Associations between distinct trajectories of body mass index (BMI) over 15 years preceding dementia and incidence of dementia

Abstract Background BMI has been reported to associate with dementia, with both increased and decreased risks. The purpose of this study was to examine longitudinal distinct trajectories of BMI over 15 years preceding dementia with incident dementia using the large national dataset from National Alzheimer’s Coordinating Center (NACC). Method We used NACC dataset from 2005 to December 2020 and performed a nested case‐control study (up to 1 to 4 match). The sample included 699 cases of all‐causes dementia with at least 5–years follow‐up time preceding dementia diagnosis, and 2120 non‐dementia controls matched on age, sex, race, education and baseline cognitive status. A group‐based trajectory model was applied to identify longitudinal distinct trajectory groups of BMI among this sample. Conditional logistic regression model was used to evaluate the effect of distinct trajectories of BMI, APOE genotype and the interaction between them with incident dementia. Result For the case group (n = 699), mean age at enrollment was 75.5 years and average education in years was 16.1, participants were more female (58.8%), non‐Hispanic white (86.8%) and healthy cognitive status at initial visit (64%). A three‐group linear trajectory model best fit BMI trajectory before the onset of dementia that featured a normal weight‐decreasing BMI group (45.5%), overweight‐decreasing BMI group (43.6%) and obese‐decreasing BMI group (10.9%). There is statistically significant linear decreasing trend over 15 years for each trajectory group (P<0.0001). The average BMI at 15 years preceding dementia for each trajectory group was 22.7, 29.9 and 36.7 respectively. Conditional logistic regression model revealed significant interaction between BMI trajectory group and APOE genotype. Comparing to obese‐decreasing and ε4 non‐carriers, the ε4 carriers in either overweight‐decreasing group or obese‐decreasing group were more likely to develop dementia [OR = 1.53, 95% CI (1.00‐2.34), P = 0.05; OR = 1.88, 95% CI (1.06‐3.34), P = 0.03]. Especially, the ε4 carriers within normal weight‐decreasing BMI group had triple the likelihood of dementia ([OR = 3.25, 95% CI (2.13‐4.94), P<0.0001]. Conclusion Our findings suggest the existence of distinct developmental trajectories of BMI preceding incident dementia and the possible effect of gene‐environment interaction on dementia. The ε4 carriers within normal weight‐decreasing BMI trajectory group exhibited the highest risk of dementia.