Résumé
BACKGROUND:Mutations in BRCA1/2 confer a high risk of breast cancer, but literature values of this risk vary. A genotype-phenotype correlation has been found in both genes, and the effect of reproductive factors differs according to mutation location. Therefore, we hypothesize that such a variation may exist for other factors related to estrogen exposure.METHODS:We used a weighted Cox regression model to assess variation in breast cancer risk with these factors using location of mutation in homogeneous breast cancer risk region of BRCA1/2 in the GENEPSO study.RESULTS:We found that late age at menarche reduced breast cancer risk by 31% and that among BRCA1 carriers, a long or a short menstrual cycle increased risk (by 65% and 73%, respectively). Among premenopausal women, overweight was associated with a 45% decrease in risk whereas underweight was associated with an increased risk (HR, 2.40). A natural menopause, mainly after age 50, was associated with a high breast cancer risk (HR, 2.46), and a significant interaction between menopause status and the location of mutations was found leading up to 10% variation in absolute risk according to the age at menopause.CONCLUSIONS:As observed in the general population, a late menarche, a long or a short menstrual cycle, over- or underweight, and being postmenopausal were associated with breast cancer risk in BRCA1/2 carriers. The association with the menopause was observed only when the mutation was located in the "high-risk" zones.IMPACT:Taking into account modifier factors, location of mutation might be important for the clinical management of BRCA1/2 mutation carriers.