Abstract
1050
Objectives: Fall-related injuries represent a leading cause for hospitalization, disability, and death in elderly patients, and those with amnestic mild cognitive impairment (MCI) have an even higher incidence of falling. In this study, we employed FDG-PET in examining the ability of alterations in regional cerebral metabolism to distinguish MCI patients at heightened risk for suffering subsequent falls.
Methods: A total of 228 FDG-PET brain scans were analyzed, prospectively acquired from 114 MCI subjects enrolled in the Alzheimer’s Disease Neuroimaging Initiative at over 50 sites in U.S. and Canada. Each subject underwent clinical evaluation and neuroimaging studies at 6 to 12 month intervals, over a period of years; 61 were documented to suffer a fall incident subsequent to the time of their baseline PET scan (MCI-F), while 53 age-matched subjects had no subsequent fall documented during long-term follow-up (MCI-N). For MCI-F subjects, FDG-PET scans acquired at baseline and the last time point preceding date of first falling were included; for MCI-N subjects, FDG-PET scans acquired at baseline and the last time available were analyzed. Standardized volume of interest (sVOI) methods were used to assess magnitude of regional metabolism as normalized to the metabolism of each subject’s pons, in 11 non-cortical and 36 cortical volumes defined in each scan. Statistical parametric mapping (spm) methods were used to assess significance of interval changes on a voxel-by-voxel basis by paired t-test analyses, and voxel clusters undergoing the most significant changes were identified after Family Wise Error (FWE)-based correction for multiple comparisons. Rates of longitudinal change were calculated as the interval difference in magnitude of metabolism in each sVOI divided by the number of months intervening between scans, for each subject.
Results: Among the 47 sVOI’s examined, thalamus underwent greatest normalized decline, with respect to both magnitude and significance, in MCI-F and MCI-N groups, consistent with general diminishment in metabolism of forebrain relative to brainstem over time. After thalamus, in MCI-F subjects cerebellar vermis was the sVOI demonstrating the most significant decline, yet was one of the two most stable brain regions in MCI-N subjects, and mean rate of metabolic decline of the vermis was 4.6 times faster in MCI-F than in MCI-N subjects. Among cortical regions, bilateral parietotemporal and bilateral posterior cingulate cortex declined the most significantly in MCI-N, corresponding to areas of expected decline in populations enriched for having pre-dementia due to Alzheimer’s disease (AD). In contrast, the four sVOI’s demonstrating most significant decline in MCI-F subjects included, in addition to left parietotemporal and right posterior cingulate cortex, right associative visual and right primary visual cortical areas. In MCI-F subjects, spm analysis identified significant decline in the precuneus area (PFWE<0.0005), while age-matched MCI-N subjects experienced significant decline in posterior and anterior cingulate metabolism (PFWE=0.01; PFWE=0.04, respectively).
Conclusions: In MCI patients, accelerated rates of decline in vermis and occipital metabolism, as well as significantly declining precuneus metabolism -- as identified through interval regional changes in FDG-PET data -- were associated with suffering subsequent fall events.