Impact of body mass index (BMI) on endocrine therapy in premenopausal breast cancer patients: An analysis of the ABCSG-12 trial

Reporter: Arpi Thukral, MD
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: June 7, 2010

Presenter: G. Pfeiler, MD, Medical University of Vienna


  • In hormone receptor positive breast cancers, drugs such as Tamoxifen and aromatase inhibitors decrease the production of estrogens, and have been shown to lead to suppression of tumor recurrence.
    • Tamoxifen is an orally active selective estrogen receptor modulator (SERM). It is an effective targeted treatment in estrogen-receptor-positive (ER +) pre- and post-menopausal breast cancer patients. The direct target of Tamoxifen is the estrogen receptor.
    • Aromatase inhibitors (AI) are a class of drugs which work by inhibiting the action of the enzyme aromatase, which converts androgens into estrogens by a process called aromatization.
  • The Austrian Breast and Colorectal Cancer Study Group Trial 12 (ABCSG-12) (Gnant, et. al. NEJM, 2009) was undertaken to examine the efficacy of gosrelin in combination with the AI anazatrazole (ANA) or Tamoxifen (Tam) +/- Zoledronic acid (ZOL) in pre-menopausal women with hormone receptor positive breast cancer.
  • Body mass index (BMI) has been shown in previous studies to be a surrogate parameter for total body aromatization.
  • It is hypothesized that overweight patients may have higher levels of aromatase enzyme availability. Therefore, BMI might have impact on the efficacy of AIs.
  • The authors of the current study hypothesized that BMI has a significant impact on efficacy of AIs, but not on Tamoxifen for the premenopausal population of the previously mentioned ABCSG-12 trial.
  • In this retrospective subset (ad-hoc) analysis of patients in the ABCSG-12, the authors investigated the influence of BMI on the effectiveness of adjuvant endocrine therapy (tamoxifen versus anastrozole in combination with goserelin) in premenopausal breast cancer patients.


  • ABCSG-12 trial methods:
    • Between 1999 and 2006, 1803 premenopausal women with Stage I or II hormone receptor positive breast cancer were randomized to Gosrelin (3.6 mg q 28 days SC) + TAM (20 mg/day PO) or Gosrelin +ANA (1 mg/day PO) +/- ZOL (4 mg IV q 6 months).
    • Essentially, the trial had four arms comparing goserelin/tamoxifen to goserelin/ anastrazole with or without zoledronic acid, and followed a 2x2 factorial design.
    • Eligibility criteria included Stage I or II breast cancer, <10 + nodes, and no chemotherapy except neoadjuvant chemotherapy.
    • The treatment duration was 3 years and patients were then followed for 2 additional years.
  • 1684 patients of the 1803 patients in the trial were included in this current analysis (Tam: N=900, ANA: N=903).
  • BMI was calculated using height and weight data on study entry from the ABCSG-12 trial
    • Standard WHO BMI definitions were used: normal 18.5-24.9 kg/m2; overweight: BMI ? 25 kg/m2.
  • Patients in both the Tam and ANA groups were further separated into 2 categories for analysis: normal weight and overweight.
  • Primary endpoints were disease-free survival (DFS) and overall survival (OS), and these were calculated by Kaplan-Meier method.
  • Results between the groups were compared using the log-rank test and Cox proportional hazard modeling.


  • Patient characteristics were well-balanced.
  • 1684 patients were included in the current analysis: 1,111 (66.0%) patients were normal weight and 573 (34.0%) were overweight.
  • Median follow-up = 5 years
  • Overweight patients treated with ANA showed significantly shorter DFS (HR 1.60, p = 0.023) translating to a 60% increased risk of disease progression.
  • In terms of OS, the ANA overweight patient group had a lower OS (HR 2.14; p = 0.008) compared to normal weight ANA patients, translating to a doubling for risk of death.
  • For the Tam group, no differences in DFS or OS were observed between overweight and normal weight.
  • A multivariate cox regression analysis was performed and results were adjusted for tumor stage, nodal stage, grade, ER/PR status, and age, and results remained the same.
  • Predictive impact of BMI:
    • Overweight patients: ANA vs. Tam
      • Superiority of Tam for DFS and OS seen; HR=1.49 for DFS and 3.03 for OS (p=0.004)
    • Normal weight patients: ANA vs. Tam
      • No difference in DFS or OS. No difference in number of events.
  • Adverse events: no difference between the 2 groups

Authors' Conclusions

  • The authors concluded that BMI significantly influences efficacy of ANA plus goserelin, but not Tam plus goserelin in the treatment of pre-menopausal breast cancer patients.
  • Overweight patients treated with ANA had a nearly 50% increase in disease progression and a tripling of rate of death that normal weight patients treated with ANA.
    • These findings may be due to an increase in total body aromatization in overweight patients.
  • The authors note that as approximately one-third of these patients were overweight with BMI ? 25 kg/m2, the impact of this study may be clinically relevant.

Clinical Implications

  • The results of this study are intriguing, however we must keep in mind that this study was an unplanned ad-hoc analysis of a larger randomized trial.
  • Also, the results of this study should be kept in proper context since other factors associated with obesity can influence prognosis in breast cancer patients.
  • Obesity is a complex process that can lead to increased estrogen, increased insulin, and increased cytokines and inflammatory markers. All of these can lead to increased breast cancer recurrence rate.
    • Although AIs reduce estrogen receptor signaling, the receptor can still be stimulated through cross-talk with the IGFR pathway which is related to obesity.
  • Questions to consider:
    • Could AIs in this group of premenopausal patients be ineffective since patients were being made “post-menopasual” with Gosrelin? Should estradiol levels have been evaluated to measure the effect of AIs on the target? Could this be a confounding factor limiting the effectiveness of AIs?
  • This study is not yet a practice changing trial, but it opens the door for studying obesity and its relationship to the effectiveness of current therapies in breast cancer. Confirmatory data from other trials demonstrating similar effects is needed.