POPULATION-BASED STUDY OF TAMOXIFEN THERAPY AND SUBSEQUENT OVARIAN, ENDOMETRIAL, AND BREAST CANCERS

Background: The success of tamoxifen in reducing the occurrence of con tralateral breast cancer among breast cancer patients in clinical tria ls has prompted the study of its use in the primary prevention of brea st cancer. Long-term risks associated with tamoxifen therapy, however, are still being evaluated, particularly with respect to subsequent ca ncer occurrence at sites other than the breast. Purpose: This populati on-based, nested case-control study investigated the risks of second p rimary cancers of the ovary, endometrium, and contralateral breast amo ng women receiving tamoxifen for breast cancer in conventional medical practice. Methods: A cohort of women diagnosed with breast cancer dur ing 1978 through 1990 was identified from a population-based cancer re gistry. Case subjects included all women in the cohort who subsequentl y developed second primary ovarian (n = 39), endometrial (n = 42), or contralateral breast (n = 234) cancer prior to 1992. Control subjects were a random sample of the cohort who did not develop a second primar y malignancy; they were matched to the case subjects on age, disease s tage, and year of initial breast cancer diagnosis (approximately two c ontrol subjects per case subject). Information on tamoxifen use as wel l as on potential risk factors for the second primary cancers was obta ined through medical record abstractions and physician questionnaires. Results: The percentage of women who had received tamoxifen was 18% a nd 20%, respectively, among ovarian cancer ease subjects and control s ubjects; 26% and 31%, respectively, among endometrial cancer case subj ects and control subjects; and 10% and 18%, respectively, among contra lateral breast cancer case subjects and control subjects. The mean dur ation of tamoxifen use was less than 2 years for all groups. The relat ive risks for ovarian and endometrial cancers in women who took tamoxi fen were relatively low but were consistent with no association (for o varian cancer, matched odds ratio [mOR] = 0.6 and 95% confidence inter val [CI] = 0.2-1.8; for endometrial cancer, mOR = 0.6 and 95% CI = 0.2 -1.9). Tamoxifen therapy was associated with a decreased risk of contr alateral breast cancer (mOR = 0.5; 95% CI = 0.3-0.9), especially if th e drug had been taken for more than 1 year (mOR = 0.4; 95% CI = 0.2-0. 9) or if the women were postmenopausal at initial breast cancer diagno sis (mOR = 0.4; 95% CI = 0.2-0.8). Conclusions and Implications: These data suggest that short durations of tamoxifen therapy are not associ ated with an increased risk of endometrial or ovarian cancer but are a ssociated with a reduction in contralateral breast cancer risk. It wou ld not be appropriate, however, to generalize these results to women w ho receive tamoxifen for longer periods.