Regional glucose metabolic abnormalities are not the result of atrophy in Alzheimer's disease.

OBJECTIVE: To determine whether the hypometabolism observed in PET images of patients with Alzheimer's disease (AD) is due entirely to brain atrophy.
BACKGROUND: Reduced brain glucose metabolism in AD patients measured using PET has been reported by numerous authors. Actual glucose metabolic values in AD may be reduced artificially because of brain atrophy, which accentuates the partial volume effect (PVE) on data collected by PET.
METHODS: Using segmented MR images, we corrected regional cerebral metabolic rates for glucose for PVEs to evaluate the effect of atrophy on uncorrected values for brain metabolism in AD patients and healthy control subjects.
RESULTS: Global glucose metabolism was reduced significantly before and after correction in AD patients compared with controls. Before PVE correction, glucose metabolic values in patients were lower than in control subjects in the inferior parietal, frontal, and lateral temporal cortex; in the posterior cingulate; and in the precuneus. These reductions remained significantly lower after PVE correction, although in the posterior cingulate the difference in metabolism between AD patients and control subjects lessened. Regional glucose metabolism of these areas with PVE correction was lower in moderately-severely demented patients than in mildly demented patients.
CONCLUSION: Reduced glucose metabolism measured by PET in AD is not simply an artifact due to an increase in CSF space induced by atrophy, but reflects a true metabolic reduction per gram of tissue.