Preclinical studies show that bisphosphonates may have antitumor activity in addition to their ability to reduce osteoclast-mediated bone resorption.1 These effects include prevention of tumor cell adhesion to bone and consequently prevention of bone metastasis,2–4 induction of tumor cell apoptosis,5 antagonism of growth factors,6 and antiangiogenic effects.7 Bisphosphonates have also shown synergistic antitumor effects in combination with chemotherapy in breast cancer models.8
Many randomized controlled trials have evaluated the role of bisphosphonates in the adjuvant setting of breast cancer and have shown a beneficial effect on bone loss prevention.9 However, the overall clinical impact of bisphosphonates on the disease status of patients with early breast cancer is still unclear and controversial.10–12
Considering the notable accumulation of randomized trials, the authors performed a meta-analysis of these to address whether the use of bisphosphonates in the adjuvant setting of breast cancer might have any effect on the natural course of the disease. Most specifically, they investigated for any beneficial effects on overall survival, prevention of disease recurrences, and occurrence of bone metastases.
The investigators searched PubMed, Cochrane Central Register of Controlled Trials, and ISI Web of Knowledge and collected all prospective randomised trials published as formal papers in peer-reviewed journals or as abstracts in the proceedings of major international congresses (ASCO Annual Meeting, San Antonio Breast Cancer Symposium, the European Cancer Conference) up to January 2009. No language or year restrictions were set. The following searching algorithm was used: [early OR adjuvant] AND [breast OR mammary] AND [tumour OR malign* OR carcinom* OR cancer] AND [biphosphonates OR bisphosphonates OR clodronate OR pamidronate OR zoledronic acid OR ibandronate]. The reference lists of all studies included in the meta-analysis were examined for other relevant articles missed by the electronic searches.
Eligibility and exclusion criteria were prespecified. The analysis included all controlled trials that randomized patients with primary breast cancer to undergo adjuvant treatment with any or no bisphosphonate treatment. Trials fulfilling these criteria were considered eligible irrespective of the study sample size and the type and dosage of bisphosphonate used. Nonrandomized studies were considered ineligible.
When multiple records were related to the same study, end point data was extracted from the report with the longest follow-up (largest number of events) to avoid duplication of information in the meta-analysis calculations.
Information recorded from each eligible trial included author names; journal; year of publication; number of centers involved; number of patients randomly assigned and analyzed per arm; years of patient enrollment; patient age; country of origin; tumor stage; menopausal status; hormonal receptor status; the exact regimens used; and dose and schedule and any additional treatments given to both arms. Study design items such as randomization mode, allocation concealment, and withdrawal descriptions were further recorded.
Two investigators extracted the relevant data, and consensus was reached on all outcomes. Authors of the original trials were contacted to identify any missing information for outcomes.
Primary outcome of this study was to evaluate whether the adjuvant use of bisphosphonates in breast cancer might have any effect on overall survival, disease recurrences, and occurrence of bone metastases compared with nonuse.
Furthermore, the investigators pooled estimates for distant metastases, visceral recurrences, and occurrence rate of locoregional relapses. Considering that treatment outcomes may vary among different types of bisphosphonates, the investigators performed subgroup analyses for deaths, disease recurrences, and bone metastases according to the bisphosphonate used (zoledronic acid/clodronate/pamidronate/risedronate).
The number of events (deaths, disease recurrences, and bone metastases) and the number of nonevents in treated and control groups were retrieved from each primary study and 2x2 tables were constructed. Odds ratios (OR) of events for treated patients with respect to those who were not and the 95% CIs were calculated.
X2 test was used to assess heterogeneity between studies (significance level set at 0.1). In the absence of heterogeneity, pooled estimates of ORs with their 95% CIs were calculated using the Mantel-Haenszel method. In the presence of heterogeneity, the DerSimonian and Laird random effects method was used to pool primary studies estimates. Statistical software STATA 8.0 (Stata Corporation, College Station, Texas) was used for statistical analysis.
Literature search identified 21 potentially eligible trials10–30 evaluating the adjuvant use of bisphosphonates compared with no use. Most of the trials were designed to analyze safety data and the impact of bisphosphonates on bone loss;14–27,29 data suitable for our analyses could be retrieved from only 13 studies.10–12,21–30
A flow chart indicating the identification of randomized controlled trials for inclusion in the meta-analysis is reported in Figure 1.
Table 1 presents the characteristics of the 13 trials that met the eligibility criteria for this study; 6 studies used the bisphosphonate zoledronic acid,11,21,22,25–27 4 trials used clodronate,12,28–30 2 trials used pamidronate,10,23 and 1 used risedronate.24
Among these trials, 3 were double-blinded,23,24,30 3 described in detail the mode of randomization,23,25,30 2 reported allocation concealment,25,30 and 9 trials reported withdrawal description.10–12,21,22,24,25,28,29 An intent-to-treat analysis was performed in all but 1 trial.29
Meta-analysis included data on 511 deaths from 9 trials (5736 patients),11,12,21–24,27,28,30 675 disease recurrences from 11 trials (5631 patients),11,12,21–28,30 and 545 bone metastases from 8 trials (5571 patients).10–12,21,23,27,28,30 Although the primary investigators were contacted, additional data could not be retrieved.
Pooled results showed no statistical significant differences with the use of bisphosphonates in early breast cancer versus nonuse for the overall number of deaths (summary OR, 0.708; 95% CI, 0.482–1.041; P = .079), disease recurrences (summary OR, 0.843; 95% CI, 0.602–1.181; P = .321), and bone metastases (summary OR, 0.925; 95% CI, 0.768–1.114; P = .413; Figure 2). A statistically significant heterogeneity was seen among trials in estimates for deaths and disease recurrences (P = .034 and P = .016 for heterogeneity, for deaths and disease recurrences, respectively). No statistically significant differences between study heterogeneity was observed for bone metastases.
Characteristics of Eligible Trials
Adjuvant treatment with bisphosphonates compared with no use was not associated with any statistically significant differences between arms for type of recurrences, including distant metastases (7 trials, 4618 patients; OR, 0.896; 95% CI, 0.674–1.192; P = .453),11,12,21,23,27,28,30 visceral recurrences (4 trials, 1693 patients; OR, 1.051; 95% CI, 0.686–1.609; P = .820),12,23,28,30 and local relapses (5 trials, 4276 patients; OR, 1.056; 95% CI, 0.750–1.487; P = .756).11,21,27,28,30 No significant between-study heterogeneity was observed.
Subgroup analyses for disease recurrences according to the type of bisphosphonate used showed a statistically significant lower risk for disease recurrences with zoledronic acid (6 trials, OR, 0.675; 95% CI, 0.479–0.952; P = .025).11,21,22,25–27 Use of zoledronic acid was not associated with any significant difference in death rate (OR, 0.642; 95% CI, 0.388–1.063) or bone metastasis rate (OR, 0.661; 95% CI, 0.379–1.151; Table 2).
This meta-analysis showed no significant differences in terms of overall survival, overall disease recurrence, distant and visceral relapses, local relapses, and bone metastases rates for the use or nonuse of bisphosphonates in the adjuvant treatment of early breast cancer. Therefore, the currently available randomized evidence does not support the hypothesis that use of adjuvant bisphosphonates may alter the natural course of breast cancer. Nonetheless, because a positive but nonsignificant trend was seen favoring bisphosphonates use for reduction in death rate (OR, 0.708; 95% CI, 0.482–1.041), further randomized studies are needed before the use of bisphosphonates in the adjuvant setting of breast cancer can be definitively supported or discouraged. Results of the ongoing phase III studies (NSABP B-34, AZURE, and SWOG 0307) will likely be crucial for better defining these issues in the next future.
Subgroup Meta-Analyses by Type of Bisphosphonates: Estimates of Effect in 23 Comparisons
Of interest, when different bisphosphonates agents were considered, the investigators noted that the use of zoledronic acid may be of benefit in decreasing the number of disease recurrences. However, these results should be interpreted cautiously because the statistical significance for this association was not adequately strong (OR, 0.675; 95% CI, 0.479–0.952; P = .025) and may be largely attributed to chance because of multitest analyses (1/21 analyses = 4.7%). Moreover, the observed decrease in the number of disease recurrences was not associated with a decrease in the occurrence of bone metastases. Conversely, the trend for overall survival improvement (OR, 0.642; 95% CI, 0.388–1.063; P = .085) noted in the zoledronic acid arm was far from reaching statistical significance. Studies with longer follow-up are needed before firm conclusions for survival setting can be made.
Overall, this meta-analysis is consistent with an earlier one by Ha and Li31 that examined the potential benefit of adjuvant oral clodronate in patients with early breast cancer. However, data from this early meta-analysis were feeble and unlikely to detect small differences between the arms because the study was relatively underpowered (only 3 trials were included involving 1653 patients and reflecting low number of events).31 Another limitation of the meta-analysis by Ha and Li31 was that the results could not be extrapolated to bisphosphonates other than clodronate.31
Therefore, this more recent meta-analysis is the first to present solid data on the use of bisphosphonates in the adjuvant treatment of early breast cancer.
Several limitations must be considered when interpreting these results. First, this meta-analysis is based on published data and a meta-analysis of individual level data might define more clearly treatment benefits.32 Second, many trials identified were reported in abstract form only, making complete data difficult to extract for analyses. Third, small time-dependent differences may have been lost in these analyses because most of the included studies did not report hazard ratios for overall and disease-free survival; therefore, this meta-analysis was mainly based on OR statistics.33 Finally, 7 studies were excluded from survival analysis because no survival data could be retrieved. Furthermore, all studies but one10 either were small or had short follow-up, and were therefore unlikely to invalidate the conclusions. Likewise, the large study by Kristensen et al.10 was unlikely to change the results because no overall survival differences were reported between the arms.
Allowing for these caveats, this meta-analysis showed that currently available randomized evidence does not support the hypothesis that using bisphosphonates in adjuvant treatment of early breast cancer likely alters the natural course of the disease, because it does not affect overall survival or disease relapse and does not prevent bone metastases. Nonetheless, patients undergoing bisphosphonate treatment seem to show a nonsignificant trend toward better outcomes. Until further evidence from new clinical trials becomes available, adjuvant bisphosphonates should not be routinely recommended as agents that may potentially alter the course of breast cancer when administered in the adjuvant setting.
Antonis Valachis, Davide Mauri, Nikolaos P. Polyzos, Lamprini Tsali, Dimitris Mavroudis, and Vassilis Georgoulias participated in the design of the study. Antonis Valachis and Davide Mauri participated in data extraction. Giovanni Casazza performed the statistical analyses. Giovanni Casazza, Davide Mauri, Lamprini Tsali, and Nikolaos P. Polyzos interpreted the results. Davide Mauri, Antonis Valachis, Nikolaos P. Polyzos, Dimitris Mavroudis, and Vassilis Georgoulias wrote the manuscript. Lamprini Tsali reviewed and commented on the final manuscript.
All authors had full access to all data in the study and take responsibility for the integrity of the data and accuracy of the data analysis.
The authors have disclosed that they have no financial interests, arrangements, or affiliations with the manufacturers of any products discussed in the article or their competitors.
ColemanRThorpeHCameronD. Zoledronic acid is well tolerated and can be safely administered with adjuvant chemotherapy—first safety data from the AZURE trial (BIG01/04) [poster]. Presented at the 29th Annual San Antonio Breast Cancer Symposium; December 14–17 2006; San Antonio Texas. Abstract 2080.
MinceyBADentchevTSloanJA. N03CC—a randomized, controlled, open-label trial of upfront vs. delayed zoledronic acid for prevention of bone loss in postmenopausal (PM) women with primary breast cancer (PBC) starting letrozole after tamoxifen [abstract]. J Clin Oncol2008;26(Suppl):Abstract 564.
Van PoznakCHannonRClackG. Managing cancer treatment-induced bone loss: 24-month results from the Study of Anastrozole with the Bisphosphonate RisedronatE (SABRE). Presented at the San Antonio Breast Cancer Symposium; December 10–14 2008; San Antonio Texas. Abstract 1137.
ShapiroCLHalabiSGibsonG. Effect of zoledronic acid (ZA) on bone mineral density (BMD) in premenopausal women who developed ovarian failure (OF) due to adjuvant chemotherapy (AdC): first results of GALGB trial 79809 [abstract]. J Clin Oncol2008;26(Suppl):Abstract 512.
HinesSLMinceyBASloanJA. N02C1: a phase III randomized, placebo-controlled, double-blind trial of risedronate for prevention of bone loss in premenopausal women undergoing adjuvant chemotherapy for breast cancer (BC) [abstract]. J Clin Oncol2008;26(Suppl):Abstract 525.
BrufskyABossermanLCaradonnaR. The effect of zoledronic acid on aromatase inhibitor associated bone loss in postmenopausal women with early breast cancer receiving adjuvant letrozole: the Z-FAST study 36-month follow-up [abstract]. Presented at the San Antonio Breast Cancer Symposium; December 13–16 2007; San Antonio Texas. Abstract 27.
EidtmannHBundredNJDeBoerR. The effect of zoledronic acid on aromatase inhibitor associated bone loss in postmenopausal women with early breast cancer receiving adjuvant letrozole: 36 months follow-up of ZO-FAST [abstract]. Presented at the San Antonio Breast Cancer Symposium; December 10–14 2008; San Antonio Texas. Abstract 44.
SchenkNLombartAFrassoladtiA. The E-ZO-FAST trial: zoledronic acid (ZA) effectively inhibits aromatase inhibitor associated bone loss (AIBL) in postmenopausal women (PMW) with early breast cancer (EBC) receiving adjuvant letrozole (Let). Eur J Cancer2007;5(Suppl 5):186–187. Abstract 2008.
DuchateauLPignonJPBijnensL. Individual patient-versus literature-based meta-analysis of survival data: time to event and event rate at a particular time can make a difference, an example based on head and neck cancer. Controlled Clinical Trials2001;22:538–547.
De BoerREidtmannHLluchA. The ZO-FAST trial: zoledronic acid effectively inhibits aromatase inhibitor associated bone loss in postmenopausal women with early breast cancer receiving adjuvant letrozole: 24 month BMD results. Breast Cancer Res Treat2007;106(Suppl 1):S36. Abstract 501.