Effects of subcortical cerebral infarction on cortical glucose metabolism and cognitive function

Abstract

Background: The mechanism of dementia in subcortical cerebral infarction is incompletely understood.

Objective: To determine how cognitive function is related to cortical metabolism in patients with subcortical infarction and a continuum of cognitive impairment.

Methods: We used positron emission tomography (PET) and the glucose metabolic tracer fludeoxyglucose F 18 to study 8 patients with subcortical stroke and normal cognitive function (S-CN), 5 patients with subcortical stroke and cognitive impairment (S-CI) who did not have dementia, 8 patients with subcortical stroke and dementia (S-D), and 11 controls with no cognitive impairment or stroke. A subset of patients had absolute regional cerebral metabolic rate of glucose (CMRglc) determined, while in all subjects regional tracer uptake normalized to whole brain tracer uptake was calculated. PET data were analyzed by constructing volumes of interest using coregistered magnetic resonance imaging data and correcting the PET data for atrophy.

Results: Global CMRglc was significantly lower in the patients with S-D than in the control and S-CN groups, with S-CI rates intermediate to those of the S-D and S-CN groups. Absolute regional CMRs of glucose were similar in the S-D and S-CI groups and in the control and S-CN groups. The regional pattern, however, showed lower right frontal regional CMRglc ratios in all stroke groups compared with the controls. There were modest correlations between performance on the Mini-Mental State Examination and whole brain CMRglc when all 4 groups were included.

Conclusions: These results demonstrate that subcortical infarction produces global cerebral hypometabolism, which is related to the clinical status of the patients. In addition, specific frontal lobe hypometabolism also appears to be a feature of subcortical infarction. Taken together, both global and regional effects on cortical function mediate the production of clinical symptoms in patients with subcortical strokes.

Figure 1

Volumes of interest from a single control subject, shown as projected onto a model of the brain surface derived from the subject’s magnetic resonance image. For clarity, medial cortical regions (hippocampus, calcarine cortex) are not shown.

Figure 2

Whole brain cerebral metabolic rates of glucose (CMRglc) in the 4 subject groups: controls without stroke, stroke–cognitively normal (S-CN), stroke–cognitively impaired (S-CI), and stroke with dementia (S-D). Bars indicate the mean and error bars represent the SEM.

Figure 3

Absolute regional cerebral metabolic rates of glucose (CMRglc) for the 4 subject groups: controls without stroke, stroke–cognitively normal (S-CN), stroke–cognitively impaired (S-CI), and stroke with dementia (S-D). Brain regions are left and right inferior parietal lobe (IPL), temporal lobe (TEMP), dorsolateral frontal cortex (DLF), orbitofrontal cortex (ORBF), hippocampus (HIP), and calcarine cortex (CALC). Error bars are SEM.

Figure 4

Regional cerebral metabolic rates of glucose (CMRglc) normalized to whole brain CMRglc (regional CMRglc ratio) for the 4 groups: controls without stroke, stroke–cognitively normal (S-CN), stroke–cognitively impaired (S-CI), and stroke with dementia (S-D); and brain regions: left and right inferior parietal lobe (IPL), temporal lobe (TEMP), dorsolateral frontal cortex (DLF), orbitofrontal cortex (ORBF), hippocampus (HIP), and calcarine cortex (CALC). Error bars are SEM.

Figure 5

Correlation between Mini-Mental State Examination (MMSE) scores and whole brain cerebral metabolic rate of glucose (CMRglc) for all 18 subjects in all 4 groups with absolute CMRglc determined.