Initially based on empirical observations, the concept of RBD has led to operational diagnostic criteria allowing further studies in various and large populations. In line with the hypothesis of a continuum from normal sad mood to major depression, the spectrum of affective disorders is thus probably more in agreement with everyday clinical practice. Moreover, recurrent and so-called subthreshold conditions have been concurrently investigated in other fields of psychopathology. Nevertheless, many methodological problems and clinical implications of RBD still require further research. The question of a precise definition of the disorder, even if a consensus seems to be obtained about some key points, remains partially to be addressed, for example, concerning the problem of the reliability of assessment. The possibility of a seasonality of the episodes needs further investigations, as do the relationships between RBD and personality disorders such as borderline personality disorder. Finally, the important question of prophylactic treatment of RBD remains unsolved, as antidepressants have failed to demonstrate any efficacy and some neuroleptics have been proposed in particular conditions. Angst's prediction that clinical relevance, impairment, distress and public health impact of affective disorders could be related not only to the duration of an episode but also to the recurrence of episodes, possibly referring to a kindling model, has therefore been confirmed. His major contribution has been to underline, in various forms of affective disorders, the importance of conducting longitudinal clinical as well as epidemiologic studies to refine our psychopathological knowledge.