Study results were presented in part at the 32nd Annual San Antonio Breast Cancer Symposium, San Antonio, TX, December 9-13, 2009 (Abstract 4083).
Postmenopausal breast cancer (BC) patients receiving adjuvant aromatase inhibitor therapy are at risk of progressive bone loss and fractures. Zoledronic acid inhibits osteoclastic bone resorption, is effective in maintaining bone health, and may therefore be beneficial in this setting.
Overall, 602 postmenopausal women with early, hormone receptor-positive BC receiving adjuvant letrozole were randomized (301 each group) to receive upfront or delayed-start zoledronic acid (4 mg intravenously every 6 months) for 5 years. The primary endpoint was the change in lumbar spine (LS) bone mineral density (BMD) at month 12. Secondary endpoints included changes in LS BMD, total hip BMD, and bone turnover markers at 2, 3, and 5 years; fracture incidence at 3 years; and time to disease recurrence.
At month 61, the adjusted mean difference in LS and total hip BMDs between the upfront and delayed groups was 8.9% and 6.7%, respectively (P < .0001, for both). Approximately 25% of delayed patients received zoledronic acid by month 61. Only 1 patient experienced grade 4 renal dysfunction; no confirmed cases of osteonecrosis of the jaw were reported. Fracture rates (upfront, 28 [9.3%]; delayed, 33 [11%]; P = .3803) and Kaplan-Meier disease recurrence rates (upfront, 9.8 [95% confidence interval (CI), 6.0-10.3]; delayed, 10.5 [95% CI, 6.6-14.4]; P = .6283) were similar at month 61.
Progressive bone loss and subsequent fractures associated with adjuvant aromatase inhibitor (AI) therapy, administered either alone or sequentially after tamoxifen in postmenopausal women with early breast cancer, can develop within the first 12 months and continue throughout the duration of AI therapy.1-3 Nonetheless, AIs are the adjuvant treatment of choice in postmenopausal women with early, hormone receptor-positive breast cancer, because they produce improved disease-free survival rates compared with tamoxifen.4-11 Many postmenopausal women will likely be routinely treated with several years of AI therapy; therefore, implementation of effective bone-loss prevention strategies is needed.
Zoledronic acid is an effective, nitrogen-containing bisphosphonate and is approved in the oncology setting for the treatment of patients with multiple myeloma or bone metastasis from solid tumors (eg, breast cancer, lung cancer, hormone-refractory prostate cancer) and hypercalcemia of malignancy.2, 12 In addition, zoledronic acid has been shown to increase bone mineral density (BMD) in premenopausal and postmenopausal breast cancer patients at risk of cancer treatment-induced bone loss.2, 3, 13-17 Recently, preclinical and clinical data have also demonstrated anticancer effects of bisphosphonates in breast and other cancer types.18-32
The prevention and treatment of bone loss with upfront (initiated simultaneously with letrozole) versus delayed (initiated with a decrease in T score to <−2 or occurrence of clinical nontraumatic fracture) zoledronic acid in early breast cancer patients receiving letrozole was evaluated in 3 similarly designed, geographically diverse studies (Z-FAST; ZO-FAST; E-ZO-FAST) and in the similarly-designed N03CC study.15 Interim (<36 months follow-up) results from the former 3 studies indicate that upfront rather than delayed-start zoledronic acid is significantly more effective in preventing bone loss.2, 3, 13, 14 This article reports the final, 5-year results of the Z-FAST study.
MATERIALS AND METHODS
The design of this 5-year, open-label, multicenter study has been described previously.2 Briefly, postmenopausal women with a history of surgically resectable, early stage (I, II, or IIIa), estrogen receptor-positive and/or progesterone receptor-positive breast cancer receiving letrozole (2.5 mg orally daily for 5 years) were randomized to receive either upfront or delayed-start zoledronic acid (4 mg intravenously [IV] every 6 months). Upfront patients received zoledronic acid after randomization; delayed-start patients received zoledronic acid only if 1) a postbaseline lumbar spine (LS) or total hip T score decreased to <−2.0; 2) any clinical nontraumatic fracture occurred; or 3) an asymptomatic vertebral fracture was identified at the month 36 follow-up. Patients were further stratified according to receipt of previous adjuvant chemotherapy (yes vs no) and baseline T score (normal, >−1.0; low, ≤−1.0 and ≥−2.0).
The primary objective of the study was to measure the difference in percentage change in LS (L1-L4) BMD from baseline to month 12 between the upfront and delayed groups. Secondary objectives included comparing the percentage change difference(s) in 1) LS BMD from baseline to months 24, 36, and 61 (the 4-week follow-up visit after end of study); 2) total hip BMD from baseline to months 12, 24, 36, and 61; and 3) changes in serum N-telopeptide (NTX) and serum bone-specific alkaline phosphatase (BSAP) concentrations from baseline to months 12, 24, 36, and 61. Other secondary objectives were fracture incidence at month 36 and time to disease recurrence.
Bone turnover marker analyses were performed in a subset of patients by a central laboratory. BSAP concentration was measured at baseline, every 3 months during year 1, every 6 months during years 2 to 3, and at months 48 and 61 (or at the final visit in cases of early discontinuation). The initial study protocol required the measurement of serum NTX concentrations; however, a change in assay kits after month 12 resulted in measured concentrations that were quantitatively different and therefore not comparable to those measured at baseline. Consequently, the endpoint (ie, percentage change from baseline) could not be determined using the values measured beyond month 12. Therefore, in July 2005, the protocol was amended to allow the collection of second-morning voided urine samples to measure urine NTX. Urine samples were collected beginning at the next scheduled visit and every subsequent 6 months for all patients active at the time of the amendment. Month 12 serum NTX results have been reported previously.2
The institutional review boards approved the study, and all study participants provided informed consent.
A complete description of all statistical analyses has been published previously.2 For the primary efficacy analysis and to evaluate BMD changes at ≥24 months, the analysis of covariance model was used to compare between-group differences. Factors of the model included treatment (upfront, delayed), chemotherapy use (yes, no), and region (formed by pooling centers); covariates included baseline BMD and years since menopause.
The time to an event (ie, disease recurrence, death, clinical fracture, criteria warranting zoledronic acid initiation [delayed group only]) was defined as the duration in months from randomization to the onset of the event. Kaplan-Meier curves were plotted to calculate disease recurrence and time to first fracture; the log-rank test compared the 2 groups, although the study was not designed nor powered to detect a difference between groups in these endpoints.
Between September 28, 2002, and December 5, 2003, 602 patients were enrolled and randomized to receive either upfront or delayed zoledronic acid (Fig. 1). Baseline patient characteristics were similar between the groups (Table 1).
Abbreviations: ECOG, Eastern Cooperative Oncology Group; SD, standard deviation.
Based on all randomized patients unless otherwise specified.
P >.05 for differences between treatment groups for all characteristics.
One patient in each group randomized but not treated.
Includes 6 upfront patients and 1 delayed-start patient <30 years old at menopause onset.
Based on local dual-energy x-ray absorptiometry scan results.
One patient with baseline T score of −2.1 included in T score ≤−1.0 and ≥−2.0 category.
Major risk factors: body mass index <20 kg/m2; current cigarette smoker or smoker within last 10 years; current other smoker or smoker within last 10 years; corticosteroid therapy; adult broken bone; hip fracture in mother.
Minor risk factors: age ≥65 years; postmenopausal; age at menopause ≤45 years; irregular menstrual cycle; treatment history for ≥1 of the following: hyperthyroidism, rheumatoid arthritis, endocrine disorder, seizure disorder, dementia, anorexia nervosa, bulimia; lack of mobility/exercise; poor/frail health; lack of vitamins/dairy intake as adult or child; family history of osteoporosis; >2 alcoholic drinks per day.
33 Patients (upfront, 19; delayed, 14) did not undergo osteoporotic risk factor assessment.
Major or minor osteoporotic risk factors, No. (%)ghi
Upfront zoledronic acid progressively increased LS and total hip BMD throughout the study; conversely, the delayed group experienced significant decreases in LS and total hip BMD at all time points (Fig. 2A and C). The least squares mean difference between groups in percentage change of BMD steadily increased from baseline to month 61 for LS (4.3% [95% confidence interval (CI), 3.7-4.9] to 8.9% [95% CI, 7.4-10.5]) and total hip (3.2% [95% CI, 2.7-3.7] to 6.7% [95% CI, 5.5-8.0]). By month 61, 74 delayed patients (24.6%) received zoledronic acid. Most (49 patients; 66.2%) met protocol-defined criteria for receiving zoledronic acid, whereas the remaining patients received zoledronic acid at the investigators' discretion.
A post hoc analysis excluded those delayed group patients who received zoledronic acid by month 61; the results showed an even greater BMD decrease by month 61 in patients receiving no zoledronic acid compared with all delayed group patients combined (LS, −5.2% vs −2.42%; total hip, −6.13% vs −4.12%) (Fig. 2B and D) and greater least squares mean differences compared with the upfront group (LS, 11.6% [95% CI, 10-13.2] vs 8.9% [95% CI, 7.4-10.5]; total hip, 8.3% [95% CI, 7-9.5] vs 6.7% [95% CI, 5.5-8]). In addition, delayed group patients who initiated zoledronic acid experienced small increases in BMD by month 61 (LS, 3.99%; total hip, 0.52%).
Zoledronic acid produced substantial increases in BMD, regardless of baseline T score, osteoporotic risk factors, or chemotherapy status. Up to month 36, increases in LS BMD were slightly greater in upfront patients with lower baseline T scores. However, patients receiving upfront zoledronic acid had similar percentage LS BMD changes from baseline to month 61 (6.18% for T score >−1; 6.22% for T score ≤−1.0 and ≥−2.0) regardless of baseline T score, suggesting that patients with low baseline T scores experience more rapid BMD improvements early in treatment than patients with higher baseline T scores. In the delayed group, larger BMD decreases were observed in patients with only 1 risk factor, likely because a greater proportion of delayed group patients with >1 baseline risk factor received zoledronic acid by the end of the study.
Delayed group patients were more likely to shift to a lower T score category at all time points; by month 61, 11% of patients with a baseline T score >−1 had a worse T score (≤−1.0 and ≥−2.0) compared with 2% of upfront patients. Likewise, 4.9% of delayed patients (vs 0% upfront patients) whose baseline T score was ≤−1 and ≥−2 experienced a decrease in T score to <−2 by month 61; the T score in 28% of upfront patients with a baseline T score ≤−1 and ≥−2 increased to >−1.
Because a 2% to 6% experimental error between any 2 BMD measurements exists, we evaluated the proportion of patients with an 8% or greater decrease in BMD to ensure any negative BMD change was accurately detected.33 The proportion of patients with a cumulative decrease of at least 8% in LS BMD from baseline to all time points was significantly lower in patients receiving upfront zoledronic acid than in delayed group patients regardless of baseline T score (Table 2). Delayed patients with normal baseline T scores were significantly more likely to have BMD decreases of at least 8% compared with patients with low baseline T scores (Table 2).
Table 2. Patients With Decrease in Lumbar Spine Bone Mineral Density (g/cm2) of ≥8% From Baselinea
Baseline T Score
Upfront Group, n=300, No. (%)
Delayed Group, n=300, No. (%)
−1 to −2
−1 to −2
−1 to −2
−1 to −2
−1 to −2
Markers of Bone Turnover
At the first collection, which occurred between months 24 and 48, the median urine NTX value was 27.0 nM BCE/mM creatinine in the upfront group and 40.0 nM BCE/mM creatinine in the delayed group. Similar levels were observed at subsequent time points, regardless of the timing of the first collection, with median urine NTX levels always being lower in the upfront group. The difference between the groups was not significant at any time, possibly because more delayed group patients received zoledronic acid treatment at later time points.
The upfront group experienced median decreases in BSAP (U/L) levels compared with baseline levels at all time points, whereas the delayed group experienced median increases at most time points (Fig. 3). The differences between the treatment groups were statistically significant (P ≤ .0026) at all postbaseline time points. An unexpected small median percentage decrease (−1.0%) in BSAP levels compared with baseline levels occurred at month 48 in the delayed group. This can be explained, however, because of the inclusion of delayed group patients who had initiated zoledronic acid during this time. In these patients, the median percentage change in BSAP from baseline to month 48 decreased by 19.5%, whereas in patients who did not initiate zoledronic acid by month 48, the median percentage change in BSAP level increased by 6.3%. Likewise, by the end of the study, BSAP levels increased by a median of 18% in delayed group patients who never received zoledronic acid during the study; however, the percentage increase in all delayed group patients was less (8.7%), because BSAP levels in patients who received zoledronic acid by the end of the study decreased by 14.2%.
Fractures occurred in slightly more delayed than upfront patients by month 61 (28 [9.3%] vs 33 [11%] patients; P = .3803); however, the difference was not statistically significant. Most fractures were associated with significant trauma (upfront group: trauma, 17 [5.7%]; no/minimal trauma, 5 [1.7]; delayed group: 21 [7%]; 6 [2%]). The median time to the first fracture was shorter in the upfront group (24.7 months) than the delayed group (34.4 months). The most common fracture sites included the rib and foot (7 patients [2.3%], each site) in the upfront group and the vertebrae (6 patients [2.0%]) in the delayed group.
Disease Recurrence or Death
Small differences in the Kaplan-Meier incidence rates of disease recurrence or death were observed between the groups at months 12, 24, 36, and 48 in favor of the upfront group; however, rates at month 61 were similar between the groups (percentage [95% CI]: upfront, 9.8 [6.0-10.3]; delayed, 10.5 [6.6-14.4]; P = .6283) (Table 3). Disease recurrence alone occurred in slightly more delayed group patients compared with the upfront group (16 patients [5.3%] vs 21 patients [7.0%]). The most common sites of recurrence were the lymph nodes and bone in the upfront and delayed patients, respectively. Death occurred in 7 upfront patients (2.3%) and 4 delayed patients (1.3%) by month 61.
Table 3. Disease Recurrence or Death at 36 and 61 Monthsa
Type of Disease Recurrence
Upfront Group, n=300
Delayed Group, n=300
Upfront Group, n=300
Delayed Group, n=300
Abbreviation: CI, confidence interval.
Time to disease progression (recurrence) was a secondary endpoint.
From any cause; no deaths were suspected to be related to study drug. The causes of death in the upfront group: acute cardiomyopathy, acute myocardial infarction, atrial fibrillation, cerebral hemorrhage, respiratory failure (2 patients), secondary ovarian cancer. The causes of death in the delayed group: cardiac failure, cardiorespiratory arrest, subdural hygroma, traumatic intracranial hemorrhage.
Lymph node metastases; unknown whether local regional or distant recurrence. Two patients had lymph node metastases; 1 patient had lymph node and chest wall metastases.
Disease recurrence or death, Kaplan-Meier % [95% CI]b
The types and rates of adverse events (AEs) were similar between the 2 groups (Table 4), most were grades 1 or 2. The incidence of serious AEs (SAEs) was also similar (upfront, 27.7% [83 patients]; delayed group, 23.7% [71 patients]). Nine patients (3.0%) in the upfront group and 3 (1.0%) in the delayed group experienced SAEs that were believed to be study drug related.
Table 4. AEs (All Grades) in >10% of Patients
Type of AE
Upfront Group, n=300, No. (%)
Delayed Group, n=300, No. (%)
Abbreviation: AE, adverse event.
Urinary tract infection
Osteonecrosis of the jaw (ONJ) was reported by investigators as study drug related in 2 upfront patients and resulted in permanent discontinuation of study medication. However, neither case was confirmed. Based on the prespecified definition of ONJ, an ONJ adjudication committee deemed 1 of the cases inconsistent with ONJ and the other indeterminate because of insufficient information.
Renal impairment (grade 1-2) or failure (grade 3-4) was reported in 6 upfront patients (2.0%) (grade 1-2, 5 patients; grade 4, 1 patient) and 4 delayed group patients (1.3%) (grade 1-2, 3 patients; grade 3, 1 patient). The patient who experienced grade 4 renal failure had several comorbidities (anemia, acute myocardial infarction, viral infection). Of the other 5 patients in the upfront group, 3 experienced renal impairment most likely linked to the study drug (2 cases necessitated dose interruption; 1 led to permanent discontinuation). Renal AEs were not suspected to be study drug related in the delayed group.
Based on the results of the current clinical trial and other recently completed studies, zoledronic acid effectively increases BMD in breast cancer patients receiving adjuvant endocrine therapy and at risk for bone loss and fracture.15, 17 The final results of the bone substudy of the Austrian Breast and Colorectal Cancer Study Group-12 trial (N = 1803), which evaluated zoledronic acid (4 mg IV every 6 months) in 404 premenopausal women with hormone-responsive breast cancer receiving goserelin and either anastrozole or tamoxifen, demonstrate significant increases in BMD (LS, +4% [P = .02]; trochanter, +3.9% [P = .07]) compared with baseline values at 5 years, even 2 years after therapy cessation.17 Conversely, women receiving endocrine therapy alone (no zoledronic acid) experienced significant decreases in BMD (LS, −6.3% [P = .001]; trochanter, −4.1% [P = .06]) at 5 years.17 Likewise, in a study comparing upfront versus delayed zoledronic acid in postmenopausal women who received 2 years of letrozole after ≤6 years of tamoxifen, Hines and colleagues observed significant losses in BMD in the delayed group (LS, −2.3%; total hip, −3.3%) compared with the upfront group (LS, 4.9%; total hip, −1.2%), findings similar to those observed in the current study.15
Regardless of baseline T score, chemotherapy status, or the presence of risk factors, we observed significant and progressive increases in LS and total hip BMD throughout the study with the initiation of zoledronic acid at study randomization; in contrast, significant decreases occurred when zoledronic acid initiation was delayed until bone loss was apparent. Over time, the rate of BMD decline in the delayed group slowed (see Fig. 2), most likely because more delayed patients received zoledronic acid by the study end. In a post hoc analysis, delayed patients who received zoledronic acid by month 60 experienced increases in LS BMD (+4%) from baseline despite the delay in therapy. Increases in LS BMD in the delayed patients who received zoledronic acid, however, were lower compared with increases in the upfront group (+6.2%); this difference may be explained by the finding that 20% of delayed patients who received zoledronic acid did so after month 36, minimizing patient exposure to zoledronic acid. When delayed patients who received zoledronic acid by month 60 were excluded from the analysis, decreases in LS BMD in the delayed subgroup who received no zoledronic acid (−5.2%) were even greater compared with the decreases in all delayed patients (−2.4%); therefore, the least squares mean difference between upfront and delayed patients who never received zoledronic acid was almost 3% more than the difference between upfront and all delayed patients. In addition, delayed patients with a low baseline T score experienced slight increases in LS BMD (1.7%), probably because most of these patients eventually met the criteria for receiving zoledronic acid. However, upfront patients with the same baseline T score experienced substantially greater increases in LS BMD (6.2%) after 5 years of zoledronic acid treatment. These findings indicate that BMD is maintained more effectively with upfront administration, but bone loss in delayed patients is likely reversible, and initiation of zoledronic acid therapy, even after bone loss has developed, is beneficial.
Zoledronic acid's beneficial effects on BMD in breast cancer patients receiving AI therapy are believed to be due to the ability of zoledronic acid to inhibit estrogen deprivation-associated osteoclast-mediated bone resorption; this hypothesis is supported by the metabolic bone marker data that favor upfront zoledronic acid.
Slightly fewer upfront patients experienced fractures compared with delayed patients; however, the differences were minimal and not statistically significant (P = .3803). The lack of a trend may have been a result of the high number of traumatic fractures in both groups (>50%) and the use of zoledronic acid in delayed patients at highest risk for fracture. These results should be interpreted cautiously, because the study was not adequately powered to detect a difference in the fracture rates.
The results from studies evaluating zoledronic acid's affects on disease recurrence and survival in breast cancer are conflicting. In the Austrian Breast and Colorectal Cancer Study Group-12 study, Gnant and colleagues observed a 36% reduction in risk of disease recurrence in premenopausal patients receiving zoledronic acid plus endocrine therapy (P = .01) compared with endocrine therapy alone.28 Conversely, the results from the 5-year AZURE study, which evaluated 3360 stage II/III breast cancer patients receiving neoadjuvant chemotherapy and/or endocrine therapy randomized to receive or not receive zoledronic acid 4 mg (every 3-4 weeks for 6 doses, then every 3 months for 8 doses, and then every 6 months for 5 doses), indicate that zoledronic acid, after a median follow-up of 59 months, has no overall effect on either disease-free survival or overall survival. However, in a preplanned subanalysis of the AZURE results, zoledronic acid did show a significant effect on overall survival in women with well-established menopause (ie, >5 years postmenopausal) (n = 1101), reducing the risk for death by 29% (hazard ratio, 0.71; P = .017).34, 35 In the current study, the disease recurrence and death rate at month 12 was lower in the upfront group compared with the delayed group (Kaplan-Meier percentage: upfront, 0.3 [95% CI, 0.0-1.0]; delayed, 2.2 [0.5-3.9]; P = .0557); however, the rates between the groups at later time points differed only slightly in favor of the upfront group. Interestingly, although no statistically significant difference in the overall disease recurrence rate was observed at month 61 (Kaplan-Meier percentage: upfront, 9.8 [95% CI, 6.0-10.3]; delayed, 10.5 [95% CI, 6.6-14.4]; P = .6283), patients in the delayed group were more likely to experience a distant recurrence (upfront, 9 [3.0%] vs delayed, 19 [6.3%]). These findings have limited implications, however, because the 40% of patients who discontinued the study early were excluded from this analysis. Furthermore, whether the use of zoledronic acid in 25% of the delayed patients affected the results of this analysis is unknown. Although the results of these studies suggest that zoledronic acid may improve survival in breast cancer patients to some degree, the results from other ongoing clinical trials (eg, Southwest Oncology Group S0307) evaluating disease-free survival in breast cancer patients receiving zoledronic acid are needed before zoledronic acid can be recommended for use in this setting.
The long-term (5 years) use of letrozole and zoledronic acid was generally well tolerated, with a safety profile similar to previous reports. No confirmed cases of ONJ were observed during the study, and only 1% of patients experienced study drug-related renal impairment.
Upfront adjuvant zoledronic acid significantly increases BMD in postmenopausal early breast cancer patients receiving an AI for 5 years, whereas patients randomized to receive delayed zoledronic acid overall progressively lost BMD during the 5 years. Patients in the delayed zoledronic acid group who were treated with zoledronic acid after bone loss had occurred did benefit by subsequent increases in BMD.
We thank the patients who participated in this study, investigators, and study coordinators; PRA International, BioImaging Technologies, and CRL Medinet; Dr. Mei Dong and Dr. Nicholas Sauter for assistance in the protocol development; and Syntaxx Communications, specifically Stephanie Butler, PharmD, and Alison Shore, MS, who provided manuscript development and editing services, respectively, on behalf of Novartis Pharmaceuticals Corporation.
Novartis Pharmaceuticals Corporation, East Hanover, New Jersey.
CONFLICT OF INTEREST DISCLOSURES
L.B. is a consultant for and received an honoraria from Amgen and Pfizer, and her place of practice receives research funding from various pharmaceutical companies; C.V.'s place of practice receives research funding from Novartis Pharmaceuticals Corporation, and he has served on an advisory board for Novartis Pharmaceuticals Corporation; G.W., L.J., S.G.E., J.H., and E.A.-A. own stock in Novartis Pharmaceuticals Corporation.