Associations of metabolic heterogeneity of obesity with the risk of dementia in middle-aged adults: three prospective studies

The associations of different obesity and metabolic phenotypes during midlife with the risk of incident dementia remain unclear. This study aimed to investigate the associations between metabolic heterogeneity of obesity and long-term risk of dementia. We conducted prospective analyses from three cohorts, including the UK Biobank (UKB), Atherosclerosis Risk in Communities (ARIC) study, and Framingham Offspring Study (FOS). Eligible participants were those aged 45-65 years with valid assessments of body mass index (BMI) and metabolic status at the study baseline. Obesity was defined as a BMI of ≥ 30.0 kg/m , while metabolic abnormality was defined as meeting ≥ 2 of the National Cholesterol Education Program-Adult Treatment Panel III (NCEP-ATP III) criteria. Metabolic heterogeneity of obesity was evaluated based on obesity and metabolic phenotypes and grouped as metabolically normal non-obesity (MNNO), metabolically abnormal non-obesity (MANO), metabolically normal obesity (MNO), and metabolically abnormal obesity (MAO). Included in this study were 295,823 participants aged 56.3 ± 5.9 years from the UKB, 12,547 participants aged 54.0 ± 5.7 years from the ARIC, and 2,004 participants aged 53.9 ± 5.9 years from the FOS. Over 4,348,208 person-years, a total of 6,190 participants (3,601 in the UKB, 2,405 in the ARIC, and 184 in the FOS) developed incident dementia. In the pooled analysis of three cohorts, metabolic abnormality was associated with a hazard ratio (HR) of 1.41 (95% confidence interval [CI]: 1.10-1.80) for dementia, while obesity was associated with an HR of 1.20 (1.03-1.41). Compared with MNNO, individuals with MANO and MAO had increased risks of dementia (pooled HR: 1.33, 95% CI: 1.04-1.71 for MANO and 1.48, 1.16-1.89 for MAO). However, there was no significant difference in the risk of dementia among MNO (pooled HR: 1.10, 95% CI: 0.98-1.24). In addition, participants who recovered from MANO to MNNO had a lower risk of dementia (pooled HR: 0.79, 95% CI: 0.64-0.97), as compared with stable MANO. Metabolic abnormality has a stronger association with dementia than obesity. Metabolically abnormal non-obesity and obesity, but not metabolically normal obesity, are associated with higher risks of incident dementia as compared with metabolically normal non-obesity. Recovering from an abnormal metabolic status to normal reduces the risk of dementia in populations without obesity. Our findings highlight the important role of metabolic status in the developm