Hyperglycemia, which may be caused or exacerbated by conventional diets, may worsen the neurological outcome from severe head injury, especially if secondary ischemic insults occur. The purpose of this study was to evaluate an experimental diet intended to replace systemic caloric and protein requirements without producing hyperglycemia. In initial studies in the laboratory, 5 experimental diets were employed in a middle cerebral artery temporary occlusion model. The effects of the diets on blood biochemistry and on infarction volume were compared in fasted animals and in animals fed a control diet. Animals fed the experimental diets had a significantly lower preischemia blood glucose concentration, a higher blood concentration of ketone bodies, and a smaller infarct volume than the animals fed a control diet. One diet chosen from the laboratory study was then evaluated in a clinical study as a randomized, open-label trial. Twenty severely head-injured patients were randomly assigned to be fed the experimental diet, EN-9305, or the control diet, Osmolyte HN, for the first 2 weeks after injury. Both treatment groups had similar blood glucose concentrations, averaging 6.33 +/- 0.21 mumol/mL (114 +/- 4 mg/dL), on day 1 prior to starting the assigned diet. Blood glucose concentration increased in the control diet group to a peak of 8.37 +/- 0.94 mumol/mL (151 +/- 17 mg/dL) on day 7 as the infusion rate of the diet was increased to the final rate. In the experimental diet group, the blood glucose concentration remained unchanged from fasting levels as the diet was advanced. Blood lactate concentration was lower, and blood ketone body concentrations were higher in the patients fed the experimental diet. Urinary nitrogen balance was better in the experimental diet group, but measures of visceral protein sparing, including serum albumin, plasma retinol binding protein, and total lymphocyte count, were not significantly different in the 2 treatment groups. Measures of cerebral anaerobic metabolism, including CSF lactate concentration and cerebral lactate production, were not significantly different in the 2 treatment groups. These studies suggest that a carbohydrate-free diet such as EN-9305 might have advantages for patients with severe head injury by replacing systemic caloric and protein requirements without producing hyperglycemia.