The authors enrolled 168 patients with closed-head trauma into a prospective trial to evaluate the effect of hyperbaric oxygen in the treatment of brain injury. Patients were included if they had a total Glasgow Coma Scale (GCS) score of 9 or less for at least 6 hours. After the GCS score was established and consent obtained, the patient was randomly assigned, stratified by GCS score and age, to either a treatment or a control group. Hyperbaric oxygen was administered to the treatment group in a monoplace chamber every 8 hours for 1 hour at 1.5 atm absolute; this treatment course continued for 2 weeks or until the patient was either brain dead or awake. An average of 21 treatments per patient was given. Outcome was assessed by blinded independent examiners. The entire group of 168 patients was followed for 12 months, with two patients lost to follow-up study. The mortality rate was 17% for the 84 hyperbaric oxygen-treated patients and 32% for the 82 control patients (chi-squared test, 1 df, p = 0.037). Among the 80 patients with an initial GCS score of 4, 5, or 6, the mortality rate was 17% for the hyperbaric oxygen-treated group and 42% for the controls (chi-squared test, 1 df, p = 0.04). Analysis of the 87 patients with peak intracranial pressures (ICP) greater than 20 mm Hg revealed a 21% mortality rate for the hyperbaric oxygen-treated patients, as opposed to 48% for the control group (chi-squared test, 1 df, p = 0.02). Myringotomy to reduce pain during hyperbaric oxygen treatment helped to reduce ICP. Analysis of the outcome of survivors reveals that hyperbaric oxygen treatment did not increase the number of patients in the favorable outcome categories (good recovery and moderate disability). The possibility that a different hyperbaric oxygen treatment paradigm or the addition of other agents, such as a 21-aminosteroid, may improve quality of survival is being explored.