Fractures in a nationwide population-based cohort of users of breast cancer hormonal therapy

Although users of aromatase inhibitors have higher total fracture risk in some randomized trials, little is known about their risk outside of clinical trials or in older higher-risk cohorts. In a population-based retrospective cohort study, we identified all older US Medicare D prescription drug insurance plan-enrolled women who had initial breast cancer surgery in 2006-2008 and began hormonal therapy (an aromatase inhibitor (AI) or tamoxifen) within the subsequent year. Total nonvertebral and hip fractures through 2012 were identified using a validated algorithm. The association of fracture outcomes with hormonal therapy type was assessed using competing risk regression models that accounted for differences in measured baseline covariates. Treatment assignment bias was reduced using inverse probability of treatment weighting computed from propensity scores. Among 23,378 women taking hormonal therapy (23.2% aged 80 or over), there were 3000 total and 436 hip fractures. Although AI users were younger and had lower comorbidity, after propensity score weighting, these and other covariates were balanced. Total nonvertebral risk was higher for users of AIs compared with tamoxifen, HR 1.11 (1.02-1.21), but the small increase in risk for hip fracture was not statistically significant, HR 1.04 (0.84-1.30). Although total nonvertebral fracture risk was higher among AI users, differences in hip fractures were not significant in a large population-based cohort of older women. Use of aromatase inhibitors by older women is associated with high risk for nonvertebral fracture that is increased compared with use of tamoxifen. Fracture risk should be assessed among patients taking these medications.