Demographic and clinical features and data on medical history and prior environmental exposure collected during an epidemiological long-term study of multiple sclerosis (MS) were tested for their possible prognostic value. Fifty-two benign MS patients were compared with 29 patients having a malignant course. A primary or secondary progressive course and cerebellar/lower brain-stem symptoms at onset indicated an unfavourable course, whereas no predictive value of sex or of any other type of onset symptomatology was found. Age at onset per se had no influence on prognosis but was associated with more rapid progression only by its relationship with a chronic progressive type of course. Prior illness, surgery, trauma and childhood exposure to defined environmental factors could not be identified as relevant for prognosis.