One hundred and six affective (76 unipolar and 30 bipolar) and 101 schizoaffective patients (45 unipolar and 56 bipolar) were investigated after a long-term course of illness, evaluating sociodemographic and general data, the long-term course of illness, disability and psychosocial alterations according to WHO/DAS, WHO/PIRS and GAS, as well as several social consequences of the illness (living situation at the end of the observation time, downward occupational drift, downward social drift, premature retirement, achievement of the expected social development). Comparing the 30 bipolar affective and 56 bipolar schizoaffective disorders, no differences were found regarding (a) sociodemographic and general data (i.e. sex distribution, age at onset, education and occupation at onset, stable heterosexual relationship, premorbid personality and social interactions, mental illness in the family, broken home, life events, season of birth and social classes) and (b) relevant patterns of the long-term course. Regarding long-term outcome, the only difference found concerned the more favourable outcome of the bipolar affective disorders according to WHO/DAS, while using GAS the difference was not statistically significant. No difference was found either between the two bipolar groups in the social consequences of the illness. The combining of both bipolar groups as "bipolar diseases" is discussed, as well as the use of the terms "bipolar disease, affective subtype" and "bipolar disease, schizoaffective subtype".