Prophylactic oophorectomy and hormone replacement therapy: protection at what price?

Abstract

The publication of the Women’s Health Initiative data within the last year has introduced more complexity into decisions faced by all women entering menopause who are trying to balance symptom management with increased risks of serious conditions, including breast cancer [1-4]. These decisions are perhaps more difficult for women with increased risk of breast and ovarian cancers based on mutations in the BRCA1 or BRCA2 breast and ovarian cancer susceptibility genes. Since prophylactic surgery reduces the risks of breast and ovarian cancers in these women, the question is no longer whether a carrier should undergo prophylactic bilateral salpingo-oophorectomy (BSO), but when to have surgery, whether or not to include the uterus, and whether to use hormone replacement therapy (HRT). The management of surgical menopause resulting from BSO requires strategies for mitigating frequently powerful menopausal symptoms, including hot flashes, loss of libido, potential cognitive effects, and changes in a woman’s body image and self-image. One effective remedy, HRT, may further increase breast cancer risk above the excess risk conferred by the mutation. This situation, in which complete data on which to base recommendations are not yet available, is perfect for decision analysis. In this issue of the Journal of Clinical Oncology, Armstrong et al [4a] take on the challenge of helping women and their physicians deal with this dilemma, by reporting the results of a Markov decision analysis estimating gains in life expectancy for a cohort of women who carry a germline mutation in BRCA1 or BRCA2 after BSO, with and without HRT. The model uses data from both singleinstitution and multicenter studies of mutation carriers, in which BSO resulted in approximately 90% and 50% reductions in the dominant risks of ovarian and breast cancers, respectively [5,6]. The potential impact of HRT on five major diseases, including breast and ovarian cancers, osteoporosis, coronary heart disease, and venous thromboembolism was culled from recent large randomized studies in post menopausal women [1-4]. Using these figures in their model, the authors derive estimates for very small reductions in life expectancy gains from BSO with HRT. They conclude that caregivers should recommend BSO at completion of childbearing for women who carry germline mutations in BRCA1 and BRCA2, and that neither the physician nor the patient should be dissuaded from considering HRT use for symptom management for fear of increased breast cancer risk. The model is well constructed and is based on current data. However, the fundamental question that must be asked before the conclusion of the analysis can be applied in general practice is whether the data on postmenopausal hormone use from a cohort of women at average population risk of breast cancer apply to the women in the model, who have a markedly increased breast cancer risk based on their mutations. Armstrong et al [4a] use sensitivity analysis (shown in Table 2 of their article) to address this question and simulate a maximum 2.5-fold relative risk of breast cancer in this population of women taking HRT. However, recent studies of oral contraceptives and breast cancer risk in women with a family history of the disease in first-degree relatives or founder mutations in BRCA1 and BRCA2, though not always consistent, have demonstrated estimated relative risks of 3.3 and 7.8, respectively [7,8]. Breast cancer risk from exogenous hormone use among women with BRCA1 and BRCA2 mutations, therefore, may exceed those estimated in the current model. Furthermore, this model was not constructed to address potential scenarios in which hormone use alters the penetrance function to shift more breast cancer risk to earlier ages, or, perhaps more plausibly, to negate or substantively diminish the risk-reducing benefit of the BSO, at least for the duration of the hormone use. In addition, there are reasons to believe that the effects of estrogen or progesterone on breast tissue may differ between carriers of BRCA1 versus BRCA2 mutations, someJOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 22 NUMBER 6 MARCH 15 2004