Weight loss in the healthy elderly might be a non-cognitive sign of preclinical Alzheimer's disease

Abstract

Weight loss has been proposed as a sign of pre-clinical Alzheimer Disease (AD). To test this hypothesis, we have evaluated the association between longitudinal changes in weight trajectories, cognitive performance, AD biomarker profiles and brain structure in 363 healthy controls from the Alzheimer´s Disease Neuroimaging Initiative (mean follow-up 50.5±30.5 months). Subjects were classified according to body weight trajectory into a weight loss group (WLG; relative weight loss ≥ 5%) and a non-weight loss group (non-WLG; relative weight loss < 5%). Linear mixed effects models were used to estimate the effect of body weight changes on ADAS-Cognitive score across time. Baseline CSF tau/AΔ₄₂ ratio and AV45 PET uptake were compared between WLG and non-WLG by analysis of covariance. Atrophy maps were compared between groups at baseline and longitudinally at a 2-year follow-up using Freesurfer. WLG showed increased baseline levels of cerebrospinal fluid tau/AΔ₄₂ ratio, increased PET amyloid uptake and diminished cortical thickness at baseline. WLG also showed faster cognitive decline and faster longitudinal atrophy. Our data support weight loss as a non-cognitive manifestation of pre-clinical AD.

Keywords: Gerotarget; PET amyloid; cerebrospinal fluid Alzheimer’s disease biomarkers; magnetic resonance imaging; pre-clinical Alzheimer’s disease; weight loss.

Figure 1. Flowchart showing the samples used in this work and the subsets utilized for analyses

Figure 2. Observed means (SEM) differences in Alzheimer's Disease Scale Assessment

Scores according to body weight trajectory groups. In blue is represented the ADAS-Cog trajectory for the non-WLG throughout time and in red is represented the ADAS-Cog trajectory for the WLG throughout time. Dots represent mean of each time-point for both groups and the error bars show the standard error of each time-point for both groups. In the lower panel it is shown the number of observation per year.

Figure 3. Group comparison of the structural changes between the weight loss group (WLG) and non-weight loss group (non-WLG)

A1. Baseline differences in cortical thickness between the weight loss group (WLG) and the non-weight loss group (non- WLG). Blue areas represent significant cortical thinning in the WLG respect to the non-WLG (FWE < 0.05). A2. Box and whisker plots showing the cortical thickness for each group at the maximum significant vertex in the right temporal cluster. B1. Differences in the symmetrized percent change (spc) over the 2-year follow-up between WLG and non-WLG. Blue indicates significant increased cortical atrophy (i.e., less spc) in the WLG (FWE < 0.05). B2. Box and whisker plots showing the symmetrized percent change for each group at the maximum significant vertex in the right frontal cluster. The colors in the box-plots are only for illustrative purposes.