ARTHRITIS & RHEUMATISM
Vol. 60, No. 11, November 2009, pp 3346–3355
Effects of Teriparatide Versus Alendronate for Treating
Glucocorticoid-Induced Osteoporosis
Thirty-Six–Month Results of a Randomized, Double-Blind, Controlled Trial
Kenneth G. Saag,1 Jose R. Zanchetta,2 Jean-Pierre Devogelaer,3 Robert A. Adler,4
Richard Eastell,5 Kyoungah See,6 John H. Krege,6 Kelly Krohn,6 and Margaret R. Warner6
Objective. To compare the bone anabolic drug
teriparatide (20 ␮g/day) with the antiresorptive drug
alendronate (10 mg/day) for treating glucocorticoid-
induced osteoporosis (OP).
Methods. This was a 36-month, randomized,
double-blind, controlled trial in 428 subjects with OP
(ages 22–89 years) who had received >5 mg/day of
prednisone equivalent for >3 months preceding screen-
ing. Measures included changes in lumbar spine and
ClinicalTrials.gov identifier: NCT00051558.
Supported by Eli Lilly and Company.
1
Kenneth G. Saag, MD: University of Alabama at Birming-
2
ham; Jose R. Zanchetta, MD: Universidad del Salvador, Buenos
Aires, Argentina; 3Jean-Pierre Devogelaer, MD: Université Catho-
lique de Louvain, Brussels, Belgium; 4Robert A. Adler, MD: McGuire
VAMC, Richmond, Virginia; 5Richard Eastell, MD: University of
Sheffield, Sheffield, UK; 6Kyoungah See, PhD, John H. Krege, MD,
Kelly Krohn, MD, Margaret R. Warner, PhD, DVM: Lilly Research
Laboratories, Eli Lilly and Company, Indianapolis, Indiana.
Dr. Saag has received consulting and speaking fees from
Novartis (more than $10,000) and from Eli Lilly, Merck, Amgen,
Procter & Gamble, and Aventis (less than $10,000 each). Dr.
Zanchetta has received consulting and speaking fees from Amgen, Eli
Lilly, Pfizer, and Servier (less than $10,000 each). Dr. Devogelaer
receives research support from Amgen, Eli Lilly, Merck, Nordic
Bioscience, Novartis, Pfizer, Procter & Gamble, Sanofi-Aventis, Ser-
vier, and Wyeth and speaking fees from Eli Lilly, Novartis, Procter &
Gamble, and Servier (less than $10,000 each). Dr. Adler has received
speaking and consulting fees from Eli Lilly, Novartis, and Merck (less
than $10,000 each), and research support from Eli Lilly, Novartis, and
Procter & Gamble. Dr. Eastell has received consulting or advisory
board fees from Amgen, Novartis, Procter & Gamble, Servier, Ono,
and GlaxoSmithKline (less than $10,000 each), lecture fees from Eli
Lilly (less than $10,000), and grant support from AstraZeneca, Procter
& Gamble, and Novartis (less than $10,000 each). Drs. See, Krege,
Krohn, and Warner own stock or stock options in Eli Lilly.
Address correspondence and reprint requests to Kenneth G.
Saag, MD, University of Alabama at Birmingham, FOT 820, 1530
Third Avenue South, Birmingham, AL 35294-3408. E-mail: ksaag@
uab.edu.
Submitted for publication December 12, 2008; accepted in
revised form July 11, 2009.
3346
DOI 10.1002/art.24879
© 2009, American College of Rheumatology
hip bone mineral density (BMD), changes in bone
biomarkers, fracture incidence, and safety.
Results. Increases in BMD from baseline were
significantly greater in the teriparatide group than in
the alendronate group, and at 36 months were 11.0%
versus 5.3% for lumbar spine, 5.2% versus 2.7% for total
hip, and 6.3% versus 3.4% for femoral neck (P < 0.001
for all). In the teriparatide group, median percent
increases from baseline in N-terminal type I procolla-
gen propeptide (PINP) and osteocalcin (OC) levels were
significant from 1 to 36 months (P < 0.01), and in-
creases in levels of C-terminal telopeptide of type I
collagen (CTX) were significant from 1 to 6 months (P
< 0.01). In the alendronate group, median percent
decreases in PINP, OC, and CTX were significant by 6
months and remained below baseline through 36
months (P < 0.001). Fewer subjects had vertebral
fractures in the teriparatide group than in the alendro-
nate group (3 [1.7%] of 173 versus 13 [7.7%] of 169; P
ⴝ 0.007), with most occurring during the first 18
months. There was no significant difference between
groups in the incidence of nonvertebral fractures (16
[7.5%] of 214 subjects taking teriparatide versus 15
[7.0%] of 214 subjects taking alendronate; P ⴝ 0.843).
More subjects in the teriparatide group (21%) versus
the alendronate group (7%) had elevated predose serum
calcium concentrations (P < 0.001).
Conclusion. Our findings indicate that subjects
with glucocorticoid-induced OP treated with teripa-
ratide for 36 months had greater increases in BMD and
fewer new vertebral fractures than subjects treated with
alendronate.
Glucocorticoids are beneficial for treating
chronic inflammatory conditions, such as rheumatoid
TERIPARATIDE VERSUS ALENDRONATE IN GLUCOCORTICOID-INDUCED OP
arthritis (RA), asthma, and systemic lupus erythemato-
sus (SLE), but bone loss is a common side effect and
fracture risk is increased (1–7). The finding that patients
taking glucocorticoids have an increased risk of fracture
at any level of bone mineral density (BMD) compared
with patients not taking glucocorticoids suggests that
glucocorticoid-induced osteoporosis (OP) is character-
ized at least in part by impaired bone quality. Currently
recommended therapies for the prevention and treatment
of glucocorticoid-induced OP include the antiresorptive
bisphosphonates alendronate and risedronate together
with supplemental calcium and vitamin D (1,8,9).
Teriparatide (recombinant human parathyroid
hormone [rhPTH(1–34)]), a bone anabolic drug, in-
creased BMD and reduced the risk of vertebral and
nonvertebral fracture in postmenopausal women with
OP (10) and increased BMD in men with hypogonadal
or idiopathic OP (11). Analyses of paired iliac crest
biopsy samples from postmenopausal women with OP
showed that teriparatide increased cancellous bone vol-
ume and connectivity, improved cancellous bone plate-
like structure, and increased cortical thickness (12).
Teriparatide induces differentiation of bone lin-
ing cells and preosteoblasts into osteoblasts, stimulates
preexisting osteoblasts to form new bone, and decreases
osteoblast and osteocyte apoptosis (13–15). Because a
primary action of glucocorticoids is to decrease bone
formation, the bone anabolic effects of teriparatide may
directly target key pathogenic mechanisms associated
with long-term glucocorticoid therapy. This hypothesis is
supported by a study which showed that areal lumbar
spine BMD, vertebral cross-sectional area, and esti-
mated vertebral compressive strength increased signifi-
cantly more in postmenopausal women with
glucocorticoid-induced OP treated with synthetic
hPTH(1–34) plus estrogen compared with those treated
with estrogen alone (16,17). Similarly, the results of the
18-month primary phase of our study showed that
lumbar spine and hip BMD increased significantly more
in subjects with glucocorticoid-induced OP treated with
teriparatide compared with those receiving alendronate
(18). This report extends the findings of this trial in
subjects with glucocorticoid-induced OP to 36 months
and includes changes in lumbar spine and hip BMD,
changes in bone turnover markers, and fracture inci-
dence.
SUBJECTS AND METHODS
Study design and participants. This was a randomized,
double-blind, double-dummy, active comparator–controlled
study conducted in 13 countries at 76 centers. There was a
3347
screening phase of ⬃1.5 months, an 18-month primary phase
(18), and an 18-month continuation phase. Investigators and
subjects remained blinded with regard to treatment through 36
months. Each subject gave informed written consent, and local
institutional review committees approved the protocol. The
study was conducted in accordance with the Declaration of
Helsinki and the Guidelines for Good Clinical Practice. The
first subject was randomized in December 2002, and the trial
was completed in January 2008.
The primary objective was the change from baseline to
18 months in lumbar spine BMD in the teriparatide versus
alendronate group (18). Prespecified secondary objectives at
36 months included lumbar spine and hip BMD, levels of bone
turnover markers, vertebral and nonvertebral fractures, and
adverse events.
Ambulatory men and women were eligible for enroll-
ment if they were at least 21 years of age and had taken
prednisone or its equivalent at a dosage of ⱖ5 mg/day for ⱖ3
months prior to screening. Subjects were required to have a
lumbar spine, femoral neck, or total hip BMD T score of ⱕ⫺2
or of ⱕ⫺1 plus a prevalent low trauma or atraumatic fracture,
as assessed by the investigator. Additional inclusion and exclu-
sion criteria have been described previously (18).
Subjects were randomly assigned to receive injectable
teriparatide (20 ␮g/day) plus oral placebo or oral alendronate
(10 mg/day) plus injectable placebo. Supplements of calcium
(1,000 mg/day) and vitamin D (800 IU/day) were provided.
Glucocorticoid use. Subjects kept a daily diary to
record their glucocorticoid use. Glucocorticoid use at baseline
was determined using the duration and dosage of only those
glucocorticoids that a subject was taking at study entry. Infor-
mation on previous usage was not obtained. The average
glucocorticoid dose at each visit was determined by averaging
the prednisone equivalent dose taken since the preceding visit.
BMD. Areal BMD (gm/cm2) was measured by dual
x-ray absorptiometry using Hologic (Hologic, Bedford, MA) or
Lunar (General Electric Medical Systems, Madison, WI) den-
sitometers. Reading of BMD scans, quality assurance, cross-
calibration adjustment, and data processing were performed by
Bio-Imaging Technologies (Newtown, PA). Lumbar vertebrae
that became fractured during the trial were excluded from the
calculation of baseline and postbaseline lumbar spine BMD.
Biochemical markers of bone turnover. Markers of
bone turnover were measured in serum obtained at approxi-
mately the same time each morning from a subset of subjects
who had fasted overnight (n ⫽ 98 subjects in the teriparatide
group and 101 subjects in the alendronate group). The bone
formation markers were bone alkaline phosphatase (bone
ALP) (Hybritech Tandem-R Ostase; Beckman Coulter, Brea,
CA) (interassay coefficient of variation [CV] 7.4–7.9%),
C-terminal type I procollagen propeptide (PICP; DiaSorin,
Stillwater, MN) (CV 5.4–8.5%), N-terminal type I procollagen
propeptide (PINP) (Intact UniQ assay; Orion Diagnostica,
Espoo, Finland) (CV 3.2–5.2%), and osteocalcin (OC) (ELSA-
OSTEO; CIS BioInternational, Bedford, MA) (CV 4.5–5.2%).
The resorption marker was C-terminal telopeptide of type I
collagen (CTX) (Serum CrossLaps; Osteometer Biotech, Her-
lev, Denmark) (CV 11.1–13.5%). Biomarkers were measured
at a central laboratory (Covance, Indianapolis, IN).
Radiographs. Spinal radiographs were obtained at
baseline, 18 months, and 36 months, and at unscheduled times
3348 SAAG ET AL

if subjects had symptoms suggestive of clinical vertebral frac-

ture. Each vertebra was graded for compression deformity,
using a visual semiquantitative method, by a Bio-Imaging
Technologies radiologist who was blinded with regard to
treatment, but not to sequence of radiographs (19). An inci-
dent vertebral fracture was defined as a vertebra not fractured
at baseline and subsequently graded as deformed (19).
For incident nonvertebral fractures, radiographs or
radiologist reports were assessed by a radiologist who was
blinded with regard to treatment. A fragility fracture was
defined as a fracture associated with trauma equivalent to a fall
from standing height or less, as assessed by the investigator.
Adverse events. Data on adverse events were collected
throughout the study. A treatment-emergent adverse event
was defined as any untoward medical event that occurred or
worsened after baseline. A serious adverse event was an event
that resulted in death, hospitalization, or significant disability/
incapacitation, or was life threatening.
Serum calcium and urate measurements. The number
of subjects with elevated serum urate levels (⬎9.0 mg/dl
[535 ␮moles/liter]) or elevated total serum calcium concentra-
tions (⬎10.5 mg/dl [2.62 mmoles/liter] and ⱖ11.0 mg/dl
[2.75 mmoles/liter]) was determined from serum collected ⬎16
hours after administration of study drugs.
Statistical analysis. Analyses were conducted on data
from all randomized and treated subjects. The study had a
power of 90% to detect a between-treatment difference of
0.015 gm/cm2 in the absolute change from baseline to last
measurement in lumbar spine BMD during the first 18 months,
Figure 1. Disposition of the subjects. ‡ ⫽ P ⬍ 0.05 versus alendro-
assuming a standard deviation of 0.04 and a 2-sided t-test with
nate.
an alpha level of 0.05. Block randomization, stratified accord-
ing to sex, investigative site, and previous use of bisphospho-
nates, was used to assign subjects to treatment in an approxi-
mate 1:1 ratio. Analysis of variance (ANOVA) was used for subgroup analyses were performed using glucocorticoid dosage
continuous variables except for markers of bone turnover, for at baseline and 36 months (3 categories: ⬍5, 5–⬍10, and
which a nonparametric method (Wilcoxon’s rank sum test) was ⱖ10 mg/day prednisone equivalent), and underlying condition
required. Categorical variables were compared between treat- requiring glucocorticoid use at baseline (4 categories: joint
ment groups using region-stratified Cochran-Mantel-Haenszel disorders [e.g., RA, osteoarthritis], other musculoskeletal dis-
or Fisher’s exact tests. ease [e.g., SLE], respiratory disorders, and other conditions)
For analyses of percent change in BMD from baseline, using ANOVA.
a treatment group comparison was made at end point, using All tests were 2-sided with the nominal significance
the last observation carried forward method (defined as 36- level set at 0.05 and were performed using SAS statistical
month or last postbaseline measurement) using ANOVA with software, version 8.2 (SAS Institute, Cary, NC).
stratification variables as covariates. If the end point reached
statistical significance, then the effect of treatments at each
time point was assessed with mixed model repeated measures. RESULTS
A predefined gate-keeping strategy controlled the overall
Type I error at an alpha level of 0.05 for determining the Subject disposition and baseline characteristics.
earliest time at which the increase in BMD differed signifi- Of the 712 subjects who were screened, 428 were
cantly between groups (20). If the 36-month comparison was randomized and received treatment (Figure 1). A total
significant at the 5% level then the 24-month comparison was of 150 (70%) of the subjects receiving teriparatide and
tested, and so on. This approach ensured that the overall Type
I error was not increased above the 0.05 level by testing at 144 (67%) of the subjects receiving alendronate entered
multiple time points. Testing of the remaining secondary the 18-month continuation phase, and 123 (57%) of the
outcomes was not adjusted for multiple comparisons. subjects receiving teriparatide and 118 (55%) of the
A within–treatment group post hoc analysis was per- subjects receiving alendronate completed the 36-month
formed to determine if the percent change in BMD was trial. The most common reasons for study discontinua-
significantly different at successive time points using the mixed
model with contrast. Covariates included were treatment, tion were subject decision and adverse event, with no
stratification variables, baseline lumbar spine BMD, time of significant difference between treatment groups. Signif-
visit, and interaction between visit and treatment. Additional icantly more teriparatide-treated subjects discontinued
TERIPARATIDE VERSUS ALENDRONATE IN GLUCOCORTICOID-INDUCED OP 3349

Table 1. Baseline characteristics of the subjects with glucocorticoid-induced osteoporosis*

Subjects taking alendronate Subjects taking teriparatide
(n ⫽ 214) (n ⫽ 214)
Age, mean ⫾ SD years 57.3 ⫾ 14.0 56.1 ⫾ 13.4
White 148 (69) 153 (71)
Women 173 (81) 172 (80)
Postmenopausal women 143 (67) 134 (63)
Treatment†
Glucocorticoid use, median (25th, 75th 2.0 (0.4, 6.5) 2.3 (0.5, 5.8)
percentiles) years‡
Prednisone equivalent, median (25th, 75th 7.5 (5.0, 10.0) 7.5 (5.0, 10.0)
percentiles)
⬍5 mg/day 20 (9) 9 (4)
ⱖ5 mg/day to ⬍10 mg/day 98 (46) 111 (53)
ⱖ10 mg/day 93 (44) 91 (43)
Prior bisphosphonate use§ 20 (9) 20 (9)
Fractures
Prevalent radiographic vertebral fracture(s)¶ 53 (25) 63 (30)
Prior nonvertebral fracture(s) 89 (42) 93 (43)
Prior nonvertebral fragility fracture(s) 43 (20) 42 (20)
BMD, least squares mean ⫾ SEM#
Lumbar spine BMD, gm/cm2 0.864 ⫾ 0.014 0.863 ⫾ 0.014
Lumbar spine BMD T score ⫺2.5 ⫾ 0.12 ⫺2.4 ⫾ 0.12
Total hip BMD, gm/cm2 0.776 ⫾ 0.013 0.763 ⫾ 0.013
Total hip BMD T score ⫺2.0 ⫾ 0.11 ⫺2.1 ⫾ 0.10
Femoral neck BMD, gm/cm2 0.721 ⫾ 0.013 0.705 ⫾ 0.013
Femoral neck BMD T score ⫺2.1 ⫾ 0.10 ⫺2.2 ⫾ 0.10
Biomechanical markers of bone turnover,
median (25th, 75th percentiles)**
Bone ALP, ␮g/liter 9 (7, 12) 9 (6, 11)
OC, ␮g/liter 14 (10, 19) 14 (9, 19)
PICP, ␮g/liter 140 (111, 177) 148 (122, 183)
PINP, ␮g/liter 39 (29, 51) 39 (29, 57)
CTX, pmoles/liter 3,331 (2,388, 5,366) 3,378 (2,050, 4,732)
Disorders requiring glucocorticoid treatment
Rheumatic disorders 161 (75) 161 (76)
Rheumatoid arthritis 111 (52) 98 (46)
Systemic lupus erythematosus 21 (10) 28 (13)
Polymyalgia rheumatica 8 (4) 10 (5)
Vasculitis 3 (1) 5 (2)
Other rheumatic disorders 18 (8) 20 (9)
Respiratory disorders 31 (14) 29 (14)
Inflammatory bowel disease 4 (2) 3 (1)
Other conditions 18 (8) 21 (10)

* Except where indicated otherwise, values are the number (%). There were no significant differences between groups.
† Glucocorticoid use was assessed in 212 subjects in the alendronate group and 213 subjects in the teriparatide group, and
prednisone equivalent was assessed in 211 subjects in each group.
‡ The duration of glucocorticoid therapy was determined based on the time that the subject was taking current glucocorticoids;
information on previous usage was not obtained.
§ Received a bisphosphonate for ⬍2 weeks within 6 months before randomization or for ⬍2 years within the previous
3 years.
¶ Prevalent radiographic vertebral fractures were assessed in 212 subjects in the alendronate group and 209 subjects in the
teriparatide group.
# Lumbar spine bone mineral density (BMD) was assessed in 207 subjects in the alendronate group and 208 subjects in the
teriparatide group, and lumbar spine BMD T score was assessed in 162 subjects in the alendronate group and 167 subjects in
the teriparatide group. Total hip BMD and femoral neck BMD were assessed in 209 subjects in each group. Total hip BMD
T score and femoral neck BMD T score were assessed in 199 subjects in each group.
** Bone alkaline phosphatase (bone ALP) was assessed in 91 subjects in the alendronate group and 81 subjects in the
teriparatide group, osteocalcin (OC) was assessed in 101 subjects in the alendronate group and 97 subjects in the teriparatide
group, C-terminal type I procollagen propeptide (PICP) was assessed in 92 subjects in the alendronate group and 82 subjects
in the teriparatide group, N-terminal type I procollagen propeptide (PINP) was assessed in 100 subjects in the alendronate
group and 98 subjects in the teriparatide group, and C-terminal telopeptide of type I collagen (CTX) was assessed in 96 subjects
in the alendronate group and 87 subjects in the teriparatide group.
3350 SAAG ET AL

due to physician decision (P ⫽ 0.028), and significantly

more alendronate-treated subjects were lost to followup
(P ⫽ 0.026).
There were no significant differences between
groups in baseline characteristics (Table 1). The age
range of the subjects was 22–89 years. In addition, there
were no significant differences in the baseline character-
istics of subjects who discontinued the study during the
first 18 months (n ⫽ 134) compared with those who
continued after 18 months (n ⫽ 294), except in the
proportion of subjects taking ⬍5 mg/day of prednisone
equivalent (4 [3.0%] of the 132 subjects who discontin-
ued and 25 [8.6%] of the 290 subjects who continued;
P ⫽ 0.037) (other data not shown).
Glucocorticoid dosage. There was no significant
difference between groups in the median glucocorticoid
dosage at baseline (Table 1) or at 36 months (6.9 mg/day
in the teriparatide group versus 7.5 mg/day in the
alendronate group; P ⫽ 0.326). The proportion of
subjects taking ⬍5, 5–⬍10, or ⱖ10 mg/day of prednisone
equivalent did not differ significantly between groups at
baseline (Table 1) or at 36 months (⬍5 mg/day in 17
[15%] of 117 subjects in the teriparatide group and 24
[22%] of 107 subjects in the alendronate group [P ⫽
0.166]; 5–⬍10 mg/day in 59 [50%] of 117 subjects in the
teriparatide group and 49 [46%] of 107 subjects in the
alendronate group [P ⫽ 0.506]; ⱖ10 mg/day in 41 [35%]
of 117 subjects in the teriparatide group and 34 [32%] of
107 subjects in the alendronate group [P ⫽ 0.671]). The
median cumulative prednisone equivalent dose was not
significantly different between groups during the first 18
months (3,444 mg in the teriparatide group versus 3,315
mg in the alendronate group; P ⫽ 0.566) or during the
final 18 months of the study (3,554 mg in the teriparatide
group versus 3,010 mg in the alendronate group; P ⫽
0.084).
Changes in BMD. The mean percent changes
from baseline in lumbar spine, total hip, and femoral
neck BMD were significantly greater in the teriparatide Figure 2. Mean percent changes from baseline in lumbar spine (A),
group than in the alendronate group at each time point total hip (B), and femoral neck (C) bone mineral density (BMD) in
and end point (Figure 2). At 36 months the mean the alendronate (ALN) group and teriparatide (TPTD) group. Num-
percent increases from baseline were 11.0% versus 5.3% bers below the graph are the numbers of subjects assessed at each time
point and for the last observation carried forward (LOCF) analysis and
for lumbar spine BMD (P ⬍ 0.001), 5.2% versus 2.7%
were identical for the total hip and femoral neck. Values are the least
for total hip BMD (P ⬍ 0.001), and 6.3% versus squares (LS) mean ⫾ SEM. ⴱ ⫽ P ⬍ 0.001; † ⫽ P ⬍ 0.01; ‡ ⫽ P ⬍ 0.05
3.4% for femoral neck BMD (P ⬍ 0.001) in the teripa- for the mean percent change from baseline in the teriparatide versus
ratide group and alendronate group, respectively. alendronate group. § ⫽ P ⬍ 0.001; 㛳 ⫽ P ⬍ 0.01, # ⫽ P ⬍ 0.05 for the
Within the teriparatide group, the mean percent changes mean percent change from baseline within each treatment group for 6
versus 3 months, 12 versus 6 months, 18 versus 12 months, 24 versus 18
in BMD were significantly greater between successive
months, and 36 versus 24 months.
time points from 3 through 36 months at the lumbar
spine, between 12 and 18 months and between 18 and 24
months at the total hip, and between 24 and 36 months mean percent changes in BMD were significantly greater
at the femoral neck. Within the alendronate group, the between 3 and 6 months, 6 and 12 months, and 18 and 24
TERIPARATIDE VERSUS ALENDRONATE IN GLUCOCORTICOID-INDUCED OP 3351

Figure 3. Median percent changes from baseline in markers of bone formation, N-terminal type I
procollagen propeptide (PINP) (A), osteocalcin (OC) (B), bone alkaline phosphatase (bone ALP) (C), and
C-terminal type I procollagen propeptide (PICP) (D), and in a marker of bone resorption, C-terminal
telopeptide of type I collagen (CTX) (E). Numbers below the graphs are the numbers of subjects assessed at
each time point. Values are the median and 25th to 75th percentiles. LOCF values are not included, as they
would overestimate the effect of teriparatide. ⴱ ⫽ P ⬍ 0.001; † ⫽ P ⬍ 0.01, ‡ ⫽ P ⬍ 0.05 for the median
percent change from baseline in the teriparatide versus alendronate group. § ⫽ P ⬍ 0.001, 㛳 ⫽ P ⬍ 0.01 for
the median percent change from baseline at each time point versus baseline within each treatment group. See
Figure 2 for other definitions.
3352 SAAG ET AL

Table 2. Incident vertebral and nonvertebral fractures in subjects and 36 months at the femoral neck. For both treatments,
with glucocorticoid-induced osteoporosis* the mean percent increases in BMD were significant
Subjects taking Subjects taking compared with baseline at each measured time point
alendronate teriparatide (P ⬍ 0.001 for teriparatide and P ⱕ 0.002 for alendro-
Fracture type (n ⫽ 214) (n ⫽ 214) P
nate, at all time points and sites).
ⱖ1 radiographic vertebral† 13 (7.7) 3 (1.7) 0.007 The differences between teriparatide and alendro-
ⱖ1 clinical vertebral‡ 4 (2.4) 0 0.037
ⱖ1 nonvertebral 15 (7.0) 16 (7.5) 0.843
nate with regard to the percent change in lumbar spine
ⱖ1 nonvertebral fragility 5 (2.3) 9 (4.2) 0.256 BMD from baseline to end point were not significantly
different across 4 underlying disease subgroups or across 3
* Values are the number (%).
† Subjects with baseline and postbaseline spinal radiographs (n ⫽ 169 glucocorticoid dose ranges at baseline or 36 months. These
subjects in the alendronate group and 173 subjects in the teriparatide results suggest that the treatment difference between
group). teriparatide and alendronate for percent change in lumbar
‡ A clinical vertebral fracture (assessed in 169 subjects in the alendro-
nate group and 173 subjects in the teriparatide group) was a new spine BMD was not affected by the underlying disease
radiographically confirmed fracture that was associated with symptoms requiring glucocorticoid therapy or by the dose of glucocor-
such as back pain. ticoid at baseline or at 36 months.
Changes in levels of biochemical markers of bone
months at the lumbar spine, between 12 and 18 months turnover. In the teriparatide group, there were signifi-
at the total hip, and between 12 and 18 months and 24 cant median percent increases from baseline in all

Table 3. Summary of treatment-emergent adverse events and predose serum calcium and urate concentrations*
Subjects taking alendronate Subjects taking teriparatide
(n ⫽ 214) (n ⫽ 214) P
Overall adverse events
ⱖ1 adverse event 184 (86) 194 (91) 0.116
ⱖ1 serious adverse event 64 (30) 70 (33) 0.518
ⱖ1 possibly treatment-related adverse event 42 (20) 58 (27) 0.074
Adverse events that occurred with a ⱖ2%
difference between groups
Higher proportion in the teriparatide group
Nausea 18 (8) 36 (17) 0.007
Headache 14 (7) 19 (9) NS
Gastritis 8 (4) 17 (8) NS
Dyspnea 6 (3) 16 (7) 0.028
Insomnia 3 (1) 12 (6) 0.017
Fatigue 4 (2) 9 (4) NS
Viral infection 0 5 (2) 0.023
Higher proportion in the alendronate group
Urinary tract infection 29 (14) 22 (10) NS
Influenza 24 (11) 18 (8) NS
Anemia 17 (8) 11 (5) NS
Dyspepsia 15 (7) 9 (4) NS
Nasopharyngitis 13 (6) 7 (3) NS
Rash 10 (5) 4 (2) NS
Joint injury 6 (3) 1 (0.5) NS
Weight loss 9 (4) 0 0.002
Predose total calcium and urate†
ⱖ1 serum calcium ⬎10.5 mg/dl‡ 14 (7) 44 (21) ⬍0.001
ⱖ2 serum calcium ⬎10.5 mg/dl§ 6 (3) 16 (8) 0.046
ⱖ1 serum calcium ⱖ11.0 mg/dl‡ 3 (1) 9 (4) 0.140
ⱖ2 serum calcium ⱖ11.0 mg/dl§ 0 1 (0.5) 1.000
ⱖ1 serum urate ⬎9.0 mg/dl¶ 11 (5) 20 (9) 0.135

* Predose was defined as ⬎16 hours after administration of study drugs. Values are the number (%). NS ⫽ not significant (P ⬎ 0.05).
† To convert serum calcium concentrations to mmoles/liter multiply by 0.25, and to convert serum urate concentrations to
␮moles/liter multiply by 59.5.
‡ Baseline and ⱖ1 postbaseline serum calcium measurements were available for 209 subjects in the alendronate group and for
211 subjects in the teriparatide group.
§ Baseline and ⱖ2 postbaseline serum calcium measurements were available for 196 subjects in each group.
¶ Baseline and ⱖ1 postbaseline serum urate measurements were available for 208 subjects in the alendronate group and 212
subjects in the teriparatide group.
TERIPARATIDE VERSUS ALENDRONATE IN GLUCOCORTICOID-INDUCED OP
markers after 1 month, and the changes were maximal at
1 month for OC and PICP, and at 6 months for bone
ALP, PINP, and CTX (Figure 3). The median percent
changes in bone ALP, OC, and PINP levels remained
significantly increased compared with baseline at 6, 18,
and 36 months, while the median percent changes in
PICP and CTX were not significantly different from
baseline at 6 (PICP only), 18, or 36 months (Figure 3). In
the alendronate group, there were significant median
percent decreases from baseline after 1 month for PICP,
PINP, and CTX levels, and after 6 months for bone ALP
and OC levels (Figure 3). The decreases from baseline
were maximal at 6 months for PINP, at 18 months for
bone ALP, PICP, and CTX, and at 36 months for OC;
however, for PINP, OC, PICP, and CTX the median
percent decreases were significantly different from base-
line from 6 through 36 months (Figure 3). There were
significant differences between the teriparatide and
alendronate groups in the median percent change from
baseline for all biomarkers at 1, 6, 18, and 36 months
(P ⬍ 0.05).
Incident vertebral and nonvertebral fractures. A
total of 3 (1.7%) of 173 subjects receiving teriparatide
compared with 13 (7.7%) of 169 subjects receiving
alendronate had 1 new radiographic vertebral fracture
over 36 months (P ⫽ 0.007) (Table 2). Most of the
vertebral fractures occurred during the first 18 months
(in 1 subject receiving teriparatide and 10 subjects
receiving alendronate). There were no significant differ-
ences between groups in the number of subjects with
new nonvertebral fractures or with new nonvertebral
fragility fractures (Table 2).
Comparison of adverse events. There were no
significant differences between groups in the number of
subjects reporting ⱖ1 adverse event, in the incidence of
serious adverse events, or in the incidence of adverse
events considered to be possibly related to study drugs as
assessed by investigators (Table 3). There was no signif-
icant difference between groups in the number of sub-
jects who died during the study (n ⫽ 9 in the teriparatide
group and n ⫽ 15 in the alendronate group; P ⫽ 0.219).
Table 3 lists the adverse events that occurred with ⱖ2%
difference between groups.
Comparison of serum calcium and urate concen-
trations. Significantly more subjects in the teriparatide
group than in the alendronate group had ⱖ1 or ⱖ2
predose serum calcium concentrations of ⬎10.5 mg/dl,
with no significant difference between groups in the
number of subjects with ⱖ1 or ⱖ2 predose serum
calcium concentrations of ⱖ11.0 mg/dl (Table 3). There
was no significant difference between groups in the
3353
number of subjects with ⱖ1 serum urate concentration
⬎9.0 mg/dl (Table 3). Adverse events related to serum
calcium or urate included hypercalcemia in 1 subject
receiving teriparatide, nephrolithiasis in 3 subjects re-
ceiving teriparatide and 2 subjects receiving alendro-
nate, and gout in 1 subject receiving teriparatide.
DISCUSSION
Among subjects with glucocorticoid-induced OP,
increases in spine and hip BMD were significantly
greater in those receiving teriparatide than in those
receiving alendronate through 36 months. New radio-
graphic vertebral fractures occurred in significantly
fewer subjects in the teriparatide group than in the
alendronate group, with no significant difference be-
tween groups in the number of subjects with new
nonvertebral fractures. In the teriparatide group, levels
of bone formation markers were significantly above
baseline at 1 month through 36 months, and in the
alendronate group, markers of bone formation and
resorption were significantly decreased from baseline by
1–6 months through 36 months.
In previous placebo-controlled trials, antiresorp-
tive therapy significantly increased spine and hip BMD
in subjects beginning glucocorticoid therapy (21) or in
subjects with established glucocorticoid-induced OP
(22–25). For example, in subjects taking glucocorticoids,
there was a 3.9% increase in lumbar spine BMD after 24
months in the alendronate 10 mg/day group versus a
0.8% decrease in the placebo group (23). In our study,
subjects with glucocorticoid-induced OP receiving 10
mg/day of alendronate had an increase in lumbar spine
BMD of 5.2% at 24 months and 5.3% at 36 months. In
comparison, subjects receiving teriparatide had an in-
crease in lumbar spine BMD of 9.8% at 24 months and
11.0% at 36 months. Thus, in subjects taking glucocor-
ticoids, the bone anabolic drug, teriparatide, was more
efficacious than alendronate for increasing BMD and
resulted in further significant increases in BMD beyond
the initial 18-month primary study end point.
The changes from baseline in bone turnover
markers in subjects taking teriparatide or alendronate
were directionally similar to those reported in subjects
not taking glucocorticoids (26,27), reflecting the bone
anabolic effects of teriparatide and the antiresorptive
effects of alendronate in subjects with glucocorticoid-
induced OP.
It has been suggested that optimal gains in BMD
with daily teriparatide treatment may be limited over
time due to the return toward baseline in bone forma-
3354
tion markers after 9–12 months of therapy (28). Similar
to previous findings with teriparatide (27,28), in our
study the bone turnover markers in subjects with
glucocorticoid-induced OP receiving teriparatide re-
turned toward baseline by 18 months. However, levels of
the bone formation markers, OC, PINP, and bone ALP,
each remained significantly increased above baseline
through 36 months, whereas the bone resorption
marker, CTX, was not significantly different from base-
line at 18 or 36 months. These changes in bone forma-
tion markers were associated with significant increases
in lumbar spine BMD between successive time points
through 36 months, indicating that gains in BMD with
daily administration of teriparatide continued even
though there was a decline in the levels of bone turnover
markers. This result may be explained by the balance in
bone turnover favoring formation over resorption. How-
ever, the precise relationship between the serum con-
centration of bone biomarkers and relative gain or loss
in BMD is unknown. Sustained anabolic effects on
cortical and cancellous bone have also been shown
during treatment with once-daily PTH(1–34) for 18–36
months in subjects with OP (29). With the antiresorptive
drug alendronate, the sustained decrease in CTX, OC,
PICP, and PINP levels through 36 months was associ-
ated with a significant increase in lumbar spine BMD
between successive time points from 3 to 12 months and
between 18 and 24 months, but with no significant
increase in lumbar spine BMD between 24 and 36
months.
Over 36 months, 16 subjects (3 receiving teripa-
ratide and 13 receiving alendronate) experienced 1 new
radiographic vertebral fracture. Most of these new ver-
tebral fractures occurred during the first 18 months of
the study (1 in the teriparatide group and 10 in the
alendronate group) (18). There was no significant dif-
ference between groups in the number of subjects with
ⱖ1 new nonvertebral fracture over 18 months (12 in the
teriparatide group versus 8 in the alendronate group) or
36 months (16 in the teriparatide group versus 15 in the
alendronate group) (18). This trial was not designed to
determine the effects of teriparatide or alendronate on
the risks of vertebral or nonvertebral fracture. In
placebo-controlled trials, both teriparatide and alendro-
nate have been shown to decrease vertebral and nonver-
tebral fracture incidence in postmenopausal women not
taking glucocorticoids (10,30,31).
Strengths of our study include the controlled,
randomized, double-blind design, the finding that
changes in lumbar spine BMD observed with alendro-
nate 10 mg/day were similar to those reported in previ-
SAAG ET AL
ous studies of alendronate in glucocorticoid-induced OP
(22,23), and the first clinical trial experience with teripa-
ratide and alendronate for up to 36 months in
glucocorticoid-induced OP. Limitations include the sub-
ject discontinuation rate of 44% at 36 months which, in
part, reflects the severity of underlying disease in sub-
jects with glucocorticoid-induced OP and the propensity
for these subjects to have multiple comorbid conditions
(32). There was no significant difference between groups
in the subject discontinuation rate, and in placebo-
controlled studies of bisphosphonate treatment in
glucocorticoid-induced OP, similar discontinuation rates
have been reported (21,22,24). Although 86 subjects
with missing or unreadable spinal radiographs were
excluded from the vertebral fracture analysis, there was
no significant difference between groups in the baseline
characteristics of the excluded subjects. Because predose
serum calcium was measured, transient hypercalcemia
after the administration of study drugs would not have
been detected (10).
In summary, subjects at high risk of fracture
associated with sustained glucocorticoid use who re-
ceived teriparatide had significantly greater increases in
spine and hip BMD compared with subjects receiving
alendronate during 36 months of therapy. Patients with
glucocorticoid-induced OP treated with teriparatide also
experienced significantly fewer new vertebral fractures,
most occurring during the first 18 months, with no
significant difference between groups in the incidence of
nonvertebral fractures. Additional key findings included
significant increases in lumbar spine BMD between 24
and 36 months in the teriparatide group, and sustained
increases in biomarkers of bone formation through 36
months in the teriparatide group with sustained de-
creases in biomarkers of bone formation and resorption
in the alendronate group. There remains an unmet need
for therapies that can substantially improve the skeletal
status of subjects with glucocorticoid-induced OP (1,33).
Our data indicate that teriparatide is efficacious and
generally well tolerated for treating subjects with
glucocorticoid-induced OP and should be considered as
a therapeutic option for subjects at high risk of fracture.
ACKNOWLEDGMENT
The authors acknowledge the clinical trial investigators
(18).
AUTHOR CONTRIBUTIONS
All authors were involved in drafting the article or revising it
critically for important intellectual content, and all authors approved
TERIPARATIDE VERSUS ALENDRONATE IN GLUCOCORTICOID-INDUCED OP
the final version to be published. Dr. Saag had full access to all of the
data in the study and takes responsibility for the integrity of the data
and the accuracy of the data analysis.
Study conception and design. Saag.
Acquisition of data. Saag, Zanchetta, Devogelaer, Adler.
Analysis and interpretation of data. Saag, Zanchetta, Devogelaer,
Adler, Eastell, See, Krege, Krohn, Warner.
REFERENCES
1. American College of Rheumatology Ad Hoc Committee on
Glucocorticoid-Induced Osteoporosis. Recommendations for the
prevention and treatment of glucocorticoid-induced osteoporosis:
2001 update. Arthritis Rheum 2001;44:1496–503.
2. Saag KG. Glucocorticoid-induced osteoporosis. Endocrinol Metab
Clin North Am 2003;32:135–57.
3. Canalis E, Mazziotti G, Giustina A, Bilezikian JP. Glucocorticoid-
induced osteoporosis: pathophysiology and therapy. Osteoporos
Int 2007;18:1319–28.
4. Rubin MR, Bilezikian JP. Clinical review 151—the role of para-
thyroid hormone in the pathogenesis of glucocorticoid-induced
osteoporosis: a re-examination of the evidence. J Clin Endocrinol
Metab 2002;87:4033–41.
5. Van Staa TP, Leufkens HG, Abenhaim L, Zhang B, Cooper C.
Use of oral corticosteroids and risk of fractures. J Bone Miner Res
2000;15:993–1000.
6. Van Staa TP, Laan RF, Barton IP, Cohen S, Reid DM, Cooper C.
Bone density threshold and other predictors of vertebral fracture
in patients receiving oral glucocorticoid therapy. Arthritis Rheum
2003;48:3224–9.
7. Kanis JA, Johansson H, Oden A, Johnell O, de Laet C, Melton LJ
III, et al. A meta-analysis of prior corticosteroid use and fracture
risk. J Bone Miner Res 2004;19:893–9.
8. Royal College of Physicians. Glucocorticoid-induced osteoporosis:
guidelines for prevention and treatment. URL: www.rcplondon.
ac.uk/pubs/books/glucocorticoid/index.asp.
9. Devogelaer JP, Goemaere S, Boonen S, Body JJ, Kaufman JM,
Reginster JY, et al. Evidence-based guidelines for the prevention
and treatment of glucocorticoid-induced osteoporosis: a consensus
document of the Belgian Bone Club. Osteoporos Int 2006;17:
8–19.
10. Neer RM, Arnaud CD, Zanchetta JR, Prince R, Gaich GA,
Reginster JY, et al. Effect of parathyroid hormone (1-34) on
fractures and bone mineral density in postmenopausal women with
osteoporosis. N Engl J Med 2001;344:1434–41.
11. Orwoll ES, Scheele WH, Paul S, Adami S, Syversen U, Diez-Perez
A, et al. The effect of teriparatide [human parathyroid hormone
(1-34)] therapy on bone density in men with osteoporosis. J Bone
Miner Res 2003;18:9–17.
12. Jiang Y, Zhao JJ, Mitlak BH, Wang O, Genant HK, Eriksen EF.
Recombinant human parathyroid hormone (1-34) [teriparatide]
improves both cortical and cancellous bone structure. J Bone
Miner Res 2003;18:1932–41.
13. Canalis E, Giustina A, Bilezikian JP. Mechanisms of anabolic
therapies for osteoporosis. N Engl J Med 2007;357:905–16.
14. Dobnig H, Turner RT. Evidence that intermittent treatment with
parathyroid hormone increases bone formation in adult rats by
activation of bone lining cells. Endocrinology 1995;136:3632–8.
15. Jilka RL, Weinstein RS, Bellido T, Roberson P, Parfitt AM,
Manolagas SC. Increased bone formation by prevention of osteo-
blast apoptosis with parathyroid hormone. J Clin Invest 1999;104:
439–46.
16. Lane NE, Sanchez S, Modin GW, Genant HK, Pierini E, Arnaud
CD. Parathyroid hormone treatment can reverse corticosteroid-
3355
induced osteoporosis: results of a randomized controlled clinical
trial. J Clin Invest 1998;102:1627–33.
17. Rehman Q, Lang TF, Arnaud CD, Modin GW, Lane NE. Daily
treatment with parathyroid hormone is associated with an increase
in vertebral cross-sectional area in postmenopausal women with
glucocorticoid-induced osteoporosis. Osteoporos Int 2003;14:
77–81.
18. Saag KG, Shane E, Boonen S, Marin F, Donley DW, Taylor KA,
et al. Teriparatide or alendronate in glucocorticoid-induced osteo-
porosis. N Engl J Med 2007;357:2028–39.
19. Genant HK, Wu CY, van Kuijk C, Nevitt MC. Vertebral fracture
assessment using a semiquantitative technique. J Bone Miner Res
1993;8:1137–48.
20. Westfall PH, Krishen A. Optimally weighted, fixed sequence, and
gatekeeping multiple testing procedures. J Stat Plan Inference
2001;99:25–40.
21. Cohen S, Levy RM, Keller M, Boling E, Emkey RD, Greenwald
M, et al. Risedronate therapy prevents corticosteroid-induced
bone loss: a twelve month, multicenter, randomized, double-blind,
placebo-controlled parallel-group study. Arthritis Rheum 1999;42:
2309–18.
22. Saag KG, Emkey R, Schnitzer TJ, Brown JP, Hawkins F, Go-
emaere S, et al. Alendronate for the prevention and treatment of
glucocorticoid-induced osteoporosis. N Engl J Med 1998;339:
292–9.
23. Adachi JD, Saag KG, Delmas PD, Liberman UA, Emkey RD,
Seeman E, et al. Two-year effects of alendronate on bone mineral
density and vertebral fracture in patients receiving glucocorticoids:
a randomized, double-blind, placebo-controlled extension trial.
Arthritis Rheum 2001;44:202–11.
24. Reid DM, Hughes RA, Laan RF, Sacco-Gibson NA, Wenderoth
DH, Adami S, et al. Efficacy and safety of daily risedronate in the
treatment of corticosteroid-induced osteoporosis in men and
women: a randomized trial. J Bone Miner Res 2000;15:1006–13.
25. Ringe JD, Dorst A, Faber H, Ibach K, Sorenson F. Intermittent
intravenous ibandronate injections reduce vertebral fracture risk
in corticosteroid-induced osteoporosis: results from a long-term
comparative study. Osteoporos Int 2003;14:801–7.
26. Szulc P, Delmas PD. Biochemical markers of bone turnover:
potential use in the investigation and management of postmeno-
pausal osteoporosis. Osteoporos Int 2008;19:1683–704.
27. McClung MR, San Martin J, Miller PD, Civitelli R, Bandeira F,
Omizo M, et al. Opposite bone remodeling effects of teriparatide
and alendronate in increasing bone mass. Arch Intern Med
2005;165:1762–8.
28. Rubin MR, Bilezikian JP. The anabolic effects of parathyroid
hormone therapy. Clin Geriatr Med 2003;19:415–32.
29. Dempster DW, Cosman F, Kurland ES, Zhou H, Nieves J,
Woelfert L, et al. Effects of daily treatment with parathyroid
hormone on bone microarchitecture and turnover in patients with
osteoporosis: a paired biopsy study. J Bone Miner Res 2001;16:
1846–53.
30. Black DM, Thompson DE, Bauer DC, Ensrud K, Musliner T,
Hochberg MC, et al. Fracture risk reduction with alendronate in
women with osteoporosis: the Fracture Intervention Trial. J Clin
Endocrinol Metab 2000;85:4118–24.
31. Liberman UA, Weiss SR, Broll J, Minne HW, Quan H, Bell NH,
et al. Effect of oral alendronate on bone mineral density and the
incidence of fractures in postmenopausal osteoporosis. N Engl
J Med 1995;333:1437–43.
32. McDonough AK, Curtis JR, Saag KG. The epidemiology of
glucocorticoid-associated adverse events. Curr Opin Rheumatol
2008;20:131–7.
33. Manolagas SC, Weinstein RS. New developments in the patho-
genesis and treatment of steroid-induced osteoporosis. J Bone
Miner Res 1999;14:1061–6.