Abstract
BACKGROUND: Patterns of cognitive change and modifiable factors for cognitive decline versus stable cognitive trajectories have rarely been described in lower-educated older adults. OBJECTIVES: We aimed to identify long-term trajectories of cognitive functioning and possible factors associated with cognitive decline. DESIGN AND PARTICIPANTS: We used data from 1,042 adults aged ≥ 60 participating in the Health, Welfare and Aging Study (SABE), São Paulo, Brazil, without cognitive impairment at baseline. Data were collected across four waves (2000-2015). Group-based trajectory modelling was used to identify cognitive trajectories. Associations with socioeconomic variables, childhood background, lifestyle, and cardiovascular risk factors were explored using weighted multinomial logistic regressions. MEASUREMENTS: The abbreviated Mini-Mental State Examination was used to measure cognition. RESULTS: Three cognitive trajectories were identified: stable (n= 754, 68.6%), mild-decline (n= 183, 20.8%), and strong-decline (n= 105, 10.7%). At baseline, respondents in the strong-decline group were more likely to be older than those with stable and mild-decline trajectories. Furthermore, participants in both the mild and strong-decline groups were more likely to have no schooling, be divorced/separated, receive less than 4 monthly wages, and be underweight (BMI < 18.5) compared to the stable group. Finally, the mild-decline group was more likely to have lived in rural areas during childhood than participants located in a stable trajectory. CONCLUSIONS: Our findings suggest that interventions to reduce cognitive decline for low-educated older adults might include strategies addressing inequalities and improving modifiable risk factor burden.