In 1977, Bier et al described the linear relation of total body glucose flux and brain size in individuals (premature infants to adults) (r=0.97). We used Bier's method to evaluate the usefulness of measurement of glucose flux by 6,6 dideuteroglucose in the certification of brain injury. During a 15 month period, a prospective study was done with 13 patients admitted to PICU. They were divided in three groups: A-control, B-Glasgow Coma Scale less than 10, C-brain death. Prior to study, enteral feedings were discontinued for 8 hours and IVF's with D5W were tapered down in 4 hours. An intravenous bolus of 6-6 dideuteroglucose was followed by a continuous infusion of the isotope. Bedside glucose monitoring was done every 30 minutes, and blood samples were drawn at 0, 120, 150, 180, 210 and 240 minutes for glucose, insulin and 6-6 dideuteroglucose enrichment. Expected glucose production rate (GPR) as a function of body weight was compared with the patient's GPR and expressed as percent of the expected. A low GPR was considered below 60%, borderline GPR was 60-80% and normal GPR was above 80%. Group A (n=3) had a mean GPR of 83%, group B (n=5) of 74% and group C (n=2) of 69%. All patients with normal neurologic outcomes had normal GPR. Patients with a low GPR died or, if survived experienced no improvement in their neurological status. Four patients that died had normal or borderline GPR. Two patients with normal or borderline GPR were discharged with moderate to severe brain damage. We conclude that the predictive value of GPR for a bad outcome in this group of patients is 100%. A normal GPR was not predictive of a good outcome. We speculate that patients with normal GPR but poor neurologic status at discharge may have a better recovery potential.