Late-life body mass index and amyloid interaction on cognitive decline in unimpaired older adults

• We examined the joint associations of BMI and amyloid with cognition in a cohort enriched for preclinical Alzheimer’s disease (AD). • Late-life obesity and amyloid pathology were independently associated with poorer baseline cognition. • Longitudinally, the BMI–cognition relationship was modified by amyloid burden. • Lower or normal BMI was associated with faster cognitive decline only when amyloid burden was high, independent of weight loss. • Future AD prevention trials with obesity interventions may benefit from preferentially enrolling younger individuals or those without substantial amyloid accumulation. The late-life “obesity paradox” of reduced Alzheimer’s disease (AD) risk is postulated to be driven by underlying preclinical/prodromal pathology. However, few studies have directly examined the joint associations of BMI and amyloid pathology with cognitive decline, especially in individuals with preclinical AD targeted in prevention trials. To determine whether late-life BMI and amyloid pathology have independent or interactive associations with cognition in clinically unimpaired older adults. Secondary analyses of A4 randomized clinical trial and the companion observational LEARN Study (median follow-up 4.7 years). Multicenter across 67 sites in US, Canada, Australia, and Japan. We included 1663 participants (Placebo n = 582, Solanezumab n = 563, LEARN n = 518) who were baseline cognitively unimpaired and medically stable, mean age 71.5 ± 4.7 years, 60% women. BMI and global amyloid burden Florbetapir PET were measured at baseline. Cognition was measured longitudinally using Preclinical Alzheimer Cognitive Composite. Higher BMI and amyloid burden were independently associated with worse baseline cognition. Longitudinally, a BMI*Amyloid*Time interaction emerged: lower/normal BMI was associated with more favorable cognitive trajectory at low amyloid levels, but with faster cognitive decline when amyloid was substantially elevated. Our cross-sectional findings support a negative association between obesity and cognitive aging up to late-life. Longitudinally, we observed an “obesity paradox”, where higher/obese BMI was associated with more favorable cognitive trajectories in the presence of advanced amyloid pathology. Together, our findings suggest that future trials targeting obesity to slow late-life cognitive decline may benefit from preferentially enrolling younger individuals or those without substantial amyloid accumulation.