Articles

Efficacy and safety of 12-weekly versus 4-weekly zoledronic

acid for prolonged treatment of patients with bone
metastases from breast cancer (ZOOM): a phase 3, open-label,
randomised, non-inferiority trial
Dino Amadori, Massimo Aglietta, Barbara Alessi, Lorenzo Gianni, Toni Ibrahim, Gabriella Farina, Fernando Gaion, Francesco Bertoldo,
Daniele Santini, Roberta Rondena, Paola Bogani, Carla I Ripamonti

Summary
Background Zoledronic acid reduces skeletal-related events in patients with breast cancer, but concerns have been Lancet Oncol 2013; 14: 663–70
raised about prolonged monthly administration. We assessed the efficacy and safety of a reduced dosing frequency of Published Online
zoledronic acid in women treated previously with monthly zoledronic acid. May 16, 2013
http://dx.doi.org/10.1016/
S1470-2045(13)70174-8
Methods We did this non-inferiority, phase 3 trial in 62 centres in Italy. We enrolled patients with breast cancer who
See Comment page 575
had one or more bone metastases and had completed 12–15 months of monthly treatment with zoledronic acid.
Osteoncology and Rare Tumors
Patients were randomly assigned with a permutated block (size four to eight) random list stratified by centre in a Center, IRCCS Scientific
1:1 ratio to zoledronic acid 4 mg once every 12 weeks or once every 4 weeks, and followed up for at least 1 year. Neither Institute of Romagna for the
patients nor investigators were masked to treatment allocation. The primary outcome was skeletal morbidity rate Study and Treatment of Cancer
(skeletal-related events per patient per year) in the intention-to-treat population. We used a non-inferiority margin of (IRST IRCCS), Meldola, Italy
(Prof D Amadori MD,
0·19. The trial is registered with EudraCT, number 2005-004942-15. T Ibrahim MD); Oncological
Department, Medical Oncology,
Findings We screened 430 patients and enrolled 425, of whom 209 were assigned to the 12-week group and 216 to the Institute for Cancer Research
4-week group. The skeletal morbidity rate was 0·26 (95% CI 0·15–0·37) in the 12-week group versus 0·22 (0·14–0·29) and Treatment, Candiolo, Turin,
Italy (M Aglietta MD);
in the 4-week group. The between-group difference was 0·04 and the upper limit of one-tailed 97·5% CI was 0·17, Department of Medical
which is lower than the non-inferiority margin. The most common grade 3–4 adverse events were bone pain (56 [27%] Oncology University Hospital of
patients in the 12-week group vs 65 [30%] in the 4-week group), nausea (24 [11%] vs 33 [15%]), and asthenia (18 [9%] vs Udine, Udine, Italy (B Alessi MD);
Oncology Department, Infermi
33 [15%]). Renal adverse events occurred in one patient (<1%) in the 12-week group versus two (1%) in the 4-week
Hospital, Rimini, Italy
group. One patient (<1%) in the 4-week group had grade 1 acute renal failure. Osteonecrosis of the jaw occurred in (L Gianni MD); Department of
four patients in the 12-week group versus three in the 4-week group. No treatment-related deaths were reported. Medical Oncology,
Median N-terminal telopeptide concentration changed from baseline more in the 12-week group than in the 4-week Fatebenefratelli-Oftalmico
Hospital, Milan, Italy
group after 12 months (12·2% vs 0·0%; p=0·011).
(G Farina MD); Department of
Medical Oncology, ULSS 15,
Interpretation Our results raise the possibility of decreasing administration of zoledronic acid to a 12-weekly regimen Camposampiero, Italy
to reduce exposure during the second year, while maintaining its therapeutic effects. However, the effects on (F Gaion MD); Institute of
Internal Medicine, Department
N-terminal telopeptide should be investigated further before changing current practice.
of Medicine, University of
Verona, Verona, Italy
Funding Novartis Farma. (F Bertoldo MD); Department of
Medical Oncology, University
Campus Bio-Medico, Rome,
Introduction Zoledronic acid (4 mg intravenously every 3–4 weeks) Italy (D Santini MD); Novartis
About 65–75% of patients with advanced breast cancer is a well-established antiresorptive drug approved for use Farma, Origgio, Italy
develop bone metastases,1 making bone the most in patients with bone-metastatic breast cancer to delay (R Rondena BSc, P Bogani BSc);
common site of breast cancer advancement.2 Malignant the onset and reduce the risk of skeletal-related events.8–10 and Supportive Care in Cancer
Unit, Department of
bone disease severely impairs normal skeletal homoeo- Although the optimum duration of treatment has not Hematology and Pediatric
stasis and can result in debilitating pain and skeletal- yet been established, zoledronic acid has been shown to Onco-Hematology, National
related events, including pathological fractures, spinal reduce skeletal complications for up to 2 years in some Cancer Institute, Milan, Italy
cord compression, severe bone pain, the need for pal- studies,10–14 and even longer in others.15–17 Because of the (C I Ripamonti MD)

liative radiotherapy or surgery, and hypercalcaemia.3 paucity of evidence for treatment beyond 2 years, clinical Correspondence to:
Prof Dino Amadori, Osteoncology
Without treatment to inhibit bone resorption, about practice guidelines of the American Society of Clinical
and Rare Tumors Center, IRCCS
two-thirds of patients with bone-metastatic breast Oncology recommend continuation of treatment until Istituto Scientifico Romagnolo
cancer develop skeletal-related events (roughly three or there is evidence of a substantial fall in general health per lo Studio e la Cura dei Tumori,
four per year per patient),4 which can limit functional status.18,19 An expert panel has recommended treatment 47014 Meldola, Italy
[email protected]
independence and reduce quality of life.5,6 Moreover, beyond 2 years based on individual risk assessment,
patients with pathological fractures have an increased since skeletal complications can occur repeatedly
risk of death.7 throughout the disease course and the risk of

www.thelancet.com/oncology Vol 14 June 2013 663

 Articles

complications exists for as long as bone metastases are than 265 μmol/L, corrected serum calcium concentration
present.20 Prolonged treatment with zoledronic acid less than 2 mmol/L or more than 3 mmol/L, presence of
increases the risk of osteonecrosis of the jaw. Therefore, active dental pathology, and recent (within 6 weeks) or
as new cancer treatments extend the life expectancy of planned dental procedures (eg, extraction, implants).
patients with advanced breast cancer,21 better strategies Concurrent cancer treatment was allowed.
for treatment with zoledronic acid are needed for long- All patients provided written informed consent.
term reduction and prevention of skeletal-related events. The study was approved by the institutional review board
We investigated whether reduced-frequency dosing of for each study site.
zoledronic acid after 1 year of standard dosing might
provide the same efficacy as the standard schedule while Randomisation and masking
improving adherence and safety. Our hypothesis is Patients were randomly assigned to treatment groups in
supported by studies22 of treatment of postmenopausal a 1:1 ratio according to a permutated block (size four to
osteoporosis, in which only a few doses of zoledronic eight) random list generated by an independent
acid suppressed bone turnover for many years because of statistician through a validated computer programme.
the potency of osteoclast inhibition and skeletal The randomisation list was stratified by centre with
accumulation of zoledronic acid. treatments in balanced blocks within each centre.
The block size was not revealed to prevent bias at
Methods randomisation. Eligible patients were allocated by the
Study design and participants investigator to the smallest available random number of
We did this phase 3, prospective, randomised, open- the list. Sealed envelopes containing the randomisation
label, non-inferiority, trial in 62 hospitals in Italy. Eligible code for each patient were produced and sent to centres:
patients were women (age ≥18 years) with stage IV breast the investigators opened them sequentially when
cancer, one or more radiologically confirmed bone assigning a new patient. Nobody involved in the study
metastases, and treated with zoledronic acid every was masked to treatment allocation.
3–4 weeks for 12–15 months before enrolment. Exclusion
criteria were: more than 3 months since the last infusion Procedures
of zoledronic acid, use of bisphosphonates other than Patients were assigned to receive either zoledronic acid
zoledronic acid, serum creatinine concentration more 4 mg once every 12 weeks or once every 4 weeks for
roughly 1 year. All patients received daily calcium (500 mg)
430 patients assessed for eligibility
and vitamin D (400–500 IUs) supplements. Patients were
not followed up beyond the duration of the study.
5 excluded
Skeletal-related events were defined as any of: vertebral
3 withdrew consent or non-vertebral pathological fracture, radiotherapy or
1 violated protocol surgery to bone, spinal-cord compression, or hyper-
1 no longer needed study drug
calcaemia of malignancy. Pathological vertebral fractures
were identified by radiography or CT scan by comparing
425 enrolled anterior, medial, and posterior vertebral heights with those
measured at baseline using the visual semiquantitative
assessment proposed by Genant and colleagues.23 A 4-mm
209 assigned to 12-weekly zoledronic acid 216 assigned to 4-weekly zoledronic acid reduction in height was indicative of fracture.23 Each
patient assessed pain intensity at rest and on movement by
60 discontinued 74 discontinued a validated 6-point Verbal Rating Scale.24 Serum N-terminal
21 adverse events 27 adverse events telopeptide concentration was assessed in patients for
11 died 10 died
10 violated protocol 8 violated protocol whom samples were collected.
9 withdrew consent 13 withdrew consent Adverse events and serious adverse events, as well
4 unsatisfactory therapeutic 4 unsatisfactory therapeutic
effect effect
as changes in laboratory test results, were monitored
3 no longer needed study drug 7 no longer needed study drug throughout the study. Renal function was tested before
1 abnormal laboratory test result 2 abnormal laboratory test result each dose of zoledronic acid according to prescribing
1 lost to follow-up 1 lost to follow-up
1 abnormal test result information. Patients with increased serum creatinine
1 administrative reasons concentration had their treatment with zoledronic acid
delayed until concentrations returned to within 10% of
149 completed study 142 completed study baseline. Patients were immediately withdrawn from the
trial if hypercalcaemia of malignancy developed within
14 days after the last infusion of study drug. Any episode of
209 analysed 216 analysed hypercalcaemia of malignancy was treated at the discretion
of the physician. All patients were advised to have a dental
Figure 1: Trial profile examination before starting treatment but no specific

664 www.thelancet.com/oncology Vol 14 June 2013

 Articles

dental assessment was defined by the protocol. Every site and the assumed standard deviation (2·04) of the original
used local guidelines for prevention of osteonecrosis of the protocol, the non-inferiority margin was reduced,
jaw. With the exception of adverse and skeletal-related according to the ratio of the two estimated SDs, to 0·19.
events, which were assessed continuously, all parameters Safety analyses included all assigned patients who
were assessed before the first dose at baseline, and at 3, 6, received at least one dose of zoledronic acid. Adverse
9, and 12 months.
The primary endpoint was overall skeletal morbidity 12-week group (n=209) 4-week group (n=216)
rate (number of skeletal-related events per patient per Mean age (SD; years) 60·4 (11·9) 59·8 (11·8)
year). Secondary endpoints were incidence of each Women 209 (100%) 216 (100%)
skeletal-related event per year, proportion of patients who
White ethnic origin 209 (100%) 216 (100%)
had skeletal-related events, time to first skeletal-related
ECOG performance status
event, bone pain, use of analgesics, N-telopeptide of
0 or 1 207 (99%) 214 (99%)
type I collagen concentration, and safety.
2 2 (1%) 2 (1%)
Mean bodyweight (SD; kg) 67·7 (13·0) 68·8 (12·4)
Statistical analysis
Mean BPI composite pain score (SD) 2·0 (1·8) 2·1 (1·9)
Efficacy analyses were done in the intention-to-treat
Analgesic use
population. To adjust for the effect of premature with-
None 131 (63%) 138 (64%)
drawal from the study, skeletal morbidity rates in each
Minor 41 (20%) 31 (14%)
treatment group were compared by an ANCOVA model
Tranquillisers and others 5 (2%) 2 (1%)
with the skeletal morbidity rate entered as the dependent
variable, treatment group as the independent variable, Mild or strong 32 (15%) 45 (21%)

and study duration as the covariate. To protect against Previous skeletal-related events 120 (57%) 124 (57%)

the assumptions of the ANCOVA model not being met, General oral health
data were also analysed with a negative-binomial model, Poor 2 (1%) 0 (0%)
with results presented as skeletal morbidity rates. Fair 19 (9%) 15 (7%)
The 95% CIs of the difference between means for each Good 165 (79%) 180 (83%)
group were calculated with the estimate of variability Excellent 22 (11%) 20 (9%)
and the least-squares means from the ANCOVA test. Missing data 1 (<1%) 1 (<1%)
Skeletal morbidity rate for the 4-weekly group was Mean time from diagnosis to study entry (SD; months)
expected to be 0·91, and a previous study4 showed that Metastatic breast cancer 23·5 (25·0) 22·6 (26·0)
the difference between zoledronic acid and pamidronate Bone metastases 18·1 (19·7) 16·7 (18·7)
was 0·67. Therefore, a difference of 0·56 was considered Number of lines of previous cancer treatment
acceptable. If the upper limit of the 95% CI did not None 20 (10%) 19 (9%)
exceed 0·56, the 12-week dosing schedule was deemed One 94 (45%) 112 (52%)
non-inferior to the 4-week schedule. Because of the Two 49 (23%) 34 (16%)
calculation method used, this assumption is equivalent More than two 46 (22%) 51 (24%)
to the use of a one-sided 97·5% CI. Skeletal morbidity Cancer treatment at baseline
rate was also calculated and summarised for each type of Chemotherapy 12 (6%) 12 (6%)
event. Multiple events were analysed by the Andersen- Endocrine 43 (21%) 48 (22%)
Gill method.25 Time to first skeletal-related event was Chemotherapy and endocrine 132 (63%) 124 (57%)
analysed by the Kaplan-Meier method and Cox pro- Chemotherapy and anti-HER2 6 (3%) 10 (5%)
portional hazards model. The Wilcoxon rank-sum test Endocrine and anti-HER2 2 (1%) 2 (1%)
was used to compare N-terminal telopeptide concen- Chemotherapy, endocrine, and anti-HER2 14 (7%) 19 (9%)
trations between treatment groups at different points; Missing data 0 (0%) 1 (<1%)
this analysis was exploratory, therefore it was not cor- On study cancer treatment
rected for multiple comparisons. Correlation between None 7 (3%) 4 (2%)
quantitative parameters (pain and analgesic scores) was Chemotherapy 20 (10%) 21 (10%)
evaluated with Spearman correlation coefficients. Endocrine 109 (52%) 105 (49%)
With a predefined non-inferiority margin of 0·56,
Anti-HER2 6 (3%) 8 (4%)
a sample size of 420 patients was estimated to be needed
Chemotherapy and endocrine 48 (23%) 54 (25%)
to detect non-inferiority for the primary outcome with
Chemotherapy and anti-HER2 7 (3%) 5 (2%)
80% power (one-sided α=0·025). At the time of the
Endocrine and anti-HER2 7 (3%) 10 (5%)
analysis, the mean skeletal morbidity rate was lower than
Chemotherapy, endocrine, and anti-HER2 5 (2%) 9 (4%)
expected and the estimated pooled standard deviation of
the study was 0·70, which was roughly a third of the Data are n (%) unless otherwise stated. BPI=Brief Pain Inventory; ECOG=Eastern Cooperative Oncology Group.
assumed SD in the original protocol (2·04). To maintain
Table 1: Baseline characteristics for the intention-to-treat population
the same ratio between the non-inferiority margin (0·56)

www.thelancet.com/oncology Vol 14 June 2013 665

 Articles

events were coded using the National Cancer Institute last patient completed treatment on Feb 17, 2010.
Common Toxicity Criteria. SAS (version 9.1) was used for Treatment groups were well-balanced for baseline
all statistical analyses except for analysis of bone marker, characteristics (table 1). Mean age was 60 years (range
for which SAS (version 9.2) was used. 28–89). The number and type of previous skeletal-related
This study is registered with EudraCT, number events was much the same in each group. Most patients
2005-004942-15. received concurrent cancer treatment. The majority of
patients had received chemotherapy plus endocrine
Role of the funding source treatment or endocrine treatment alone at baseline.
The sponsor contributed to the study design, did the Concurrent cancer treatments were mainly hormone
statistical analysis, contributed to the writing of the treatment and chemotherapy plus hormone treatment.
report, and agreed with the decision to submit the paper Previous and concomitant cancer treatments were typical
for publication. The corresponding author had full access of the antineoplastic drugs commonly used to treat
to all the data and the final responsibility to submit for women with stage IV breast cancer and the treatment
publication. groups did not differ greatly (data not shown). General
oral health was either good or excellent in almost all
Results participants in both groups. 291 of 425 (68%) patients
Enrolment started on Feb 20, 2006. 425 patients were completed the study—ie, were monitored for 1 year.
enrolled and randomly assigned to treatment (209 to the Median follow-up (considering the time to event or
12-weekly group, 216 to the 4-weekly group; figure 1). The withdrawal from the study) was 337 days (IQR 205–346)
overall, with little difference between groups. The largest
12-week group 4-week group centre enrolled 28 patients, and only 14 centres enrolled
(n=209) (n=216) more than ten patients. Thus, the analysis was done
Skeletal morbidity rate (95% CI)* 0·26 (0·15–0·37) 0·22 (0·14–0·29) without considering the centre as a factor.
Number of skeletal-related events during study (%) Overall skeletal morbidity rate was 0·26 (95% CI
Any 31 (15%) 33 (15%) 0·15–0·37) in the 12-week group and 0·22 (95% CI
Radiation to bone 22 (11%) 24 (11%) 0·14–0·29) in the 4-week group (table 2). The between-
Pathological fracture (nonvertebral) 7 (3%) 5 (2%) group difference was 0·04 and the upper limit of the one-
Pathological fracture (vertebral) 3 (1%) 4 (2%) tailed 97·5% CI was 0·17, indicating that the 12-week
Spinal-cord compression 2 (1%) 1 (<1%) schedule was non-inferior to the 4-week schedule.
Surgery to bone 2 (1%) 1 (<1%) According to the negative binomial model, the skeletal
Hypercalcaemia of malignancy 1 (<1%) 4 (2%) morbidity rate ratio (4-week group vs 12-week group) was
Incidence of skeletal-related events during study (per year; 95% CI)
0·97 (95% CI 0·60–1·57; p=0·896). Even with this model,
Any 0·26 (0·15–0·37) 0·22 (0·14–0·29)
the results were very close in the two groups. The upper
Radiation to bone 0·17 (0·08–0·26) 0·14 (0·09–0·20)
limit of the CI—assuming a reference rate of 0·22 events
per year—was 0·35 in the 12-week group, with a
Pathological fracture (nonvertebral) 0·08 (0·00–0·16) 0·03 (0·00–0·06)
difference of 0·13, which was within the non-inferiority
Pathological fracture (vertebral) 0·02 (0·00–0·05) 0·02 (0·00–0·04)
margin. Excluding patients with hypercalcaemia of
Spinal-cord compression 0·01 (0·00–0·02) 0·00 (0·00–0·01)
malignancy, the between-group difference was 0·05 and
Surgery to bone 0·02 (0·00–0·05) 0·00 (0·00–0·01)
the upper limit of one-tailed 97·5% CI was 0·18, which
Hypercalcaemia of malignancy 0·01 (0·00–0·02) 0·02 (0·00–0·04)
was lower than the non-inferiority margin.
Median VRS Pain score†‡
Much the same proportions of patients in each group
At rest
had on-study skeletal-related events (overall and by type).
Baseline 1 (120/152; 79%) 1 (146/170; 86%)
Overall, 31 of 209 (15%) patients in the 12-week group and
3 months 2 (130/155; 84%) 2 (151/169; 89%)
33 of 216 (15%) in the 4-week group had an on-study
6 months 2 (124/147; 84%) 1 (143/167; 86%)
skeletal-related event (table 2; p=0·898). Andersen-Gill
9 months 1 (111/146; 76%) 1 (133/152; 88%)
multiple event analysis did not show a significant
End of study 2 (125/160; 78%) 1 (129/152; 85%) difference between groups (hazard ratio [HR] 1·01, 95%
On movement CI 0·62–1·64; p=0·972). At 12 months, the Kaplan-Meier
Baseline 2 (115/152; 76%) 2 (134/170; 79%) estimate of the skeletal-related events was 22·1% (95% CI
3 months 2 (113/155; 73%) 2 (134/169; 79%) 12·8–36·6) in the 12-week group and 18·3% (95% CI
6 months 2 (107/147; 73%) 2 (132/167; 79%) 13·3–25·0) in the 4-week group (log-rank p=0·912;
9 months 2 (100/146; 68%) 2 (110/152; 72%) figure 2). Median time to first on-study skeletal-related
End of study 3 (109/160; 68%) 2 (109/152; 72%) event could not be calculated because of the very low event
*Upper limit of one-tailed 97·5% CI of difference=0·17. †Assessed with a validated, 6-point Verbal Rating Scale. ‡Data
rate. Skeletal morbidity rates for individual events were
in parentheses are number of patients with score <4/number of patients with data available. much the same between groups (table 2).
We detected no relevant differences between groups for
Table 2: Efficacy outcomes (intention-to-treat population)
bone pain or analgesic use during follow-up. In particular,

666 www.thelancet.com/oncology Vol 14 June 2013

 Articles

the median pain at rest and pain on movement scores

100
were less than 4 at all points in both groups, and most
patients had a score of less than 4. The lowest percentage 90
of patients with scores of less than 4 was for pain on
80
movement at the last visit in the 12-week group (table 2).
Median analgesic use was zero at all points in both 70

No skeletal-related events (%)

groups, and most patients had scores of less than 4.
The lowest proportion of patients with scores of less than 60
4 was at the last visit in the 4-week group. Pain at rest and
50
on movement were both positively correlated with
analgesic use (0·50 for both). 40
Most patients received the protocol-specified number of
30
doses: 155 of 209 (74%) in the 12-week group received four
doses (cumulative exposure 16 mg) and 136 of 216 (63%) 20
in the 4-week group received 12 doses (cumulative
exposure 48 mg). Adverse events were mainly deemed 10 12-week group
4-week group
related to concomitant treatments— eg, chemotherapy—
0
or from complications of the underlying cancer (table 3). 0 3 6 9 12
Adverse events more common in the 4-week group than Time from randomisation (months)
Number at risk
in the 12-week group (184/216 [85%] vs 159/209 [76%]), 12-week group 209 193 163 137 15
probably driven by the significantly higher incidence of 4-week group 216 198 171 143 11
gastrointestinal events in the 4-week group (91/216 [42%]
Figure 2: Kaplan-Meier estimate of time to first on-study skeletal-related event in the intention-to-treat
vs 65/209 [31%]). The most frequent treatment-related
population
adverse event was bone pain (table 3). No events—besides
renal adverse events and osteonecrosis of the jaw—were
deemed related to zoledronic acid. Gastrointestinal We assessed concentration of N-terminal telopeptide in
disorders were also common and included nausea, 220 patients (111 in the 12-week group, 109 in the 4-week
vomiting, abdominal pain, diarrhoea, and constipation. group). At baseline, median concentrations were low
Such adverse events are known to occur during treatment (9·55 nmol/L in the 12-week group and 9·60 nmol/L in
of bone-metastatic malignancies with zoledronic acid. the 4-week group). In the 12-week group, N-terminal
Grade 3 or 4 adverse events occurred in just under half the telopeptide concentration increased at 3 months and all
patients in each group, 28% of which were malignant points thereafter, indicating partial recovery of bone
neoplasm progression (table 3). turnover (figure 3). In the 4-week group, N-terminal
21 of 209 (10%) patients in the 12-week group versus telopeptide concentration did not change, suggesting
29 of 216 (13%) in the 4-week group reported serious persistent inhibition of bone turnover with 4-weekly
adverse events. No dose reductions were reported. Two zoledronic acid (figure 3). Values for premature
of 209 participants (1·0%) in the 12-week group and discontinuations for end-of-study visit were excluded
nine of 216 (4·2%) in the 4-week group discontinued from the analysis because prematurely discontinued
zoledronic acid because of a serious or clinically sig- patients were not replaced.
nificant adverse event. No treatment-related deaths Median percentage change from baseline in N-terminal
were reported. telopeptide concentration was significantly lower in the
Adverse events of interest—including renal adverse 4-week group versus the 12-week group at 6 months
events and osteonecrosis of the jaw—were rare in both (–2·3% vs 12·2%; p=0·0111), 9 months (–2·2% vs 10·6%;
groups. Renal adverse events occurred in one patient p=0·0472), and 12 months (0·0% vs 12·2%; p=0·0465).
(<1%) in the 12-week group and two patients (1%) in the Because few patients had skeletal-related events, the
4-week group. One patient (<1%) in the 4-week group had relation between skeletal morbidity rate and N-terminal
non-serious acute grade 1 renal failure. A post-hoc telopepetide concentration was not formally analysed.
analysis of deterioration of renal function did not show However, available data did not show any obvious pattern.
any clinically meaningful differences between groups
(data not shown). More than 90% of patients in both Discussion
groups had normal serum creatinine concentrations Our findings indicate that a 12-weekly zoledronic acid
throughout the study, and no grade 3 or 4 increases in regimen was not inferior to a 4-weekly schedule in patients
serum creatinine concentration were reported. Overall, with bone-metastatic breast cancer who had previously
seven cases of osteonecrosis of the jaw were reported completed 1 year of on-label treatment with zoledronic
(four in the 12-week group [2%] vs three [1%] in the acid. The skeletal morbidity rate was low and much the
4-week group). All were reported as serious adverse same in both groups. However, one should note that,
events, as per the protocol. although significant, the proximity of the upper limit of

www.thelancet.com/oncology Vol 14 June 2013 667

 Articles

The skeletal-morbidity rate we report is much lower than

12-week group 4-week group
(n=209) (n=216) the assumed rate of 0·91 based on data from previous
trials.10,11 This discrepancy can be attributed in part to
Any 159 (76%) 184 (85%)
differences in concomitant cancer treatments and baseline
Treatment-related adverse events occurring in ≥5% of participants
characteristics. First, new treatments for metastatic breast
Bone pain 56 (27%) 65 (30%)
cancer (eg, taxanes) were available when ZOOM started,
Nausea 24 (11%) 33 (15%)
which could have delayed disease progression and
Asthenia 18 (9%) 33 (15%)
contributed to the lower-than-expected skeletal morbidity
Pyrexia 22 (11%) 28 (13%)
rate. By contrast, the treatment groups in our study were
Vomiting 14 (7%) 23 (11%)
very homogeneous with respect to previous and
Upper abdominal pain 16 (8%) 20 (9%)
concomitant cancer treatments and the results could not
Anaemia 13 (6%) 22 (10%) have been biased by any difference between the two
Diarrhoea 12 (6%) 17 (8%) groups. Second, the timing of initiation of zoledronic acid
Headache 14 (7%) 15 (7%) and history of skeletal-related events varied between ours
Neutropenia 11 (5%) 18 (8%) and previous studies. None of the patients in a study by
Abdominal pain 12 (6%) 15 (7%) Rosen and colleagues10 had previously been treated with
Constipation 12 (6%) 15 (7%) bisphosphonates, and more than two-thirds had a history
Cough 14 (7%) 12 (6%) of skeletal-related events. In the ZOOM trial, all patients
Pain 11 (5%) 15 (7%) had received 1 year of on-label treatment with zoledronic
Dyspnoea 10 (5%) 15 (7%) acid before enrolment, and their skeletal-related event
Arthralgia 9 (4%) 13 (6%) history was described retrospectively. Previous skeletal-
Fatigue 10 (5%) 12 (6%) related events are an important characteristic because
Back pain 7 (3%) 13 (6%) subsequent skeletal-related events occur more often after
Increased γ-glutamyltransferase 8 (4%) 12 (6%) the first event.26 Therefore, the timing of when
Pain in extremity 8 (4%) 12 (6%) bisphosphonate treatment is started is important.
Paraesthesia 11 (5%) 8 (4%) American Society of Clinical Oncology guidelines strongly
Anorexia 5 (2%) 12 (6%) recommend starting bisphosphonate treatment at first
Rash 3 (1%) 12 (6%) radiological evidence of bone metastases, even in the
Vertigo 2 (1%) 11 (5%) absence of symptoms.18,19 The rationale for early treatment
CTCAE grade 3 or 4 92 (44%) 95 (44%) is that bone metastasis—even when asymptomatic and of
CTCAE grade 3 or 4 occurring in ≥1% of participants small size—is an incurable disease process that continually
Malignant neoplasm progression 58 (28%) 60 (28%) disturbs bone remodelling, weakens bone, and can result
Anaemia 2 (1%) 6 (3%) in one or more skeletal complications, which negatively
Neutropenia and febrile neutropenia 5 (2%) 13 (6%) affect quality of life and increase the risk of subsequent
Asthaenia 4 (2%) 5 (2%) skeletal-related events. Moreover, retrospective analyses of
Pain 4 (2%) 2 (1%) large health claims and Spanish population databases
Increased γ-glutamyltransferase 1 (<1%) 7 (3%)
showed that early initiation of treatment with zoledronic
Tumour marker increase 4 (2%) 1 (<1%)
acid correlates with improved clinical outcome compared
Bone pain 11 (5%) 13 (6%)
with delayed treatment.14,27 In our study, 57% of patients in
each treatment group had a skeletal-related event during
Osteonecrosis 3 (1%) 3 (1%)
the period of monthly zoledronic acid treatment that
Dyspnoea 3 (1%) 2 (1%)
preceded study entry.
Serious adverse events 21 (10%) 29 (13%)
Baseline concentrations of N-terminal telopeptide were
Data are n (%). CTCAE=common terminology criteria for adverse events. similarly low in both treatment groups, consistent with the
antiresorptive action of standard zoledronic acid dosing
Table 3: Adverse events
and concentrations reported in postmenopausal women
receiving oestrogen-replacement treatment.28 The
the one-tailed 97·5% CI to the adjusted non-inferiority continued suppression of N-terminal telopeptide concen-
margin suggests that the non-inferiority of 12-weekly ver- trations during treatment in the 4-week group accords with
sus 4-weekly zoledronic acid dosing might not be robust. the pharmacokinetic and pharmacodynamic profile of
All the statistical models used to assess efficacy zoledronic acid. However, the increased concentrations of
produced consistent results. The result of the negative N-terminal telopeptide after baseline in the 12-week group
binomial model accords with the ANCOVA, confirming suggest that accumulation of zoledronic acid in the
that the results for the primary outcome were very close skeleton does not ensure extended suppression of bone
in the two groups, but within the non-inferiority margin. turnover as reported in postmenopausal women with
No significant between-group differences occurred for osteoporosis.22,29 The recovery of bone turnover is probably
any of the secondary efficacy endpoints. a result of the rapid removal of stored zoledronic acid

668 www.thelancet.com/oncology Vol 14 June 2013

 Articles

through increased osteoclast activity at sites of active bone 80 12-week group

lesions, which could hamper the therapeutic effect. In 4-week group
view of the association between maintenance of
suppression of N-terminal telopeptide and reduced risks 60
of skeletal-related events,30 these data raise concerns about

Change in N-terminal telopeptide (%)

long-term reductions of the skeletal morbidity rate with
reduced-frequency zoledronic acid. These concerns are 40
particularly important because patients with bone-
metastatic breast cancer are now surviving longer now
20
than they have in the past. Results from the phase 3
* * *
BISMARK trial,31 comparing a marker-directed zoledronic
acid regimen (using urinary N-terminal telopeptide 0
measurements) to on-label zoledronic acid in patients with
bone-metastatic breast cancer, suggest that zoledronic acid
might be needed every 3–4 weeks when bone resorption –20
markers are increased, and that delaying a more intensive
zoledronic acid schedule until N-terminal telopeptide
concentrations increase might not be the best course of –40
0 3 6 9 12
action, particularly in patients with skeletal-related events.
Time (months)
Zoledronic acid was well-tolerated. Adverse events Number with assessable data
were consistent with the known safety profile of 12-week group 85 79 71 68
4-week group 89 79 64 65
zoledronic acid, and were as expected for this study
population taking the drug long-term. We report no new Figure 3: Median change in N-terminal telopeptide concentration
or unexpected safety findings, and no treatment-related Bars are IQRs. *p<0·05 by Wilcoxon rank-sum test.
deaths. The incidence of renal adverse events was low
and similar between treatment groups, which might be
Panel: Research in context
because we monitored renal safety according to the
drug’s prescribing information. Extension of treatment Systematic review
intervals did not decrease the occurrence of When we proposed the study, we were not aware of any phase 2 or 3 published trials that
osteonecrosis of the jaw. This fact would not exclude the investigated different schedules of zoledronic acid for treatment of bone metastases from
possibility of keeping 4-weekly zoledronic acid as the breast cancer. We searched Medline with the terms “bone metastases [AND] breast cancer
standard care. However, the study was not long enough [AND] zoledronic acid” for reports published in English, between March, 1996, and June,
to detect an increase in osteonecrosis of the jaw, which 2005, and found no evidence of work on this topic. American Society of Clinical Oncology
might be expected after 2 years of treatment with guidelines recommend continuing treatment with biphosphonates in patients with
zoledronic acid. No long-term follow-up was planned breast cancer with bone metastases until there is evidence of a substantial decline in the
for this study; longer studies might be needed to show patient’s general health. These recommendations are based on the fact that multiple
any benefit from 12-weekly zoledronic acid on skeletal complications can occur repeatedly over the course of disease and that
occurrence of osteonecrosis of the jaw. bisphosphonates continue to reduce these events in the second year of treatment.
The ZOOM trial has some limitations. First, this study Another consideration is that bisphosphonates accumulate in the bone and the
was open-label with differences in frequency of clinic pharmacological effect might last for a period after treatment has been discontinued.
visits between groups, which could introduce a reporting Therefore, it might be possible to administer zoledronic acid less frequently provided that
bias for adverse events in the 4-week group. Second, we at least 1 year of treatment with the standard schedule has been completed.
did not collect data about grading or hormone receptor
Interpretation
status. Third, we did not prespecify imaging frequency,
To our knowledge, the ZOOM trial is the first randomised trial to show that outcomes of
which might have reduced the detection of skeletal-
12-weekly zoledronic acid are not inferior to those for 4-weekly zoledronic acid for patients
related events, especially asymptomatic pathological
previously treated with the drug every month for 12–15 months. These results could have a
fractures. Fourth, this study was done in a single country
favourable effect on both patient management and health-care costs. However, despite
and the local standard of care may have affected the
positive results for skeletal-related events, the differences in bone-marker concentrations
results. Finally, although the predefined non-inferiority
between the two treatment groups, along with some minor trial weaknesses, require
margin was based on previous evidence, it might seem
further assessment before a 12-weekly schedule can be recommended.
large for a non-inferiority trial. However, our results led
us to adjust the non-inferiority margin to the much more
restrictive level of 0·19. Studies are needed to elucidate the roles of bisphos-
The approval of denosumab has raised the question of phonates and RANK ligand inhibitors in the manage-
whether bisphosphonates or RANK ligand inhibitors are ment of bone metastases.
a better option for prevention of skeletal-related events To our knowledge, the ZOOM trial provides the first
not only in breast cancer, but also in prostate cancer. evidence for reduced-frequency zoledronic acid dosing in

www.thelancet.com/oncology Vol 14 June 2013 669

 Articles

patients with breast cancer who have completed 1 year of 12 Saad F, Gleason DM, Murray R, et al. Long-term efficacy of
monthly treatment, started at diagnosis of bone metastasis zoledronic acid for the prevention of skeletal complications in
patients with metastatic hormone-refractory prostate cancer.
(panel). Ongoing phase 3 trials with a similar design and J Natl Cancer Inst 2004; 96: 879–82.
patient population as ZOOM (eg, OPTIMIZE-2) might 13 De la Haba J, Rodriguez-Lescure A, Baena JM, et al. Final analysis
provide further information about outcomes with reduced- of dose-density with zoledronic acid treatment in metastatic breast
cancer patients: ZARAS study. Proc Am Soc Clin Oncol 2011;
frequency zoledronic acid dosing after at least 1 year of 29 (suppl): abstr 630.
standard treatment. As cancer treatments continue to 14 Hatoum HT, Lin SJ, Smith MR, et al. Zoledronic acid and skeletal
prolong survival and improve quality of life, assessments complications in patients with solid tumors and bone metastases:
analysis of a national medical claims database. Cancer 2008;
should be done of the efficacy and safety of a reduced- 113: 1438–45.
frequency zoledronic acid dosing regimen after the second 15 Duck L, Delforge M, Doyen C, et al. Long term treatment with
year of on-label zoledronic acid treatment. These long- zoledronic acid (≥ 2 years) in patients with metastatic bone disease
or multiple myeloma: 6-month results from the LOTUZ study.
term studies should also explore the relation between bone Proc Am Soc Clin Oncol 2009; 27 (suppl): poster N20.
turnover markers and skeletal-related events. 16 Henk H, Teitelbaum A, Kaura S. Evaluation of the clinical benefit
Contributors of long-term (beyond 2 years) treatment of skeletal-related events in
advanced cancers with zoledronic acid. Curr Med Res Opin 2012;
DA and TI designed the study. DA, MA, BA, LG, TI, GF, FG, and CIR
28: 1119–27.
enrolled patients and collected data. All authors interpreted data, wrote
17 Morgan GJ, Davies FE, Gregory WM, et al. Effects of induction and
the report, and approved the final version of the Article.
maintenance plus long-term bisphosphonates on bone disease in
Conflicts of interest patients with multiple myeloma: the Medical Research Council
RR and PB are employees of Novartis. DA has served as a consultant to Myeloma IX Trial. Blood 2012; 119: 5374–83.
Italfarmaco, Eisai, and GlaxoSmithKine. TI has served as a consultant to 18 Hillner BE, Ingle JN, Chlebowski RT, et al. American Society of
Novartis, Amgen, and Ipsen. DS has received payment for lectures from Clinical Oncology 2003 update on the role of bisphosphonates and
Novartis, Roche, Janssen, AstraZeneca, Amgen, and Pfizer. LG has bone health issues in women with breast cancer. J Clin Oncol 2003;
served as a consultant to GlaxoSmithKline, Roche, and Novartis. FB has 21: 4042–57.
served as a consultant to Amgen, Dompè, Eli Lilly, and Abiogen. All 19 Van Poznak CH, Temin S, Yee GC, et al. American Society
of Clinical Oncology executive summary of the clinical practice
other authors declare that they have no conflicts of interest.
guideline update on the role of bone-modifying agents in metastatic
Acknowledgments breast cancer. J Clin Oncol 2011; 29: 1221–27.
Novartis Pharmaceuticals provided financial support for medical editorial 20 Aapro M, Abrahamsson PA, Body JJ, et al. Guidance on the use
assistance. We thank Marithea Goberville (ProEd Communications) for of bisphosphonates in solid tumours: recommendations of an
editorial assistance. international expert panel. Ann Oncol 2008; 19: 420–32.
21 Hortobagyi GN. Moving into the future: treatment of bone
References metastases and beyond. Cancer Treat Rev 2005; 31: 9–18.
1 Coleman RE. Bisphosphonates: clinical experience. Oncologist 2004;
22 Black DM, Reid IR, Boonen S, et al. The effect of 3 versus 6 years of
9: 14–27.
zoledronic acid treatment of osteoporosis: a randomized extension
2 Bussard KM, Gay CV, Mastro AM. The bone microenvironment in to the HORIZON-Pivotal Fracture Trial (PFT). J Bone Miner Res
metastasis; what is special about bone? Cancer Metastasis Rev 2008; 2012; 27: 243–54.
27: 41–55.
23 Genant HK, Wu CY, van Kuijk C, Nevitt MC. Vertebral fracture
3 Coleman RE. Skeletal complications of malignancy. Cancer 1997; assessment using a semiquantitative technique. J Bone Miner Res
80: 1588–94. 1993; 8: 1137–48.
4 Lipton A, Theriault RL, Hortobagyi GN, et al. Pamidronate prevents 24 Caraceni A, Cherny N, Fainsinger R, et al. Pain measurement tools
skeletal complications and is effective palliative treatment in and methods in clinical research in palliative care: recommendations
women with breast carcinoma and osteolytic bone metastases: long of an expert working group of the European Association of Palliative
term follow-up of two randomized, placebo-controlled trials. Cancer Care. J Pain Symptom Manage 2002; 23: 239–55.
2000; 88: 1082–90.
25 Andersen PK, Gill RD. Cox’s regression model for counting
5 Weinfurt KP, Castel LD, Li Y, Timbie JW, Glendenning GA, processes: a large sample study. Ann Stat 1982; 10: 1100–20.
Schulman KA. Health-related quality of life among patients with
26 Major PP, Cook R. Efficacy of bisphosphonates in the management
breast cancer receiving zoledronic acid or pamidronate disodium
of skeletal complications of bone metastases and selection of
for metastatic bone lesions. Med Care 2004; 42: 164–75.
clinical endpoints. Am J Clin Oncol 2002; 25: S10–18.
6 Costa L, Major PP. Effect of bisphosphonates on pain and quality
27 Rodriguez-Lescure A, de la Haba J, Baena JM, et al. Second interim
of life in patients with bone metastases. Nat Clin Pract Oncol 2009;
analysis of dose density with zoledronic acid treatment in metastatic
6: 163–74.
breast cancer patients. ZARAS study. 12th St Gallen International
7 Saad F, Lipton A, Cook R, Chen YM, Smith M, Coleman R. Breast Cancer Conference; St Gallen, Switzerland; March 16–19,
Pathologic fractures correlate with reduced survival in patients with 2011. Abstr P323.
malignant bone disease. Cancer 2007; 110: 1860–67.
28 Scariano JK, Glew RH, Bou-Serhal CE, Clemens JD, Garry PJ,
8 Zometa (zoledronic acid) injection (package insert). East Hanover: Baumgartner RN. Serum levels of cross-linked N-telopeptides and
Novartis Pharmaceuticals Corporation, 2009. http://www. aminoterminal propeptides of type I collagen indicate low bone
accessdata.fda.gov/drugsatfda_docs/label/2009/021223s018lbl.pdf mineral density in elderly women. Bone 1998; 23: 471–77.
(accessed Jan 11, 2012).
29 Grey A, Bolland MJ, Horne A, et al. Five years of anti-resorptive
9 Holen I, Coleman RE. Bisphosphonates as treatment of bone activity after a single dose of zoledronate—results from a randomized
metastases. Curr Pharm Des 2010; 16: 1262–71. double-blind placebo-controlled trial. Bone 2012; 50: 1389–93.
10 Rosen LS, Gordon D, Kaminski M, et al. Zoledronic acid versus 30 Lipton A, Cook R, Saad F, et al. Normalization of bone markers is
pamidronate in the treatment of skeletal metastases in patients with associated with improved survival in patients with bone metastases
breast cancer or osteolytic lesions of multiple myeloma: a phase III, from solid tumors and elevated bone resorption receiving
double-blind, comparative trial. Cancer J 2001; 7: 377–87. zoledronic acid. Cancer 2008; 113: 193–201.
11 Rosen LS, Gordon D, Kaminski M, et al. Long-term efficacy and 31 Coleman RE, Wright J, Houston S, et al. Randomized trial of
safety of zoledronic acid compared with pamidronate disodium in marker-directed versus standard schedule of zoledronic acid for
the treatment of skeletal complications in patients with advanced bone metastases from breast cancer. Proc Am Soc Clin Oncol 2012;
multiple myeloma or breast carcinoma: a randomized, double-blind, 30 (suppl): abstr 511.
multicenter, comparative trial. Cancer 2003; 98: 1735–44.

670 www.thelancet.com/oncology Vol 14 June 2013