Weight Loss is a Preclinical Signal of Cerebral Amyloid Deposition and Could Predict Cognitive Impairment in Elderly Adults

Background: Higher late-life body mass index (BMI) was associated with reduced risk of Alzheimer's disease (AD), which might be explained by a reverse causal relationship. Objective: To investigate whether weight loss was a preclinical manifestation of AD pathologies and could be a predictor of cognitive impairment. Methods: A total of 1,194 participants (mean age = 73.2 [range: 54 to 91] years, female = 44.5%) from the Alzheimer's Disease Neuroimaging Initiative (ADNI) were grouped according to AD biomarker profile as indicated by amyloid (A) and tau (TN) status and clinical stage by clinical dementia rating (CDR). BMI across the biomarker-defined clinical stages was compared with Bonferroni correction. Pearson correlation analysis was performed to test the relationship between the amyloid change by PET and the BMI change. Multiple regression models were used to explore the influences of amyloid pathologies on BMI change as well as the effects of weight loss on longitudinal changes of global cognitive function. Results: BMI was significantly decreased in AD preclinical stage (amyloid positive [A+] and CDR = 0) and dementia stage (A+/TN+ and CDR = 0.5 or 1), compared with the healthy controls (A-/TN- and CDR = 0, p < 0.005), while no significant differences were observed between preclinical AD and AD dementia. Amyloid PET change was inversely correlated with BMI change (p = 0.023, 0 = -14). Individuals in amyloid positive group exhibited faster weight loss (time x group interaction p = 0.019, 0 = -0.20) compared to the amyloid negative group. Greater weight loss predicted higher risk of developing cognitive disorders. Conclusion: Elders who experienced greater weight loss might belong to preclinical stage of AD and could be targeted for primary prevention of the disease.