AUTHOR=Takeuchi Hikaru , Kawashima Ryuta TITLE=Effects of Body Mass Index on Brain Structures in the Elderly: Longitudinal Analyses JOURNAL=Frontiers in Endocrinology VOLUME=13 YEAR=2022 URL=https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2022.824661 DOI=10.3389/fendo.2022.824661 ISSN=1664-2392 ABSTRACT=

The relationship between obesity and neurocognitive consequences is complex. Here we investigated associations between body mass index (BMI) and subsequent changes in brain structures, cognitive changes, and the onset of dementia after adjustment of a wide range of potential confounding variables using a large prospective cohort data of UK Biobank. After correcting for confounding factors, higher BMI was associated with greater retention in visuospatial memory performance (decline in error numbers) [beta = -0.019 (CI:-0.027~-0.016), N = 39191], increase in depression tendency scores [beta = 0.036(0.027~0.045)] as well as decreased risk of incident dementia [increasing BMI by 1 is associated with HR of 0.981 (CI:0.969~0.992), N = 398782], but not changes in fluid intelligence or reaction time. Whole brain multiple regression analyses (volumetric analyses: N = 1253, other analyses: N = 1241) revealed positive associations between BMI and subsequent changes in regional gray matter volume (rGMV) in multiple areas, regional white matter volume changes in widespread white matter (WM) tracts, fractional anisotropy changes in several tracts, and intracellular volume fraction (ICVF) and orientation dispersion (OD) in widespread areas, and isotropic volume fraction (ISOVF) in a few areas, and negative associations between BMI and subsequent changes in rGMV in the bilateral medial temporal lobe areas, mean, axial and radial diffusivity, and ISOVF in widespread areas. These results are mostly consistent with the view that less BMI precedes greater neurocognitive aging or atrophy, with a few exceptions including OD findings and the rGMV finding of the medial temporal lobes as most of significant longitudinal associations of higher BMI were opposite to those seen in higher age and dementia. Future epidemiological studies should consider separating effects of higher BMI itself from potential confounders.