If upper respiratory tract infections could be avoided, episodes of optic neuritis following virus infections would be prevented. High doses of glucocorticoids may modify the course of optic neuritis and multiple sclerosis. In the first analysis of the 1992 multicenter optic neuritis study, multiple sclerosis developed in fewer patients treated with high-dose methylprednisolone than in groups treated with 60 mg/day of placebo or prednisone followed by a rapid taper (Beck and Cleary 1993b). However, after refining the diagnostic criteria for multiple sclerosis, the same group found no effect of steroids on development of multiple sclerosis. In another study, the mode of steroid administration and steroid discontinuation profoundly altered the probability of future attacks of optic neuritis (Herishanu et al 1989). Patients with optic neuritis were treated for 3 days with 1 g of intravenous methylprednisolone and no taper. Sixty-six percent had recurrent bouts of optic neuritis, and 83% of the patients developed multiple sclerosis within 18 months. In contrast, recurrent optic neuritis developed in only 14% of untreated patients, and in only 33% of patients treated with 60 mg of oral prednisone for 10 days followed by a month taper. No untreated patients, and 7% of patients treated with prednisone, developed multiple sclerosis during a 6-year follow-up. These studies, and similar provocative effects after sudden steroid withdrawal in experimental autoimmune encephalitis, suggest that rapid discontinuation of steroid therapy may be dangerous (Reder et al 1994).