Get Newsletter
AlzRisk Paper Detail

Reference: Roberts, 2006
Cohort: Rochester Epidemiology Project
Risk Factor: Hormone Therapy

Average Follow-up Time Detail
The Mayo Clinic, which provides care to many Olmsted County residents, began using a system of medical record keeping in 1910, and in 1966, this system was expanded to encompass non-Mayo health care providers serving county residents. As a result, nearly all county residents are represented in some form within the REP. The study includes women who developed AD between January 1, 1985 and December 31, 1989, and matched controls of the same age ( + or - 1 year) who resided in Rochester in the year of onset of dementia in the matched case (index year).

This study and Waring et al., 1999 are from the same cohort (i.e., Rochester Epidemiology Project), but they cover different intervals (1980-84 vs 1985-89).

Exposure Detail
HRT exposure was defined as postmenopausal use of oral or
parenteral estrogen for 6 months or longer. Women who used only estrogen creams or suppositories were considered nonusers. Each episode of HRT starting from the perimenopausal period or the onset of menopause through the index year was recorded for cases and controls. Menopause was defined as the absence of menses for a period of at least 1 year. All exposure information was obtained from the medical records by a trained nurse abstractor.

Results are reported for five separate exposure categories in this paper:

1) Any v. minimal use (≥ 6 months v. < 6 months or never)
2) Duration of use (never, 0.5-3 years, > 3 years)
3) Total cumulative dose (never, ≤ 756 mg, > 756 mg)
4) Age at initiation ( never, ≤ 49.5 years, > 49.5 years)
5) Time between menopause and HRT initiation (never, ≤ 2 years, > 2 years)

The results reported here compare time between menopause and HRT initiation (≤ 2 years, > 2 years) minimal or never HRT use (< 6 months or never).

Ethnicity Detail
In 1990, 96% of the Olmsted County population were white, many of northern European origin.

Age Detail
In both cases and controls, the median age at index was 84.0 years (25th%=79.0, 75th%=88.0).

Screening and Diagnosis Detail
Screening Method:

AD Diagnosis:
DSM IV Diagnostic and Statistical Manual IV

All medical records from any care provider in the community were abstracted for information relevant to the diagnosis of dementia or AD. The hospital version of the International Classification of Diseases, Adapted (H-ICDA) was used to identify those with codes that could possibly indicate dementia (see methods in Rocca et al., 1998. Incidence of dementia and Alzherimer's disease. A reanalysis of the data from Rochester, Minnesota 1975-1984. Am J epidemiol;148:51-62). Diagnosis of AD was based on the DSM-IV criteria.

Covariates & Analysis Detail
Analysis Type:
Conditional logistic regression

AD Covariates:
TMtype of menopause

Controls were matched to cases based on age (+ or - 1 year), and resided in Rochester in the year of onset of dementia in the matched case.

Possible confounding by reproductive factors, vascular risk factors, use of NSAIDS, or education was examined using separate multivariable models with each potential confounder entered into a model with HRT and type of menopause. These covariates were not included in the models that produced the results shown here.