RESUMO
BACKGROUND: Type 2 diabetes mellitus (T2DM) is associated with cognitive impairment. However, most of the evidence has been based on self-reported T2DM, and undiagnosed diabetes has not been considered as a separate category. We aimed to examine the extent to which undiagnosed diabetes modifies the association between diabetes and cognitive impairment in a representative sample of Brazilian adults aged 50 years and older. METHODS: We analyzed baseline data from 1944 participants of the Brazilian Longitudinal Study of Aging (ELSI-Brazil) conducted from 2015 to 2016. Diabetes was evaluated based on self-reported doctor diagnosis and glycosylated hemoglobin levels. Participants were classified as diabetics (D), undiagnosed diabetics (UDD), or nondiabetics (ND). Cognitive function was assessed by word list learning and verbal fluency tests. Three multiple logistic regression models were used to evaluate the changes in the strength of the associations. RESULTS: Participants with diabetes had 49% greater odds of exhibiting impaired memory than nondiabetics (odds ratio [OR] = 1.49; 95% CI: 1.01-2.20). By combining UDD and ND, the association between diabetes and impaired memory was attenuated by 2.0%, losing its statistical significance (OR = 1.46; 95% CI: 0.98-2.17). By combining UDD and D, the association was attenuated by 7.4% (OR = 1.38; 95% CI: 1.01-1.90). No significant association was found between T2DM and impaired verbal fluency. CONCLUSION: This study found an association between T2DM and impaired memory but not with impaired verbal fluency. When UDD individuals are considered diabetics, this association is attenuated; when UDD individuals are considered as ND, this association is attenuated to the extent that it loses its statistical significance, affecting thus the clinical interpretation.
