CLINICAL TRIALS JBMR

Arzoxifene for Prevention of Fractures and Invasive

Breast Cancer in Postmenopausal Women
Steven R Cummings, 1 Michael McClung, 2 Jean-Yves Reginster, 3 David Cox, 4 Bruce Mitlak, 4 John Stock, 4
Messan Amewou-Atisso , 4 Trevor Powles , 5 Paul Miller , 6 José Zanchetta , 7 and Claus Christiansen8
1
San Francisco Coordinating Center, California Pacific Medical Center Research Institute and the University of California,
San Francisco, CA, USA
2
Oregon Osteoporosis Center, Portland, OR, USA
3
CHU Centre Ville and Sart-Tilman, University of Liège, Liège, Belgium
4
Eli Lilly and Company, Indianapolis, IN, USA
5
Parkside Oncology Clinic, London, United Kingdom
6
Colorado Center for Bone Research, Lakewood, CO, USA
7
Instituto de Investigaciones Metabólicas, Buenos Aires, Argentina
8
Nordic Bioscience A/S, Herlev, Denmark

ABSTRACT
Arzoxifene is a selective estrogen receptor modulator (SERM) that has been shown to be more potent in preclinical testing than currently
available agents. Its effects on clinical outcomes are not known. In a randomized, blinded trial, women aged 60 to 85 years with
osteoporosis, defined as a femoral neck or lumbar spine bone mineral density T-score of
2.5 or less or a vertebral fracture, and women
with low bone mass, defined as a bone density T-score of
1.0 or less and above
2.5, were assigned to arzoxifene 20 mg or placebo
daily. The primary endpoints were new vertebral fracture in those with osteoporosis and invasive breast cancer in the overall population.
After 3 years, the cumulative incidence of vertebral fractures in patients with osteoporosis was 2.3% lower in the arzoxifene group than
in the placebo group, a 41% relative risk reduction [95% confidence interval (CI) 0.45–0.77, p < .001]. In the overall population, the
cumulative incidence of invasive breast cancer over 4 years was reduced by 1.3%, with a 56% relative reduction in risk (hazard
ratio ¼ 0.44, 95% CI 0.26–0.76, p < .001); there was no significant decrease in nonvertebral fracture risk. Arzoxifene increased the
cumulative incidence of venous thromboembolic events by 0.7%, with a 2.3-fold relative increase (95% CI 1.5–3.7). Like other SERMs,
arzoxifene decreased vertebral fractures and invasive breast cancer while the risk of venous thromboembolic events increased. ß 2011
American Society for Bone and Mineral Research.

KEY WORDS: OSTEOPOROSIS; FRACTURE; VERTEBRAL FRACTURE; BREAST CANCER; ARZOXIFENE; SERMS; RANDOMIZED; CONTROLLED TRIAL

Introduction potent than raloxifene at decreasing bone resorption and

improving bone mass in postmenopausal women.(4,5) Based on

S elective estrogen receptor modulators (SERMs) have been

developed with an aim to reduce the risk of major conditions
that account for much of the morbidity and mortality in
these initial results, the Generations Trial was designed to test
whether arzoxifene 20 mg/day would safely reduce the risk of
fractures and invasive breast cancer in postmenopausal women
postmenopausal women: vertebral and nonvertebral fractures, with low bone mass or osteoporosis.
breast cancer, and cardiovascular disease. Available SERMs have
important limitations. Tamoxifen use is limited to treatment
and prevention of breast cancer, and it increases the risk of Methods
endometrial cancer,(1) whereas raloxifene decreases the risk of
Study design
vertebral fractures and invasive breast cancer but has not
been shown to reduce the risk of nonvertebral fractures or The Generations Trial was an international, randomized, placebo-
cardiovascular disease.(2,3) Arzoxifene is a SERM that is more controlled trial testing the hypotheses that arzoxifene 20 mg/day

Received in original form April 5, 2010; revised form June 29, 2010; accepted July 14, 2010. Published online July 23, 2010.
Address correspondence to: Steven R Cummings, MD, San Francisco Coordinating Center, Suite 5700, 185 Berry Street, San Francisco, CA 94107, USA.
E-mail: scummings@sfcc-cpmc.net
Journal of Bone and Mineral Research, Vol. 26, No. 2, February 2011, pp 397–404
DOI: 10.1002/jbmr.191
ß 2011 American Society for Bone and Mineral Research

397
would reduce the risk of new vertebral fractures in postmeno-
pausal women with osteoporosis and invasive breast cancer in
women with osteoporosis or low bone mass. Equipoise about the
comparison of arzoxifene and placebo was based on the fact that
it is not possible to predict the effect of SERMs on hip and
nonvertebral fractures, endometrial cancer, and cardiovascular
disease and that they sometimes have unexpected effects, so the
balance of benefits and risk can be established only by a placebo-
controlled trial.(6–8) All patients were counseled about their bone
density and presence of a vertebral fracture and were informed
about alternative treatments. Patients gave written informed
consent, and the protocol was approved by the institutional
review boards at each investigative site.
Patients
Women 60 to 85 years of age were enrolled in stratum A if they
had a femoral neck or lumbar spine bone mineral density (BMD)
T-score of
2.5 or less or in Stratum B if their femoral neck and
lumbar spine T-scores both were
1.0 or less and above
2.5.
Women with a vertebral fracture were assigned to stratum A
regardless of their BMD. Approximately 10% of women who had
been randomized into stratum B were found subsequently to
have a vertebral fracture. These women were combined with
women in stratum A for analysis of women ‘‘with osteoporosis,’’
and the remainder of subjects in stratum B were analyzed as
‘‘women with low bone mass.’’
Women were excluded if they had conditions that influence
bone metabolism: unexplained or abnormal vaginal bleeding
within 6 months; a history of breast cancer or estrogen-
dependent neoplasia; any history of venous thromboembolism,
stroke, or transient ischemic attack; liver disease; impaired kidney
function; a 25-hydroxyvitamin D level of less than 10 ng/mL;
endocrine disorders besides type 2 diabetes and hypothyroidism
that required pharmacologic therapy; and bisphosphonate
therapy in the previous year if treatment had been less than
1 year or within 3 years if bisphosphonate therapy had lasted
more than 1 year, use of aromatase inhibitors, estrogens,
calcitonin, or SERMs within the previous 3 months, sodium
fluoride within the past 6 months, or more than 1 month of
systemic corticosteroids in the previous 6 months. Use of vaginal
estrogen was allowed at the lowest possible dose and for a total
duration not exceeding 3 months.
Treatments
All patients received daily supplements containing approxi-
mately 500 mg of elemental calcium and 400 to 600 IU of vitamin
D. Patients were randomly assigned to tablets of 20 mg
arzoxifene or an identical placebo daily. Compliance with
treatment was the number of tablets taken (number of tablets
dispensed minus number of tablets returned) as a percentage of
the number of tablets dispensed.
Any patient who developed a vertebral fracture or who lost
more than 6% BMD at the spine or 9% at the femoral neck in 1
year or 13% or 15%, respectively, by year 3 was referred to her
physician with this information to consider alternative treatment
for osteoporosis. Those treated with nonhormonal therapy, such
as a bisphosphonate, were allowed to continue taking their study
398 Journal of Bone and Mineral Research
medication, whereas those who started a hormone therapy,
such as estrogen, discontinued the study medication. The study
drug was discontinued if the patient developed invasive or
noninvasive breast cancer, venous thromboembolism, endome-
trial hyperplasia or cancer, or used estrogen-containing products
or SERMs. Those who stopped taking the study drug were
encouraged to continue participation in the trial.
Endpoints
The primary endpoints were new vertebral fractures in patients
with osteoporosis and invasive breast cancer in women with
either osteoporosis or low bone mass. Nonvertebral fractures in
women with osteoporosis and vertebral fracture in women with
low bone mass were secondary endpoints. Secondary endpoints
for the overall population included coronary heart disease,
stroke, clinical vertebral fracture, cognitive function, and hip
fracture. These endpoints were analyzed in this order; statistical
testing continued with the following endpoint only if the effect
with arzoxifene 20 mg/day on the preceding endpoint was
statistically significantly greater than with placebo. Additional
endpoints included back pain, days of disability, and all clinical
fractures.
Adjudication committees that were blinded to treatment
assignments confirmed breast cancers [including invasive, ductal
carcinoma in situ (DCIS), lobular carcinoma in situ (LCIS), and
atypical ductal hyperplasia], nonvertebral fractures, endometrial
cancers and endometrial hyperplasia, coronary events, stroke,
and venous thromboembolism from medical records, reports of
imaging studies, and histologic reports.
For women in stratum A, new vertebral fractures were
assessed by lateral spinal radiographs at 6, 12, 24, and 36 months;
for patients in stratum B, lateral spinal radiographs were assessed
at 36 months. Spinal radiographs were assessed independently
by two readers at a central facility (Synarc) using a semiquanti-
tative scale from normal (grade 0) to severe (grade 3).(9) A new
vertebral fracture was defined as a change from normal to grade
of 1 to 3 on any subsequent radiograph. If the readers disagreed,
the result was based on the grading of a third reader.
Nonvertebral fractures were confirmed by review of a report
by a radiologist or by review of the original radiograph.
Pathologic fractures, as well as fractures of fingers, toes, skull, or
facial bones, were excluded from the analysis because they are
not associated with low bone density.(10) We did not exclude
fractures attributed to a high degree of trauma.(11)
BMD at the proximal femur and lumbar spine was measured
by dual-energy X-ray absorptiometry (DXA) and at 12, 24, 36, and
60 months in all patients and additionally at 3 and 6 months in a
subset of 1022 women with osteoporosis. Levels of N-terminal
propeptide of type 1 procollagen (P1NP; Orion Diagnostica,
Espoo, Finland), a marker of bone formation, and CTX (Crosslaps
ELISA Nordic Bioscience Diagnostics AS, Herlev, Denmark), a
marker of bone resorption, were measured in fasting samples of
serum at baseline, 3, 6, and 12 months in the same subset of
women.
Bilateral mammograms and breast examinations were
performed at screening and yearly thereafter and at early
termination if more than 6 months had passed since the previous
CUMMINGS ET AL.
mammogram. Cardiovascular events consisted of strokes and
coronary events, including cardiac death (owing to acute
myocardial infarction, sudden cardiac death, or death due to
a coronary artery procedure), nonfatal myocardial infarction,
hospitalized acute coronary syndromes other than myocardial
infarction, and myocardial revascularization. Venous throm-
boembolic events included deep vein thrombosis, pulmonary
embolus, retinal vein thrombosis, and superficial venous
thromboses treated with systemic anticoagulants. Levels of
low-density lipoprotein (LDL) and high-density lipoprotein (HDL)
cholesterol and high-sensitivity C-reactive protein were mea-
sured in fasting samples in a subset of 365 women at baseline
and at 6, 12, and 36 months. Women who reported vaginal
bleeding underwent endometrial sampling; biopsies were read
centrally according to World Health Organization (WHO)
criteria.(12)
Adverse events
Adverse events were classified according to the Medical
Dictionary for Regulatory Activities (MedDRA) system. As pro-
spectively planned, all serious adverse events with p < .05 for
differences between treatment groups are reported. Other
adverse events are reported if they occurred with at least 1%
difference between treatment groups and with p < .01 for the
difference between groups.
Statistical analysis
A sample size of 2275 per treatment group and per stratum
provided a 94% power to detect a 50% relative reduction in the
risk of new vertebral fracture at a two-sided a of 0.01 and 90%
power to detect a 50% relative reduction in the risk of new
invasive breast cancer at a two-sided a of 0.04.
Comparisons of baseline characteristics between treatment
groups used a one-way ANOVA for continuous variables and a
Pearson’s chi-square test for categorical variables. Primary
analysis of vertebral fracture was based on data available at
36 months, with a further analysis based on 48-month data. All
other analyses were based on data available at 48 months.
All analyses were based on the intention to treat. Comparisons
between treatment groups used Pearson’s chi-square test or a
Fisher’s exact test for the percentages with vertebral fracture.
Cox proportional hazards models were used to calculate hazard
ratios for nonvertebral fracture, all clinical fractures, and invasive
breast cancer. Before the study results were known, several
subgroup analyses were planned for vertebral and nonvertebral
fracture: ethnicity, diagnosis of osteoporosis versus low bone
mass, geographic region, baseline presence of vertebral fracture,
maximum severity of the prevalent vertebral fractures, and
lumbar spine BMD and femoral neck BMD above and below T-
scores of
2.5. Differences in the change in BMD between the
groups were assessed using analysis of covariance (ANCOVA)
models with a fixed effect for treatment. For biochemical markers
of bone metabolism, differences in within- and between-
treatment-group changes were assessed using the signed-rank
test and Wilcoxon rank-sum tests, respectively. Difference in the
percentages of women with adverse events were analyzed using
Pearson’s chi-square test or a Fisher’s exact test.
ARZOXIFENE
Analyses and writing
The sponsor designed and oversaw execution of the protocol
and was responsible for management and quality control of
data. An independent data-monitoring committee reviewed
unblinded data at least annually. The sponsor performed the
analyses, and the authors received all analyses that they
requested. An external author (SC) planned the analyses for
the article before treatment assignment was unblinded and
wrote the first draft of the ‘‘Results’’ and ‘‘Discussion.’’ The
authors who were not employed by the sponsor were not paid
for their work. They approved the manuscript for publication and
vouch for the manuscript’s accuracy.
Results
A total of 232 sites in 23 countries enrolled 9354 patients, 4678
with osteoporosis and 4676 with low bone mass (Table 1).
Assessment of outcomes at 4 years was completed for 78% in
each group. Most of the subjects who discontinued, 12.8%
overall, did so for ‘‘personal reasons.’’ Only 3.9% of subjects in the
placebo group and 2.9% in the arzoxifene group received a
bisphosphonate during the trial.
Fractures, bone density, and bone turnover
In patients with osteoporosis, arzoxifene was associated with a
2.3% ( p < .001) absolute reduction in the 3-year cumulative
incidence of radiographic vertebral fractures, the primary
outcome, representing a 41% reduction in relative risk [95%
confidence interval (CI) 0.45–0.77, p < .001]. The reduction in risk
was similar over 4 years and in each year of treatment (Fig. 1).
Treatment with arzoxifene did not reduce the risk of non-
vertebral fractures: Overall, the cumulative incidence was 7.6% in
the placebo group and 7.1% in the arzoxifene group [hazard ratio
(HR) ¼ 0.94, 95% CI 0.81–1.10). Results were similar for women
with either osteoporosis or low bone mass (Table 2). There were
no statistically significant reductions in risk of nonvertebral
fractures for any of the predefined subgroups defined by age,
baseline presence or severity of vertebral fractures, or bone
density; specifically, for subgroups defined by femoral neck BMD
T-score, treatment group HRs were 1.07 (95% CI 0.79–1.45) for
those with a T-score of
2.5 or less and 0.89 (95% CI 0.69–1.15)
for those with a T-score of greater than
2.5 in the overall
population. Reductions in risk of vertebral fracture were similar in
those subgroups (HRs ¼ 0.71 and 0.50, respectively, p ¼ .20 for
the interaction).
At 12 months, women in the arzoxifene group had a 42.1%
(95% CI
46.1 to
38.3) reduction in CTX level and a 33.5% (95%
CI
36.5 to
30.5) reduction in P1NP level compared with
women in the placebo group. Differences in the 3-year changes
in bone density were 2.6% at the total hip, 2.8% at the femoral
neck, and 2.9% at the lumbar spine (Table 3).
Other endpoints
In the overall population, the cumulative rate of invasive breast
cancer was 1.0% with arzoxifene versus 2.3% with placebo, with a
Journal of Bone and Mineral Research 399
Table 1. Baseline Characteristics of Study Patients

Overall Osteoporosis Low bone mass

Arzoxifene Arzoxifene Arzoxifene

Placebo 20 mg/day Placebo 20 mg/day Placebo 20 mg/day
Variable (n ¼ 4678) (n ¼ 4676) (n ¼ 2612) (n ¼ 2640) (n ¼ 2066) (n ¼ 2036)
Region
Europe 1442 (30.8) 1434 (30.7) 657 (25.2) 678 (25.7) 785 (38.0) 756 (37.1)
North America 751 (16.1) 755 (16.2) 259 (9.9) 255 (9.7) 492 (23.8) 500 (24.6)
South America 1845 (39.4) 1845 (39.5) 1272 (48.7) 1272 (48.2) 573 (27.7) 573 (28.1)
Other 640 (13.7) 642 (13.7) 424 (16.2) 435 (16.5) 216 (10.5) 207 (10.2)
Age, years (SD) 67.4 (5.5) 67.5 (5.6) 67.6 (5.6) 67.7 (5.7) 67.2 (5.4) 67.1 (5.4)
BMI, kg/m2 26.7 26.7 26.0 (4.4) 25.9 (4.4) 27.7 (4.6) 27.8 (4.5)
Smoking, n (%) 431 (9.2) 435 (9.3) 249 (9.5) 248 (9.4) 182 (8.8) 187 (9.2)
Alcohol consumption, n (%) 1109 (23.7) 1148 (24.6) 402 (15.4) 409 (15.5) 707 (34.2) 739 (36.3)
Hysterectomy, n (%) 1102 (23.6) 1110 (23.7) 581 (22.2) 590 (22.4) 521 (25.2) 520 (25.5)
Diabetes, n (%) 342 (7.3) 375 (8.02) 204 (7.8) 236 (8.9) 138 (6.7) 139 (6.8)
Preexisting vertebral fracture, n (%) 706 (15.1) 717 (15.3) 706 (27.0) 717 (27.2) 0 0
Bone mineral density
Lumbar spine
Mean (SD), g/cm2 0.86 (0.14) 0.86 (0.14) 0.78 (0.10) 0.78 (0.10) 0.97 (0.10) 0.97 (0.10)
T-score, mean
2.22 (1.18)
2.26 (1.19)
2.93 (0.88)
2.95 (0.92)
1.34 (0.87)
1.37 (0.83)
Femoral neck
Mean (SD), g/cm2 0.71 (0.09) 0.71 (0.09) 0.67 (0.09) 0.67 (0.09) 0.76 (0.07) 0.76 (0.07)
T-score
1.86 (0.70)
1.86 (0.72)
2.15 (0.67)
2.15 (0.70)
1.50 (0.56)
1.49 (0.56)
Total hip
Mean (SD), g/cm2 0.79 (0.11) 0.79 (0.11) 0.74 (0.10) 0.74 (0.11) 0.84 (0.08) 0.84 (0.09)
T-score
1.37 (0.87)
1.37 (0.89)
1.73 (0.84)
1.74 (0.86)
0.92 (0.69)
0.90 (0.69)

56% relative decrease in risk (HR ¼ 0.44, 95% CI 0.26–0.76,
p < .001).
The cumulative 4-year incidence of stroke was not significantly
different in the arzoxifene group (1.0%) versus placebo group
(0.9%), nor was the cumulative incidence of coronary events
(1.5% with arzoxifene versus 1.6% with placebo) (Table 4).
Fig. 1. Percent of study patients with osteoporosis who experienced a
vertebral fracture over the course of the study (3-year findings). Four
humdred and fourteen women with osteoporosis were captured in
stratum B and thus only had lateral spinal radiographs at 36 months.
Findings presented by year of study are for women in stratum A only.
Compared with placebo, arzoxifene decreased LDL cholesterol
levels by 10.2% ( p < .001) with a nonsignificant 2.8% increase in
HDL cholesterol ( p ¼ .10) and a 5.6% decrease in C-reactive
protein level ( p ¼ .37).
Safety
Arzoxifene was associated with a 2.3-fold increase in risk of
venous thromboembolic events (95% CI 1.5–3.7). The absolute
risk difference in the 4-year cumulative incidence was 0.7%
(Table 4). Over 4 years, 9 cases of endometrial cancer were
confirmed with arzoxifene and 4 with placebo ( p ¼ .16). Two of
the endometrial cancers in the arzoxifene group were serous
adenocarcinomas; all others were endometrioid carcinomas.
More cases of uterine polyps occurred in the arzoxifene group
than in the placebo group [65 (1.39%) versus 34 (0.73%),
p ¼ .002]. The cumulative incidence of reports of vaginal
bleeding was similar with placebo (2.8%) and arzoxifene
(3.2%, p ¼ .25). Serious adverse events of acute cholecystits
were more common with arzoxifene (Table 4), and there were
five cases of lung metastases, from diverse cancers, in the
arzoxifene group but none in the placebo group. Five cases of
osteonecrosis at various skeletal sites were reported with
arzoxifene; none involved jaw bones. Four of the five had at
least one risk factor for osteonecrosis: Two were taking
corticosteroids, three had prior trauma at the site, and three
had osteoarthritis.

400 Journal of Bone and Mineral Research CUMMINGS ET AL.

Table 2. Four-Year Fracture Findings by Treatment Group

Population Placebo Arzoxifene 20 mg/day

Relative riskb Rate differencea
a
Endpoint n (%) Rate n (%) Rate (95% CI) p Value (95% CI)
Osteoporosis
Vertebral fracture 139 (5.8) 13.9 82 (3.4) 8.0 0.58 (0.45–0.76) <.001 5.9 (3.0–8.8)
Nonvertebral fracturec 208 (8.0) 20.7 203 (7.0) 20.0 0.96 (0.79–1.17) .71 0.72 (
3.2–4.7)
Low bone mass
Vertebral fracture 40 (2.3) 5.1 27 (1.6) 3.5 0.69 (0.42–1.11) .13 1.6 (
0. 5–3.7)
Nonvertebral fracture 146 (7.1) 18.0 131 (6.4) 16.4 0.91 (0.72–1.16) .45 1.6 (
2.5–5.7)
Overall
Vertebral fracture 179 (4.3) 10.1 109 (2.6) 6.1 0.61 (0.48–0.77) <.001 4.0 (2.1–5.8)
All clinical fractures 388 (8.3) 21.5 375 (8.0) 20.8 0.97 (0.84–1.12) .66 0.7 (
2.3–3.7)
Clinical vertebral fracture 42 (1.0) 2.3 47 (1.1) 2.6 1.12 (0.74–1.69) .60
0.3 (
1.3–0.7)
Nonvertebral fracture 354 (7.6) 19.5 334 (7.1) 18.4 0.94 (0.81–1.10) .45 1.1 (
1.7–3.9)
Hip fracture 26 (0.6) 1.4 20 (0.4) 1.1 0.77 (0.43–1.38) .38 0.32 (
0.4–1.0)
a
Per 1000 woman-years.
b
Risk ratios were calculated for vertebral fractures and hazard ratios for other fracture endpoints.
c
Nonvertebral fractures include hip, wrist, pelvis, humerus, clavicle/ribs, lower leg, femur, patella, sacrum, scapula, sternum, tarsus calcaneus, forearm,
metacarpal, metatarsal, carpal bones, coccyx, and ankle.

Hot flushes, muscle cramps, vaginal discharge, and vulvova-
ginitis were reported more often in women in the arzoxifene
group than in the placebo group (Table 4). Cough, upper
respiratory infections, and pneumonia were reported more
frequently with arzoxifene, and there was a 0.3% absolute
difference in risk of serious adverse events of chronic obstruc-
tive pulmonary disease (Table 4). There were no statistically
significant increases in the incidence of asthma.
Discussion
Treatment with arzoxifene for 4 years significantly reduced the
risk of vertebral fractures in women with osteoporosis and
invasive breast cancer in postmenopausal women with low bone
mass or osteoporosis. It did not reduce the risk of nonvertebral
fractures, and neither did raloxifene and bazedoxifene.(13,14)
Although there have been no direct comparisons, the 2.5% to 3%
improvements in bone density and 30% to 40% reduction in risk
of vertebral fracture are similar to those reported for other
SERMs.(14,15) Lasofoxifene 0.5 mg/day reduced the risk of
nonvertebral fractures, although it also improved bone density
at the hip and spine by only 3% and reduced markers of bone
turnover to a similar degree as arzoxifene.(15) In contrast,
estrogen and tibolone have slightly greater effects on spine (4%
to 6%) and femoral neck (3% to 5%) bone density over 3 years
than do SERMs.(2,8,16) Together, these findings suggest that for
agents acting through the estrogen receptor, a greater effect on
bone density is required to consistently and significantly reduce
the risk of nonvertebral fractures.
Arzoxifene reduced the risk of invasive breast cancer to a
similar degree to that seen with tamoxifen, raloxifene, and
0.5 mg/day of lasofoxifene,(1,3,13,15) suggesting that reduction in
breast cancer may be a class effect of SERMs. However,
significant breast cancer risk reduction with bazedoxifene was
not evident based on the preliminary study results.(17)
The absence of an effect of arzoxifene on cardiovascular
events in postmenopausal women with osteoporosis resembles
that of raloxifene.(3,18) In contrast, conjugated estrogen with
medroxyprogesterone increases the risk of coronary heart
disease and stroke, and tibolone and tamoxifen increase the
risk of stroke(1,6,8); in a recent study of lasofoxifene, the

Table 3. Changes in Biochemical Markers of Bone Resorption (CTX), Bone Formation (P1NP), and Bone Density Between Placebo and
Arzoxifene Groups in a Subset of Women with Osteoporosis

Percent change from baseline

Placebo Arzoxifene 20 mg/day Difference between p Value

(n ¼ 505) (n ¼ 517) groups (95% CI) for difference
CTX, ng/mL, median (interquartile range)
1.5 (
18.0–31.4)
40.6 (
58.7 to
22.0)
42.1 (
46.1 to
38.3) <.001
p ¼ 0.027 p < .001
P1NP, mg/L, median (interquartile range)
6.3 (
21.9–12.5)
41.6 (
52.3 to
22.6)
33.5 (
36.5 to
30.5) <.001
p ¼ .002 p < .001
Lumbar spine BMD, mean (SD) 0.9 (0.2) p < .001 3.7 (0.2) p < 0.001 2.9 (2.3 to 3.4) <.001
Femoral neck BMD, mean (SD)
1.2 (0.2)p < .001 1.7 (0.2) p < .001 2.8 (2.3 to 3.4) <.001
Total hip BMD, mean (SD)
1.2 (0.2) p < .001 1.4 (0.2) p < .001 2.6 (2.2 to 3.1) <.001

ARZOXIFENE Journal of Bone and Mineral Research 401

Table 4. Adverse Events and Safety Findings for the Overall Study Population

Placebo (n ¼ 4678) Arzoxifene (n ¼ 4676) p Value

a
Serious adverse events 1015 (21.7) 1083 (23.2) .09
Death 98 (2.1) 105 (2.3) .62
Cholecystitis acute 3 (0.1) 11 (0.2) .02
Osteonecrosis 0 5 (0.11) .03
Metastases to lung 0 5 (0.11) .03
Chronic obstructive pulmonary disease 9 (0.2) 21 (0.5) .03
Discontinuation of study due to AE 160 (3.4) 185 (4.0) .17
AEs with specified difference between treatment groupsb
Vulvovaginitisc 510 (10.9) 688 (14.7) <.001
Vaginal discharged 95 (2.0) 149 (3.2) <.001
Muscle spasmse 349 (7.5) 657 (14.1) <.001
Constipation 270 (5.8) 336 (7.2) .005
Headache 307 (6.6) 235 (5.0) .001
Upper respiratory tract infection 282 (6.0) 348 (7.4) .006
Cough 263 (5.6) 351 (7.5) <.001
Pneumonia 179 (3.8) 236 (5.1) .004
Hot flush 190 (4.1) 393 (8.4) <.001
Hypercholesterolaemia 260 (5.6) 121 (2.6) <.001
Adjudicated endpointsf
Endometrial cancer 4 (0.1) 9 (0.2) .16
Endometrial hyperplasia 2 (0.0) 4 (0.1) .45
Coronary event or stroke 113 (2.4) 116 (2.5) .83
Coronary event 74 (1.6) 72 (1.5) .88
Stroke (without TIA) 42 (0.9) 47 (1.0) .59
Venous thromboembolismg 27 (0.6) 63 (1.3) <.001
Pulmonary embolism 7 (0.1) 16 (0.3) .06
Deep vein thrombosis 9 (0.2) 26 (0.6) .004
a
Number (%).
b
All adverse events reported in at least 1% difference between treatment groups and with a significant difference between treatment groups at p < .01.
c
Special search category encompassing several MedDRA preferred terms related to vaginal infection.
d
Term included in vulvovaginitis special search category.
e
Majority of muscle spasms were recorded as leg cramps.
f
Not mutually exclusive.
g
Venous thromboembolism includes pulmonary embolism, deep vein thrombosis, retinal vein thrombosis and other venous thrombolic events.

incidences of coronary events and stroke were lower with
lasofoxifene than with placebo.(15) The reasons for these
differences between treatments are not certain; all decrease
LDL cholesterol levels, but their effects on C-reactive protein
(CRP) levels differ: Estrogen and tibolone increase levels of CRP,
raloxifene and arzoxifene have no effect on CRP levels, and
lasofoxifene decreases CRP levels, suggesting that the different
effects of these treatments on risk of cardiovascular disease
might be due in part to differences in their effect on
inflammation.
Treatment with arzoxifene was associated with an increased
risk of venous thromboembolic events, as has been observed for
all agents that have any agonistic effects on the estrogen
receptor. Although uncommon and not reaching statistical
significance, there were more cases of endometrial cancer and
endometrial hyperplasia with arzoxifene. Arzoxifene shares
several other adverse effects with other SERMs, including hot
flushes and muscle cramps. Vaginal discharge and vulvovaginitis
have been reported with some other SERMs, including
lasofoxifene, and tibolone, possibly mediated through estro-
genic effects on the vaginal epithelium.(8,15)
Treatment with arzoxifene was associated with increased
reports of upper respiratory infections, pneumonia, coughing,
and serious adverse events of chronic obstructive pulmonary
disease. Bronchial epithelium and alveolar macrophages express
estrogen receptors,(19,20) and depletion or inhibition of estrogen
receptors increases expression of inflammatory markers in the
lung, including tumor necrosis factor a (TNF-a). Previous trials of
SERMs, estrogen, and tibolone have not reported increases in
the risk of allergies, inflammation, or infection in lungs.
Treatment with arzoxifene was associated with a small increased
risk of metastases to the lung but not primary lung cancer, but
these cases involved diverse types of cancer, and there is no
known biologic basis for pulmonary susceptibility to metastasis,
and therefore, it might be a chance finding.
Estrogen has lithogenic effects on bile,(21) which may explain
the increased risk of cholecystitis with arzoxifene and some other
SERMs.

402 Journal of Bone and Mineral Research CUMMINGS ET AL.

 This trial illustrates the difficulty of predicting the clinical
effects of agents that interact with hormone receptors present in
many tissues. Preclinical models for the effect of treatment on
bone, endometrium, cardiovascular disease, and other condi-
tions may not be reliable indicators of the clinical effects of
treatments on those outcomes and may fail to detect effects on
other systems and conditions.
We conclude that arzoxifene shares many of the character-
istics of other SERMS. These include measurable reduction in
bone resorption, modest increases in BMD, and a decreased risk
of vertebral but not nonvertebral fractures. Like other SERMs,
arzoxifene also reduced the risk of invasive breast cancer but
increased the risk of thromboembolic events and gynecologic
adverse events.
Disclosures
SRC receives consulting fees from Amgen, Eli Lilly and Company,
and Pfizer and honoraria/speaking fees from Novartis and Eli Lilly
and Company. MM receives research grants and consultant fees
from Amgen, Eli Lilly and Company, Merck, Novartis, and Procter
& Gamble and speaker’s bureau fees from Amgen, Eli Lilly and
Company, Novartis, and sanofi-aventis. J-YR receives consultant
and advisory board fees from Servier, Novartis, Negma, Eli Lilly
and Company, Wyeth, Amgen, GlaxoSmithKline, Roche, Merck,
Nycomed, NPS, Theramex, and UCB and lecture fees from Merck
Sharp & Dohme, Eli Lilly and Company, Rottapharm, IBSA,
Genevrier, Novartis, Servier, Roche, GlaxoSmithKline, Teijin, Teva,
Ebewee Pharma, Zodiac, Analis, Theramex, Nycomed, and Novo-
Nordisk and grant support from Bristol Myers Squibb, Merck
Sharp & Dohme, Rottapharm, Teva, Eli Lilly and Company,
Novartis, Roche, GlaxoSmithKline, Amgen, and Servier. DC, BM,
JS, and MA-A are full-time employees and stockholders of Eli Lilly
and Company. TP has a consultancy with Eli Lilly and Company
and Pfizer. PM has recent fundings from Procter & Gamble,
sanofi-aventis, Roche, Eli Lilly and Company, Merck, Novartis, and
Amgen and has served on speaker boards, advisory boards, and
as a consultant for Procter & Gamble, sanofi-aventis, Merck, Eli
Lilly and Company, Amgen, NPS, Novartis, Roche, and
GlaxoSmithKline. JZ receives consulting fees and participates
on paid advisory boards for Amgen, Eli Lilly and Company, Merck,
Pfizer, and Servier. CC is chairman of Nordic Bioscience A/S and
has served as consultant for Roche, Wyeth-Ayerst, Eli Lilly and
Company, Novartis, Novo Nordisk, Proctor & Gamble, Groupe
Fournier, Besins EscoVesco, MSD, Chiesi, Boehringer Mannheim,
and Pfizer.
Acknowledgments
We would like to thank Dr Sherie A Dowsett, Eli Lilly and
Company, for input on scientific content and creation of figure
and tables. The study was supported by Eli Lilly and Company.
The trial is registered at Clinicaltrial.gov number NCT00088010.
The following investigators participated in the trial: Argentina:
J Zanchetta, C Mautalen, Z Man, C Magaril; Australia: G Nicholson,
J Graham, J Eisman, T Roberts, G Jones, A Lowy, R Prince, P
Phillips, M Williamson, D Carey, J Wark, M Seibel, J Karrasch;
ARZOXIFENE
Belgium: J-Y Reginster, J-J Body, T Besse-Hammer, J-P Devoge-
laer, S Boonen, J-M Kaufman; Brazil: F Bandeira, M Castro, C
Fernandes, C Zerbini, L Russo, V Borba, J C Brenol, I Barbosa; Chile:
S Cheviakoff, JC Montero; Colombia: J Molina, C Cure, W Otero, P
Velez, A Terront; Denmark: M Holm-Bentzen, P Alexandersen, P
Simonsen, HC Hoeck, B Langdahl, K Brixen, L Hyldstrup; Estonia: I
Valter, K Maasalu, KVahula, I Kull; Finland: H Kröger, J Heikkinen, S
Piippo, M Välimäki; France: P Delmas, G Weryha, P Bourgeois, R
Chapurlat; Hungary: P Lakatos, L Kiss, L Korányi, B Gömör, G Poor,
I Szombati, B Spengler, A Balogh, K Takács, I Kiss; India: AC
Ammini, U Sriram, N Shah, KM Kumar, R Mehrotra, R Patni, MS
Seshadri, K Seshadri, H Kumar, S Godbole, B Stinivasan, M
Thomas; Israel: S Ish-Shalom, J Foldes, B Kaplan, D Zeltser, Y
Yaeger, M Hirsch; Mexico: P Garcia-Hernandez, H Avila-Armengol,
J Morales Torres, G Mendoza Benavente, P De la Pena-Rodriguez,
G Huerta-Yanez; New Zealand: N Gilchrist, A Fenton, J Singh, N
Graham, P Jones, M Watson, L Fitzmaurice; Norway: U Syversen, A
Høiseth, J Halse, J Stakkestad, P Norheim; Poland: E Czerwinski, B
Jachimczak, R Lorenc, E Marcinowska-Suchowierska, K Hos-
zowski, L Szczepanski, A Sawicki, Z Zieniewicz, J Brzezicki, A
Racewicz, W Tlustochowicz, S Jeka, K Lipinski; Romania: H Bolosiu,
C Codreanu, R Chirieac, I Zosin; Puerto Rico: A Reymunde, J
Morales Ramirez, E Barranco; South Africa: P Dalmeyer, M Davey,
C Davis, T de Villiers, GC Ellis, S Lipschitz, B Ascott-Evans, AJ de
Weerd, E Freysen, S Brown; Spain: S Palacios Gil Antuano, F
Hawkins Carranza, M Diaz Curiel; Sweden: O Ljunggren, G Toss;
USA: B Lubin, J Vargo, J Gabriel, M Berry-Corporon, N Koval, S
Klein, D Brandon, A Kivitz, D Fiske, H Bone, B Mizock, J Race, M
Reynolds, D Young, M Sebai, R Rosenberg, N Campbell, A Barajas,
L Popeil, R Pucillo, M Haider, G Ledesma, SPM Lillestol, M Low-
enstein, M McCartney, P Miller, D Podlecki, J Berg, R Tamayo, R
Weinstein, F Burch, J Clark, A Guevara, W Jennings, D Johns, H
McIlwain, V Piziak, S Schwartz, B Berwald, P Stein, N Lunde, JD
Rowe, RS Cohen, JJ Fanciullo, J Fiechtner, L Glaser, T Hoggarth,
JNE Lewiecki, S Mattson, L Peterson, R Kagan, S Songcharoen, S
Silverman, RR Tirado-Bernardini, J Wise, J Fidelholtz, R Kaplan, H
Knapp, B Miller, S Nattrass, J Neutel, J Poludniak, C Rosen, HJJ
Peniston, R Trapp, G Yeoman, C Recknor, K Strader, G Woodson,
M Neuffer, M DiGiovanna, M Greenwald, P Levin, L Wynstock, S
Elliott, S Diem, J McKenney, H Bays, LPR Griffin, R Lies, M Keller, M
McClung, T Antony, S Kafka, V Miller, W Garland, L Sherman, R
Malamet, Z Osborne, E Gallup, S Bart, K Kempf, R Nurnberg, T
Anderson, S Khosla, J Engelbrecht, A Kivitz, D Osterling, J Peller, R
Emkey, R Kroll, E Schwartz, T Truitt, W Powell, R Takacs, F Schaerf,
P Liu, K Ensrud, M Omizo.
References
1. Fisher B, Costantino JP, Wickerham DL, et al. Tamoxifen for preven-
tion of breast cancer: report of the National Surgical Adjuvant Breast
and Bowel Project P-1 Study. J Natl Cancer Inst. 1998;90:1371–1388.
2. Ettinger B, Black DM, Mitlak BH, et al. Reduction of vertebral fracture
risk in postmenopausal women with osteoporosis treated with
raloxifene: results from a 3-year randomized clinical trial. JAMA.
1999;282:637–645.
3. Barrett-Connor E, Mosca L, Collins P, et al. Effects of raloxifene on
cardiovascular events and breast cancer in postmenopausal women.
N Engl J Med. 2006;355:125–137.
Journal of Bone and Mineral Research 403
 4. Palkowitz AD, Glasebrook AL, Thrasher KJ, et al. Discovery and
synthesis of [6-hydroxy-3-[4-[2-(1-piperidinyl)ethoxy]phenoxy]-2-
(4-hydroxyphenyl)]benzothiophene: a novel, highly potent, selective
estrogen receptor modulator. J Med Chem. 1997;40:1407–1416.
5. Zanchetta J, Palacios S, Cox D, Stock J, Dowsett S, Kendler D.
Arzoxifene versus raloxifene: 12-month effects on bone mineral
density, bone turnover markers, and safety parameters in post-
menopausal women with osteoporosis. Osteoporos Int. 2009;20
(Suppl 1): S14.
6. Rossouw JE, Anderson GL, Prentice RL, et al. Risks and benefits of
estrogen plus progestin in healthy postmenopausal women: princi-
pal results From the Women’s Health Initiative randomized con-
trolled trial. JAMA. 2002;288:321–333.
7. Gennari L, Merlotti D, Valleggi F, Martini G, Nuti R. Selective estrogen
receptor modulators for postmenopausal osteoporosis: current state
of development. Drugs Aging. 2007;24:361–379.
8. Cummings SR, Ettinger B, Delmas PD, et al. The effects of tibolone
in older postmenopausal women. N Engl J Med. 2008;359:697–
708.
9. Genant HK, Wu CY, van KC, Nevitt MC. Vertebral fracture assessment
using a semiquantitative technique. J Bone Miner Res. 1993;8:1137–
1148.
10. Stone KL, Seeley DG, Lui LY, et al. BMD at multiple sites and risk of
fracture of multiple types: long-term results from the Study of
Osteoporotic Fractures. J Bone Miner Res. 2003;18:1947–1954.
11. Mackey DC, Lui LY, Cawthon PM, et al. High-trauma fractures and low
bone mineral density in older women and men. JAMA. 2007;298:
2381–2388.
12. Kurman RJ, Norris HJ. Endometrial hyperplasia and related cellular
changes. In: Kurman RJ, ed. Blaustein’s Pathology of the Female
Genital Tract. 4th ed. New York: Springer-Verlag; 1994:411–427.
404 Journal of Bone and Mineral Research
13. Cummings SR, Eckert S, Krueger KA, et al. The effect of raloxifene on
risk of breast cancer in postmenopausal women: results from the
MORE randomized trial. Multiple Outcomes of Raloxifene Evaluation.
JAMA. 1999;281:2189–2197. Erratum in JAMA 1999;282:2124.
14. Silverman SL, Christiansen C, Genant HK, et al. Efficacy of bazedox-
ifene in reducing new vertebral fracture risk in postmenopausal
women with osteoporosis: results from a 3-year, randomized,
placebo-, and active-controlled clinical trial. J Bone Miner Res.
2008;23:1923–1934.
15. Cummings SR, Ensrud K, Delmas PD, et al. Lasofoxifene in postme-
nopausal women with osteoporosis. N Engl J Med. 2010;362:686–696.
16. Cauley JA, Robbins J, Chen Z, et al. Effects of estrogen plus progestin
on risk of fracture and bone mineral density: the Women’s Health
Initiative randomized trial. JAMA. 2003;290:1729–1738.
17. Archer DF, Pinkerton JV, Utian WH, et al. Bazedoxifene, a selective
estrogen receptor modulator: effects on the endometrium, ovaries,
and breast from a randomized controlled trial in osteoporotic post-
menopausal women. Menopause. 2009;16:1109–1115.
18. Barrett-Connor E, Grady D, Sashegyi A, et al. Raloxifene and cardi-
ovascular events in osteoporotic postmenopausal women: four-year
results from the MORE (Multiple Outcomes of Raloxifene Evaluation)
randomized trial. JAMA. 2002;287:847–857.
19. Ivanova MM, Mazhawidza W, Dougherty SM, Minna JD, Klinge CM.
Activity and intracellular location of estrogen receptors alpha and
beta in human bronchial epithelial cells. Mol Cell Endocrinol. 2009;
305:12–21.
20. Vegeto E, Cuzzocrea S, Crisafulli C, et al. Estrogen receptor-alpha as a
drug target candidate for preventing lung inflammation. Endocrinol-
ogy. 2010;151:174–184.
21. Cirillo DJ, Wallace RB, Rodabough RJ, et al. Effect of estrogen therapy
on gallbladder disease. JAMA. 2005;293:330–339.
CUMMINGS ET AL.