Abstract P1-13-02: Benefit/risk of invasive breast cancer adjuvant tamoxifen or aromatase inhibitor use by age, race/ethnicity, and co-morbidity

Background: Older age, minority race/ethnicity, and co-morbidities adversely influence early stage IBC outcome. Further, as AI compared to tamoxifen use has different effects on IBC recurrence and other health outcomes including hip fracture, endometrial cancer, stroke, pulmonary embolism and, perhaps especially, coronary heart disease (CHD), an approach to summarize relative benefit/risk of these agents is needed. Methods: Incidence rates for health outcomes by age and race/ethnicity, absent AI or tamoxifen use, were obtained from the placebo arms of the three primary Women9s Health Initiative clinical prevention trials (n = 33,072). AI and tamoxifen effects on IBC distant recurrence (relative risk [PR] 0.82) were estimated from a meta-analysis of the ATAC and Big-1-98 trial results (n = 14,149) and on other health outcomes from published meta-analysis of side effects in seven AI vs tamoxifen trials (n = 30,023); RR of 1.47 for hip fracture, 0.34 for endometrial cancer, 0.84 for stroke, 0.55 for pulmonary embolism, 1.26 for CHD (Amir et al JNCI 2011;103:1299). Following the methodology of Freedman et al (JCO 2011;29:2327), mortality weights were assigned health outcomes (5 year mortality risk of 0.2 for MI, 0.8 for IBC distant recurrence, etc) to assess net all cause mortality benefit/risk for AI compared to tamoxifen by recurrence risk, age (decade), race/ethnicity, hysterectomy (yes/no) and, in separate analyses, in women with diabetes and in women with cardiovascular disease [CVD] history. Results: In these analyses, clinical outcome of women with early stage IBC was unfavorable (more deaths per 1000 women/years) with tamoxifen compared to AI adjuvant use regardless of age, race/ethnicity, diabetes or CVD history even with 10-year distant recurrence risk of only 10%. AI superiority was substantially greater in women with a uterus and, in exploratory analyses, when assuming a CVD relative risk for AI use of 1.0, rather than a CHD relative risk of 1.26 used in all other analyses, as the AI CHDrisk elevation is controversial. The net benefit of AI compared to tamoxifen influence on clinical outcomes are described in a series of tables to be presented which quantify benefit/risk in particular women groups. The example below illustrates the excess number of deaths/1000 woman years for tamoxifen vs. AI by recurrence risk, age, and hysterectomy status. For a 60-69 year old woman with a uterus at 30% 10 year recurrence risk, there would be 382 more deaths per 1000 women/yr for tamoxifen vs. AI use. With prior hysterectomy, there would still be 287 more deaths for tamoxifen use. Conclusion: We developed an index to quantify the benefit risk for adjuvant AI vs. tamoxifen use. Even assessing a RR of 1.26 for CHD for AI use, tamoxifen compared to AI use had unfavorable outcome in all examined groups. This index can complement clinical evaluation in comparing use of these two adjuvant therapy approaches in women of different ages and racial/ethnic groups. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P1-13-02.