The effect of sedation induced by intravenous diazepam on cerebral glucose metabolic activity was examined with [18F]2-fluoro-2-deoxy-D-glucose (FDG) and positron emission tomography (PET) in five patients with probable Alzheimer's disease. Each subject was studied on 2 separate days: on one occasion at rest with eyes patched and ears open, and on the second when sedated with intravenous diazepam titrated to maintain stage II sleep by clinical and EEG criteria. Similar patterns of glucose uptake were observed in both the presence and the absence of sedation, but overall glucose utilization was depressed an average of 20% and was closely correlated with the amount of diazepam administered prior to the injection of FDG. The predominant temporoparietal hypometabolism and relative sparing of frontal metabolism observed in this disease are therefore not explained by differences in anxiety or activity level in this patient group. Utilization of diazepam sedation for PET study appears to be safe and may permit the study of patients otherwise unable to cooperate with FDG-PET procedures.