Background: The purpose of this study was to evaluate the cross-sectional and longitudinal association of oral estrogen replacement therapy (ERT) and cognitive function in an older, nondemented sample of women.
Methods: In a prospective cohort of 9651 white women aged 65 years and older enrolled in the Study of Osteoporotic Fractures, a modified Mini-Mental Status Exam (mMMSE), and digit symbol substitution and Trails B tests were administered twice, 4 to 6 years apart. History and current use of oral ERT was documented. Age, educational attainment, and activity limitations were the primary covariates in the analyses; in addition, stroke and depression scores were adjusted in subsets of women with available data.
Results: Current and past users of ERT had better initial scores on the mMMSE than did never users, P < .05 and .001, respectively, with better scores for current estrogen hormone users being most apparent among the older and less educated women. The percentages of women scoring < or = 23 of a possible 26 on the mMMSE were 14.3 for current users, 14.5 for past users, and 20.5 for never users, P < .001. However, only past users exhibited smaller declines upon retesting in mMMSE and Trails B performance, P < .05, than did never users. Educational attainment predicted both initial test scores and change scores and was, next to age, the most powerful predictor of cognitive function.
Conclusions: Current oral ERT does not protect against age-related declines in cognitive function in older nondemented women, whereas formal education does protect, even though it had been completed many years earlier. The influence of education in late-life on cognitive function should be tested.