Objective: We studied the relationship between the use of estrogen replacement therapy (ERT) and cerebral magnetic resonance imaging (MRI) abnormalities among older women.
Design: A population-based prospective study (Cardiovascular Health Study).
Setting: Four regions in the United States.
Participants: A total of 2133 (62.9% of the eligible) women aged 65 to 95 years (mean age 74.8), on whom MRI was performed in 1992-1994.
Measurements: Presence of global brain atrophy, white matter changes, small infarct-like lesion (ILL) (<3 mm), MRI infarcts (> or =3 mm, mostly small and asymptomatic), and cognitive function as measured by Mini-Mental State Exam (MMSE), and by ERT use (current/past/never), adjusted for a number of socioeconomic, lifestyle, and reproductive covariates.
Results: Current use of ERT was reported by 15% and past use by another 23% of participants; 35% of all women had MRI infarcts. The prevalence of MRI infarcts did not differ in current or past users from those who had never used ERT (nonusers). Bifrontal distance, the largest distance between frontal horns, and the size of ventricles were larger among current ERT users compared to past users or nonusers (P (trend) = .01), adjusted for all other covariates, but no dose-response relationship to current or past ERT use was found. Duration of estrogen use was not associated with any atrophy measure. Cortical atrophy measure, sulcal widening, or white matter disease did not differ significantly by ERT use or duration of use. Central measures of atrophy, bifrontal distance, and ventricular size were significantly associated with cognition as measured by MMSE.
Conclusions: Current ERT users had much more clinically significant central atrophy than nonusers, but the implications remained unclear.