We compared high-dose dexamethasone (initial dose, 3 mg per kilogram of body weight) with placebo in a randomized, double-blind trial involving 38 patients with culture-positive, specifically defined severe typhoid fever. The patients in the two treatment groups ranged in age from 5 to 54 and were comparable at the outset. All patients received chloramphenicol. The case-fatality rate of 10 per cent (2 of 20 patients) in the dexamethasone group was significantly lower than the fatality rate of 55.6 per cent (10 of 18) in the placebo group (P = 0.003). There was no significant difference in the incidence of complications among the survivors in either group. Delirium, obtundation, and stupor were grave prognostic signs that were useful for predicting which patients were at high risk of dying before they became comatose or went into shock. Dexamethasone is unnecessary for most patients with typhoid but is recommended for all patients with suspected typhoid fever who are delirious, obtunded, stuporous, comatose, or in shock.