Racial Differences in BMI and Cognitive Performance in Persons Aged 80 and Older

Abstract Background There is evidence that late‐life BMI is protective against cognitive decline. As Alzheimer’s disease (AD) study samples do not reflect the general population, it is not clear if there are racial differences this finding. We investigated late‐life BMI and cognitive performance in a biracial sample. Method Using ANOVA or chi‐square, demographic, physical, and cognitive test scores were compared in persons age 80 and older by race. Cognitive groups were based on global CDR. Normal cognition (NC) was defined per global CDR = 0 and impaired cognition per global CDR = 0.5,1,2, or 3. Mixed models adjusted for age, sex, and education with BMI as the predictor evaluated cognitive test scores by race for study visits from 2005 to 2020. The significance level for all tests was p < 0.05. Result Data for 9,581 persons was analyzed. The sample was mostly white (86.6% vs. 13.41%, <.0001). AA were younger (86.4 vs. 86.9, <.0001), had higher BMI (26.8 vs. 25.5, <.0001), fewer years of education (13.4 years vs. 15.3 years, <.0001), and lower MOCA scores (18 vs 21, <.0001) than whites. In AA, mixed models showed associations between higher BMI and executive function (Trial Making Test‐B:β = ‐1.51,p<.01) and episodic memory (Craft delayed: β = .07,p = .01). In whites, there was an association between higher BMI and picture naming (Multilingual Naming: β = .09,p = <.01), attention (Trail Making Test‐A:β = ‐.12, p = 0.3), and working memory (Number span, backward: β = .03,p = .01). Conclusion Late‐life BMI is associated with improved cognitive domain performance that vary by race. Further study is needed to prospectively replicate the findings in multiethnic cohort and explore factors underlying the differences.