal research examining risk for the development of bipolar disorders is lacking. The present study examined the predictive validity of affective temperaments, using the TEMPS-A, at a three-year follow-up assessment.MethodsThe study interviewed 112 participants (77% of the original sample) at a three-year follow-up of 145 non-clinically ascertained young adults psychometrically at-risk for bipolar disorders, who previously took part in a cross-sectional examination of affective temperaments and mood psychopathology.ResultsAt the reassessment, 29 participants (26%) met criteria for bipolar spectrum disorders, including 13 participants who transitioned into disorders during the follow-up period (14% of the originally undiagnosed sample). Cyclothymic/irritable and hyperthymic temperaments predicted both total cases and new cases of bipolar spectrum disorders at the follow-up. Cyclothymic/irritable temperament was associated with more severe outcomes, including DSM-IV-TR bipolar disorders, bipolar spectrum psychopathology, major depressive episodes, and substance use disorders. Hyperthymic temperament was associated with bipolar spectrum psychopathology and hypomania, whereas dysthymic temperament was generally unassociated with psychopathology and impairment.LimitationsThe present sample of young adults is still young relative to the age of onset of mood psychopathology.ConclusionsThese results provide the first evidence of the predictive validity of affective temperaments regarding risk for the development of bipolar psychopathology. Affective temperaments provide a useful construct for understanding bipolar psychopathology.