Tamoxifen is an anti-estrogen used in the treatment of breast cancer and to reduce the incidence of breast cancer in high risk women. Although the brain is an estrogen target organ and several studies have found a beneficial effect of estrogen on cognitive function, the effect of tamoxifen on cognition has not been reported. Therefore, we initiated a follow-up study of women who had participated in a study of breast cancer to assess the effect of tamoxifen treatment on cognitive function. We recruited previously interviewed patients who were cases in a population-based case-control study of 2,653 women with primary breast cancer diagnosed between 1987 and 1996 at ages 55-72 years in Los Angeles County, California, USA. In November 1997, each case was mailed a follow-up questionnaire. Cognitive function was assessed by (1) clock drawing. (2) copying a box drawing, and (3) narrative writing to describe a pictured scene. Women reporting treatment with tamoxifen were categorized as standard-term users (4-5 years), short-term users (< 4 years) or long-term users (6 + years) and compared to never users. Tamoxifen users were also classified as past or current users. Differences in the mean cognitive test scores were tested after adjusting for age, age at diagnosis, stage of disease, radiation therapy, chemotherapy, race, education, marital status, previous use of oral contraceptives, type of menopause, age at last menstrual period, previous use of hormone replacement therapy, and depressive symptoms using analysis of covariance. All p-values for differences in the proportion of women who had errors on the tests are 2-sided and adjusted for age, stage of disease at diagnosis, and chemotherapy. Information from 1,163 women aged 57-75 years of age was analyzed; 710 had taken tamoxifen. There was little difference between women who had used tamoxifen for the standard five years and never users on the three cognitive tests. However, more women who had used tamoxifen for the standard term reported seeing their physician for memory problems than non-users (3.8% vs 1.5%, p = 0.04). This was especially true for current users of standard-term (8.0%, p = 0.003). Current users also had a significantly lower mean complexity score (p = 0.03) on the narrative writing task. No differences were seen between past users and non-users. Our study suggests that current use of tamoxifen may adversely effect cognition. Further study of tamoxifen and cognition is needed so that healthy women considering tamoxifen for the primary prevention of breast cancer have comprehensive information about the side effects of the treatment.