Abstract S2-6: Aromatase Inhibitors and Risk of Myocardial Infarction, Stroke and Fracture

Abstract Randomized trials have demonstrated that aromatase inhibitors (AI's) are more effective than tamoxifen in preventing breast cancer recurrence. Since many women with early stage disease have a modest risk of recurrence, it is important to consider carefully the potential side effects of treatments in this setting. Trials have suggested increased risk of fracture is the only serious side effect of AI's, but study participants are often healthier than the general population, and little is known about toxicities of AI's in nontrial populations. We sought to evaluate the association between use of AI's and the incidence of myocardial infarction (MI), stroke, and fracture in community-based populations. Methods: Using encounter and pharmacy data from WellPoint plans in the HealthCore Integrated Research Database, 44,463 women aged ≥50 years with at least 2 diagnosis codes for breast cancer between 2001 and 2007 were identified and followed through the end of 2008. We used propensity score methods to identify matched controls from a cohort of 88,101 women aged ≥50 years enrolled in WellPoint plans who had no diagnosis codes for breast cancer. We identified receipt of aromatase inhibitors and tamoxifen and used Cox proportional hazards models with time-varying treatment variables to assess whether treatment with aromatase inhibitors or tamoxifen was associated with MI, stroke, and fracture among women with breast cancer and to assess the association of breast cancer with the outcomes of interest. Results: Among 44,049 breast cancer patients and 44,049 matched controls, the average age was 67 and 87% of cases and controls had no baseline comorbidities. After a median follow-up of 899 days for cases and 1004 for controls, 1.3% of women had an MI, 2.9% an ischemic stroke, 1.5% a hip fracture and 9.5% any fracture. Breast cancer patients treated with AI's had a significantly lower adjusted hazard of ischemic stroke and a significantly higher adjusted hazard of fracture compared with breast cancer patients not receiving an AI or tamoxifen. Table. Adjusted hazard ratio (95% CI) associated with breast cancer and hormonal therapy Treatment with AI or tamoxifen was not associated with MI or any fracture. Patients treated with tamoxifen had a significantly lower adjusted hazard of stroke and hip fracture compared with breast cancer patients not receiving hormonal therapy. Conclusion: The side effect profile of AI's in this community-based population was similar to that seen in clinical trials and suggests that these drugs are associated with few serious side effects. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr S2-6.