ESTROGEN REPLACEMENT THERAPY IN OLDER WOMEN - COMPARISONS BETWEEN ALZHEIMERS-DISEASE CASES AND NONDEMENTED CONTROL SUBJECTS

Objectives: We hypothesized that oral estrogen replacement therapy wou ld be less common among elderly women meeting criteria for Alzheimer's disease (AD) than among nondemented elderly women. For women with AD, we hypothesized that estrogen users would perform better on a cogniti ve task than would nonusers. Design: A case-control study of estrogen replacement therapy, in which hierarchical procedures were used to con trol for potentially confounding effects of age and education. When co gnitive performances were compared between estrogen users and nonusers with AD, the duration of dementia symptoms was an additional control variable. Setting: Alzheimer's Disease Research Center at the Universi ty of Southern California, Los Angeles. Subjects: Subjects were a volu nteer sample of consecutively enrolled elderly women, recruited primar ily from the community, who met clinical criteria for probable AD (n=1 43) or met criteria for nondemented control status (n=92). Seventy cas e patients who have subsequently died met histopathologic criteria for AD; one other demented woman who did not meet the autopsy criteria fo r AD was excluded from all analyses. Main Outcome Measures: Current us e of estrogen replacement at the time of enrollment as reported by con trol subjects or by the primary caregivers of AD case patients. Among cases, performances on a brief cognitive screening instrument were com pared between estrogen users (n=10) and nonusers (n=128) for whom this information was available. Results: Alzheimer's disease case patients were significantly less likely than control subjects to use estrogen replacement (7% vs 18%), but groups did not differ with regard to the total number of prescription medications or to the most frequently pre scribed class of drug (thyroid medication). Demented case patients usi ng estrogen did not differ significantly from those not using estrogen in terms of age, education, or symptom duration, but their mean perfo rmance on a cognitive screening instrument was significantly better (M ini-Mental State examination scores of 14.9 vs 6.5).