A patient reported to me that her gynaecologist recommended twice-yearly infusion of 4 mg zoledronate as prophylaxis against bone metastasis. The patient was diagnosed two years ago with stage T1N1, herceptin receptor-positive and hormone receptor-negative breast cancer. Each infusion costs the patient about 500 euro. Is there any support for this procedure?

Dr. A. THONKE (Internal Medicine, Naturopathy, Sports Medicine)
D-65779 Kelkheim
Conflict of interest: none

The biphosphonates clodronate (BONEFOS, etc.), ibandronate (BONDRONAT), pamidronate (AREDIA, generics) and zoledronate (ZOMETA) have been licensed for the prevention of skeletal complications (pathological fractures, vertebral compression, etc.) in advanced neoplastic disease with involvement of the bone, as well as for the treatment of tumour-induced hypercalcaemia. In the case of clodronate and zoledronate in particular, there has been a discussion for quite some time about different cancer-inhibiting effects they may have, for example that they promote apoptosis of cancer cells or inhibit their invasion and proliferation (1). For clodronate, several studies have been performed in women with primary operable or advanced breast cancer (see a-t 1998; no. 9: 82-3), which gave conflicting results. On meta-analysis, however, they showed no significant effect for the biphosphonate either as adjuvant therapy or for metastatic breast cancer on overall survival or survival without bone metastasis or invasion of other organs (2).

For zoledronate, in addition to completed animal studies, only pilot studies with small numbers of patients with various advanced solid tumours or multiple myeloma have previously been performed. Recently, however, Austrian researchers have published a large study - the ABCSG*-12 study (1) - evaluating the use of the aromatase inhibitor anastrozole (ARIMIDEX) compared with the gold standard tamoxifen (NOLVADEX, generics) in premenopausal women with hormone receptor-positive non-metastatic breast cancer. This open-label study, performed between 1999 and 2008, concurrently tested adjuvant treatment with zoledronate using a factorial design (a-t 2008; 39: 119).

Starting 8 weeks after surgery, 1,800 women received the LH-RH agonist goserelin (ZOLADEX) 3.6 mg s.c. every 4 weeks along with either tamoxifen (20 mg/day) or anastrozole (1 mg/day), either with or without zoledronate (4 mg i.v. every 6 months**) during the study period of 3 years. According to the article in the New England Journal of Medicine (1), the primary endpoint of the study was disease-free survival, and involved a comparison of anastrozole and tamoxifen on the one hand and on the other zoledronate compared with endocrine therapy alone. The effect of the treatments on overall survival was a secondary endpoint. In addition an exploratory analysis was made of the effect of zoledronate on bone metastasis-free survival. After a mean follow-up of 4 years (data current as of 31 March, 2008) anastrozole showed no benefit over tamoxifen in either disease-free or overall survival. On the numerical results, the aromatase inhibitor performed worse in both analyses (see table). Zoledronate on the other hand increased disease-free survival significantly, but did not improve overall survival. A statistically significant effect on the rate of bone metastasis likewise was not established (1.8% vs. 2.5% under endocrine therapy alone) (1).

This seemingly positive result for biphosphonate gave rise to a situation where breast cancer guidelines - those of the Gynaecological Oncology Study Group (AGO) for example - recommended zoledronate as a therapy option in premenopausal non-metastatic breast cancer (3). A limitation, however, was noted - the adjuvant therapy used in the study did not match the gold standard: The guidelines recommend 5 years of tamoxifen therapy (4,5), whereas in the ABCSG-12 study the anti-oestrogen agent was taken for only 3 years. However, the study has a further, very striking defect which substantially limits its power.

According to information published in the U.S. study register ClinicalTrials.gov, in February 2006 (i.e. barely 7 years into the 9-year study) the primary endpoint as originally proposed was solely a comparison of anastrozole and tamoxifen in their effect upon disease-free survival and additionally upon overall survival. A substudy was to be performed for zoledronate, primarily to determine only its effect upon bone density (6). This is also described in this way in an intermediate analysis on bone density published in 2007 by the researchers (7). Possible effects of zoledronate on disease-free and overall survival were originally intended to be studied as secondary endpoints (6). In July 2008 however - more than 3 months after the end of data collection - the primary and secondary endpoints were changed in ClinicalTrials.gov (8) to match how they appear in the current article (1). The article in the New England Journal of Medicine hides the fact that the primary oncological endpoint was substantially altered in such a way that an originally negative study - judged according to the study protocol - becomes at least a partially positive study (1). The obfuscating data in the Web appendix for the study ("the study protocol is not completely precise with regard to the definition of the primary goal") are enough to indicate that the primary endpoint was subsequently altered, but even here one gets no clear information on how it was originally defined and when and why the changes were made.

  In the ABCSG-12 study the aromatase inhibitor anastrozole (ARIMIDEX) provided no benefit for disease-free and overall survival in premenopausal women with non-metastatic breast cancer compared with the gold standard tamoxifen (NOLVADEX, generics).

  Concomitant adjuvant therapy with the biphosphonate zoledronate (Zometa) tested in a factorial design significantly increases disease-free survival. An effect on overall mortality and the rate of bone metastasis was not supported.

  Subsequent alteration of the primary oncological endpoint allowed to create a partially positive study from what was a negative study without this change. Initially, only anastrozole and tamoxifen were to be compared in their effect on disease-free survival and overall survival. The primary endpoint for zoledronate was its effect upon bone density. This fact is concealed in the publication of the study; information is lacking about the time when the design was altered. The authors still describe the originally intended design in another article published a year before the end of the 9-year study.

  The ABCSG-12 study also lacks power because the endocrine therapy used does not match the gold standard: the study proposed only a 3-year regimen of tamoxifen instead of the recommended 5 years.

  The result for zoledronate therefore needs confirmation in a properly conducted, blinded study of adequate size and duration to determine its effect upon overall mortality, too.

  In the current state of knowledge we recommend that women with non-metastatic breast cancer abstain from the use of zoledronate outside clinical trials.

(R = randomized study, M = meta-analysis)

1

GNANT, M. et al.: N. Engl. J. Med. 2009; 360: 679-91

2

HA, T.C., LI, H.: Brit. J. Cancer 2007; 96: 1796-801

3

AGO: Diagnosis and treatment of patients with primary and metastatic breast cancer, Bisphosphonates, March 18, 2009; http://www.ago-online.de/download/g_mamma_09_1_0_26_bisphosphonates.pdf

4

AGO: Diagnosis and treatment of patients with primary and metastatic breast cancer, adjuvant endocrine therapy in premenopausal patients, March 18, 2009; http://www.ago-online.de/download/g_mamma_09_1_0_ 01_adj_endo_therapy_premenopausal_patients.pdf

5

Deutsche Krebsgesellschaft: Interdisziplinäre S3-Leitlinie Mammakarzinom, Stand 2008; http://www.krebsgesellschaft.de/download/interdiszipl._s3-ll_mamma_080211.pdf

6

ClinicalTrials.gov: Studie NCT00295646, February 23, 2006 http://clinicaltrials.gov/archive/NCT00295646/2006_02_23

7

GNANT, M. et al.: J. Clin. Oncol. 2007; 25 : 820-8

8

ClinicalTrials.gov: Studie NCT00295646, update July 3, 2008
http://clinicaltrials.gov/archive/NCT00295646/2008_07_03/changes

*

ABCSG = Austrian Breast and Colorectal Cancer Study Group

**

The zoledronate dose was lowered in the course of the study, so now the data on which it was based as well as the original dose vary. The current article (1) states that the regimen used was 8 mg/4 weeks. The later change was made because other studies using the new dose found that aromatase inhibitor-induced bone loss was prevented. In an intermediate analysis for bone density (administered dose: 8 mg/6 months) the reason given is "reports of renal function impairment under the 8-mg dose" (7).