The red man syndrome (RMS) is the most common toxicity of vancomycin therapy. A prospective trial to investigate the frequency, causation, and possible prophylaxis of this syndrome was conducted. Thirty-three patients were observed during their first two doses (1 g/60 min) of vancomycin. Before dose 1, they were randomized to double-blinded pretreatment with either diphenhydramine (50 mg) or placebo. Patients were examined frequently, and histamine levels were obtained at 0, 30, and 60 min during dose 1. Those with first-dose reactions were rerandomized for pretreatment and had histamine levels drawn during a second infusion. Of 17 patients with placebo pretreatment, 8 (47%) had RMS. None of the 16 pretreated with diphenhydramine had a first-dose reaction (P = .003). Three of the eight first-dose reactors had a second-dose RMS reaction; in one of these three, it was more severe than the dose 1 RMS despite diphenhydramine pretreatment. RMS events were associated with elevated plasma histamine; this was especially true of severe reactions.