Precise information was obtained by a computerized analysis of the treatment, prognosis and clinical outcome of 108 representative cases selected from a series of 160 patients with Horton's disease (HD) diagnosed between 1970 and 1981. At the end of the study, 18 patients had died (16,6 p. 100), the commonest cause of death being a cerebrovascular accident. Death occurred during the first two years of the illness in half these cases. Despite these early complications of the disease, from a statistical point of view, the patients did not die of HD or the complications of its treatment. The 5 years survival rate 80,1 p. 100 (actuarial analysis). Most deaths were from natural causes; an intercurrent disease was often implicated; the lower probability of survival for men (69,9 p. cent at 5 years, compared to 85,4 p. cent for women) is compatible with the lower life expectancy which is well established. The roles of intercurrent disease and sex were statistically significant. On the other hand, there was no difference in the probability of 5 year survival when the different presenting of the disease, the terrain (HLA-DR4) and choice of corticosteroid therapy were compared. The initial dose of prednisone or prednisolone should exceed 0,5 mg/kg/day, if relapse is to be avoided. No additional benefits were observed with doses in excess of 1 mg/kg/day except in cases with complications from the onset (ocular, coronary insufficiency, stenosis of the main large arteries, without being able to prejudge their mechanism).(ABSTRACT TRUNCATED AT 250 WORDS)