Purpose: This cross-sectional study sought to identify diabetes accurately in a study population of 3681 women age 75 and older and to determine the association of diabetes with cognitive performance.
Methods: A previously validated test, the Telephone Interview of Cognitive Screening-Modified (TICSm) was given to assess cognitive status. A diabetes case identification database, medical record review and self-report were used to determine diabetes cases. 489 (13.3%) of the women in the study were classified with diabetes and 3192 without diabetes.
Results: t-tests and linear regression analyses determined that diabetic women had a mean TICSm score 1.4 points lower (i.e. more impaired) than non-diabetic women. Using linear regression to adjust for age, education, and vascular disease, diabetic women showed a 1.1 lower score on the TICSm. Similar adjustments were made for potential confounding variables such as depression, hormone replacement therapy (HRT), high body weight, smoking, alcohol use and exercise and diabetics again showed a 1.0 lower score.
Conclusion: This study, which utilizes highly rigorous case identification methodology, provides further evidence that diabetes is associated with significantly worse cognitive performance in the elderly.