Seventy-one patients with Guillain-Barré syndrome (GBS) were retrospectively selected from within a defined area (Cantabria) in northern Spain, from 1975 to 1988. Excluding two non-resident cases, epidemiological analysis was based on 69 cases. The annual incidence rates were stable during the 14-year period of study with an average incidence of 0.95 (age-adjusted, 0.86) cases per 100,000 population. No significant difference was found for sex, urban or rural residence and there was no significant seasonal clustering. Antecedent event were recorded in 57% of patients, the most frequent events being upper respiratory infection and gastroenteritis. No association between use of gangliosides and the syndrome was found. Eight patients had variant syndromes including Fisher's syndrome (2 cases), and axonal (4 cases) and sensory (2 cases) GBS. Recurrences occurred in 3 cases. Excluding nine patients with incomplete follow-up and two with Fisher's syndrome, clinical analysis was based on 60 cases. Patients were divided into three groups as a function of their peak weakness. Significant features of the severe group were a requirement for ventilation, presence of bulbar palsy or dysautonomia and a longer duration of the plateau phase. However, it was not possible at an early stage of the clinical course to predict future motor deficit. Four (6.7%) patients belonging to the severe group died during the acute phase of the disease. No specific treatment for GBS was given. Outcome was assessed by means of serial examination up to 24 months after the onset of symptoms using a functional scale. At 3, 6 and 24 months 70%, 46% and 12% of patients, respectively, had a poor outcome.(ABSTRACT TRUNCATED AT 250 WORDS)