Effects of denosumab on bone mineral density and bone turnover in postmenopausal women transitioning from alendronate therapy

Abstract

Patients treated with bisphosphonates for osteoporosis may discontinue or require a switch to other therapies. Denosumab binds to RANKL and is a potent inhibitor of bone resorption that has been shown to increase bone mineral density (BMD) and decrease fracture risk in postmenopausal women with osteoporosis. This was a multicenter, international, randomized, double-blind, double-dummy study in 504 postmenopausal women ≥ 55 years of age with a BMD T-score of −2.0 or less and −4.0 or more who had been receiving alendronate therapy for at least 6 months. Subjects received open-label branded alendronate 70 mg once weekly for 1 month and then were randomly assigned to either continued weekly alendronate therapy or subcutaneous denosumab 60 mg every 6 months and were followed for 12 months. Changes in BMD and biochemical markers of bone turnover were evaluated. In subjects transitioning to denosumab, total hip BMD increased by 1.90% at month 12 compared with a 1.05% increase in subjects continuing on alendronate (p < .0001). Significantly greater BMD gains with denosumab compared with alendronate also were achieved at 12 months at the lumbar spine, femoral neck, and 1/3 radius (all p < .0125). Median serum CTX levels remained near baseline in the alendronate group and were significantly decreased versus alendronate (p < .0001) at all time points with denosumab. Adverse events and serious adverse events were balanced between groups. No clinical hypocalcemic adverse events were reported. Transition to denosumab produced greater increases in BMD at all measured skeletal sites and a greater reduction in bone turnover than did continued alendronate with a similar safety profile in both groups. Copyright © 2010 American Society for Bone and Mineral Research

Introduction

Osteoporosis is a chronic and progressive condition characterized by decreased bone mass and microarchitectural deterioration, leading to increased bone fragility and a consequential increased risk of fracture.1–3 An estimated 9 million new osteoporosis-related fractures occurred worldwide in the year 2000,4 and as many as 75 million people in the United States, Europe, and Japan are affected by osteoporosis.1 Because numerous therapies exist to treat osteoporosis, and since patients switch between therapies, the outcomes after changing therapies are important to investigate. In particular, recognizing the long skeletal retention of bisphosphonates, the safety and efficacy of different antiresorptive agents given after alendronate therapy is of interest to clinicians.

The RANKL inhibitor denosumab is a fully human monoclonal antibody under investigation as a therapy for osteoporosis. RANKL is essential for the formation, activity, and survival of osteoclasts.5–9 By binding RANKL, denosumab potently reduces bone resorption with accompanying increases in bone mineral density (BMD). In previous studies, denosumab treatment for up to 4 years increased BMD significantly at the lumbar spine, total hip, one-third radius, and total body compared with placebo10–13 and reduced the risk of vertebral, nonvertebral, and hip fractures significantly over 3 years in postmenopausal women with osteoporosis.14

Bisphosphonates, including alendronate, are the most widely prescribed class of therapy for osteoporosis.15, 16 Bisphosphonates also reduce bone resorption, but through a different mechanism of action than denosumab.17 In a head-to-head study in treatment-naive subjects, denosumab produced significantly greater gains in BMD at the total hip, lumbar spine, and other sites than branded alendronate.18 The present study was conducted in postmenopausal women previously treated with alendronate to evaluate the effects of transitioning to denosumab on safety, BMD, and bone remodeling in comparison with continued branded alendronate therapy.

Materials and Methods

Subjects

Ambulatory postmenopausal women at least 55 years of age with a lumbar spine or total hip BMD measurements corresponding to a T-score of −2.0 or less and −4.0 or greater and who had been receiving alendronate treatment equivalent to 70 mg/week for at least 6 months were eligible. Women were excluded if they had current hyper- or hypothyroidism, current hyper- or hypoparathyroidism, elevated transaminases, significantly impaired renal function (creatinine clearance ≤ 35 mL/min as estimated by the Cockcroft and Gault method),19 hyper- or hypocalcemia, serum 25-hydroxyvitamin D levels < 20 ng/mL (<50 nmol/L) or any other condition that could result in impaired calcium metabolism, or any metabolic bone disease that could interfere with interpretation of the findings. Women who were intolerant of alendronate therapy or for whom it was contraindicated or who had taken any bisphosphonate other than alendronate within 1 year of screening also were excluded. Women were excluded if they had ever received intravenous bisphosphonates, fluoride (except for dental treatment), or strontium ranelate; had received parathyroid hormne (PTH) or PTH derivatives within 1 year; had received any Selective Estrogen Receptor Modulator (SERM), anabolic steroids, systemic hormone replacement, calcitonin, calcitriol, or other vitamin D derivatives within 3 months; or had height, weight, or girth measurements that precluded accurate dual-energy x-ray absorptiometry (DXA) assessments. All subjects provided written informed consent before undergoing any study-related procedures. The study was conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonisation of Good Clinical Practice Guidelines, and the study protocol was approved by an institutional review board for each study site.

Study design

The Study of Transitioning from Alendronate to Denosumab (STAND) was a 1-year international, multicenter, randomized, double-blind, double-dummy, parallel-group, phase 3 trial that was conducted between October 2, 2006 and March 27, 2008. After a 1-month run-in period during which all subjects received open-label, branded alendronate (Fosamax) 70 mg once weekly, subjects were randomly assigned to receive subcutaneous denosumab injections 60 mg once every 6 months or to continue receiving branded alendronate 70 mg once weekly. Subjects assigned to denosumab also received placebo tablets once weekly, and subjects assigned to alendronate also received placebo subcutaneous injections every 6 months to maintain the blind. Randomization was stratified by length of prior alendronate therapy: 6 to <12 months, 12 to 24 months, or >24 months. Randomization and treatment assignment were based on a randomization schedule that used randomly permuted blocks and that was prepared by the sponsor before the start of the study. All subjects were supplied with 1000 mg elemental calcium and at least 400 IU vitamin D daily.

Study procedures

After screening, study visits occurred at baseline (day 1) and months 1, 3, 6, 9, and 12, with additional study visits occurring 5, 10, or 15 days after the day 1 visit and 5, 10, or 15 days after the month 6 visit. BMD of the lumbar spine, proximal femur (“total hip”), femoral neck, and one-third radius was measured by DXA at screening, month 6, and month 12. Spine and hip scans were performed in duplicate. The mean of the duplicate measurements was used for analyses. All DXA scan data were submitted to a central imaging vendor (Synarc) for blinded analysis. Fasting morning blood samples were collected at all study visits for measurement of biochemical markers of bone turnover in serum. Serum concentrations of type 1 C-telopeptide (CTX-I) and intact N-terminal propeptide of type I procollagen (PINP) were assayed by the University of Liege CHU using the Serum CrossLaps ELISA (Nordic Bioscience Diagnostics A/S, Herlev, Denmark) and UniQ PINP RIA (Orion Diagnostika GmbH, Wedel, Germany), respectively. Hematology assessments were performed at baseline, month 6, and month 12, and serum chemistry assessments were performed at baseline, post-day-1, month 1, month 6, post-month-6, and month 12 by a central laboratory (Quintiles Laboratories, Ltd.) using standard procedures. Evaluation of anti–denosumab antibody formation in subjects receiving denosumab was performed at baseline and months 6 and 12 by PPD development and the sponsor using previously described methods.12 Adherence with oral study drug was evaluated at the end of the open-label run-in period and at months 1, 3, 6, 9, and 12 by counting tablets. Adverse events were recorded at each study visit. Fractures were reported as adverse events, and review of radiographs was not adjudicated.

Endpoints

The primary efficacy endpoint was the percent change in total hip BMD from baseline to month 12. The secondary efficacy endpoints were the percent change from baseline in serum CTX-I at month 3 and the percent change from baseline in lumbar spine BMD at month 12. Other endpoints included the percent change from baseline in BMD at the femoral neck and one-third radius at month 12. Safety endpoints included adverse events, changes in safety laboratory analytes, serum calcium levels, and vital signs.

Statistical analysis

The primary efficacy analysis included all randomized subjects who had a baseline measurement and at least one postbaseline measurement. Data from subjects in this subset were analyzed according to randomized treatment groups. Safety analysis included all subjects who received at least 1 dose of study drug and were analyzed according to actual treatment received. A run-in safety subset included all subjects who entered the 1-month run-in period and were given alendronate, regardless of whether they received study drug during the 12-month treatment period. A repeated-measures model was used as the primary analysis method for the percent change in BMD at months 6 and 12. The model included treatment, time of BMD assessment, treatment-by-time-of-BMD-assessment interaction, baseline BMD value, time-on-prior-alendronate-therapy stratum, DXA instrument type, and baseline-BMD-value-and-instrument-type interaction. Least significant change (LSC) in the BMD measurements for the spine and hip was calculated using the duplicate scans from baseline based on the precision of the DXA measurements using the root mean square CV (rms-CV%) multiplied by 2.77, as described previously.18, 20 Analyses of changes from baseline in bone turnover markers were carried out in the observed data subset; missing values were not imputed. Measurements of available samples that were below the lower limit of quantification were considered equal to the value of the lower limit of quantification for all analyses. Median values for bone turnover markers are reported because baseline values and percentage changes were skewed. A Wilcoxon rank-sum test adjusting for time-on-prior-alendronate stratum was used to assess the significance of the treatment difference at each time point for bone turnover markers.

The primary hypothesis was that denosumab was not inferior to alendronate with respect to the mean percentage change in total hip BMD at month 12. The noninferiority margin for the difference in percent change in total hip BMD between the two treatment groups was −0.35%. The −0.35% margin was based on a comparison of data from two studies presenting 12-month changes in total hip BMD in subjects who had previously received at least 6 months of alendronate therapy and then continued or discontinued therapy.11, 21 The difference in the 12-month percent change in total hip BMD between subjects who continued and subjects who discontinued alendronate therapy in these studies was 1.66%, with a confidence interval (CI) for the difference between the groups of 0.71%−2.61%. The noninferiority margin represented half the lower bound of this confidence interval. With 235 subjects in each treatment group, a two-group, 2.5% one-sided t test would provide 90% power to confirm noninferiority for the difference in means between those transitioning to denosumab and those continuing on alendronate based on an assumed difference of 0.25% between means for the primary endpoint and common standard deviation of 2.0%. Assuming a dropout rate of 5% over the 12-month study, the planned enrollment was 250 subjects per group. A step-down multiple-testing procedure was used to control the family-wise error rate. Specifically, (1) if noninferiority of denosumab for total hip BMD at month 12 was demonstrated by a lower bound of the 95% CI to be greater than −0.35%, then (2) superiority for the percent reduction in serum CTX-I at month 3 would be tested, and if the p value was < .05, superiority would be stated, and (3) superiority of total hip BMD at month 12 would be tested; if the p value was < .05, then superiority would be demonstrated, and (4) noninferiority at the lumbar spine would be tested and demonstrated if the lower bound of the 95% CI was greater than −0.22. The noninferiority margin for the lumbar spine was determined as described above for the total hip. Analyses were confirmed by independent statistical analyses performed at Vanderbilt University Department of Biostatistics.

Results

From a total of 1020 women screened, 504 women were enrolled in the study, and 481 (95.4%) completed 12 months of follow-up. Rates of discontinuation were similar between the denosumab and continued alendronate groups, as were the reasons for early withdrawal (Fig. 1). Compliance with study medication was very good; 94% of subjects in each group received both injections (denosumab or matching placebo) and at least 80% of the tablets (alendronate or matching placebo) through month 12.

Figure 1. Participant flow diagram. Subjects continued to undergo screening after entering the 1-month open-label run-in period. Randomization was considered the time of enrollment.

Baseline demographics and clinical characteristics were similar between the two treatment groups (Table 1). Subjects had a mean age of 67.6 years, with an average of 19.3 years since menopause. Subjects had been treated with bisphosphonate therapy for a median of 36 months (range 6 to 192 months) in the period immediately before screening. Forty-three percent of subjects had received generic alendronate before screening. Average BMD T-scores at the total hip and lumbar spine were −1.80 and −2.63, respectively, and 50% had previously had an osteoporosis-related fracture. Median baseline serum levels for the bone turnover markers CTX-I and PINP were 0.204 ng/mL and 22.11 µg/L, respectively, reflective of bone marker reductions from prior bisphosphonate therapy.

Table 1. Baseline Demographics and Subject Disease Characteristics

	Alendronate 70 mg/week (N = 251)	Denosumab 60 mg q6 m (N = 253)	All (N = 504)
Abbreviations: SD, standard deviation; BMI, body mass index; BMD, bone mineral density; CTX-I, type 1 C-telopeptide; PINP, intact N-terminal propeptide of type 1 procollagen; Q1, Q3, interquartile range a Defined as osteoporotic vertebral or nonvertebral fractures; no assessment of vertebral fractures was performed for this study. p > .05 for all comparisons.
Age, mean ± SD, years	68.2 ± 7.7	66.9 ± 7.8	67.6 ± 7.8
BMI, mean ± SD, kg/m^2	24.7 ± 4.0	24.2 ± 3.8	24.4 ± 3.9
Years since menopause, mean ± SD, years	19.9 ± 9.9	18.8 ± 9.2	19.3 ± 9.6
Prior time on bisphosphonates, median (range), months	34.5 (6–192)	36.0 (6–133)	36.0 (6–192)
History of osteoporotic fracturea	117 (47%)	134 (53%)	251 (50%)
Total hip BMD T-score, mean ± SD	−1.81 ± 0.74	−1.79 ± 0.82	−1.80 ± 0.78
Lumbar spine BMD T-score, mean ± SD	−2.62 ± 0.79	−2.64 ± 0.75	−2.63 ± 0.77
Serum CTX-I, median (Q1, Q3), ng/mL	0.207 (0.132, 0.320)	0.187 (0.127, 0.291)	0.204 (0.130, 0.309)
Serum PINP, median (Q1, Q3), µg/L	22.52 (15.81, 31.16)	21.24 (16.04, 28.79)	22.11 (15.97, 29.82)

Bone mineral density

BMD at the total hip increased by 1.90% (95% CI 1.61%−2.18%) at month 12 in subjects transitioned to denosumab compared with a 1.05% (95% CI 0.76%−1.34%) increase from baseline in subjects continuing on alendronate therapy (Fig. 2). The difference between treatment groups was 0.85% (95% CI 0.44%−1.25%) greater with denosumab; the lower limit of the confidence interval excluded the prespecified noninferiority margin (−0.35%), thus showing the noninferiority of denosumab compared with alendronate. Superiority testing demonstrated the BMD increase with denosumab at the total hip was statistically superior to the change with alendronate (p < .0001).

Figure 2. Percentage change from baseline in bone mineral density of the (A) total hip, (B) lumbar spine, (C) femoral neck, and (D) one-third radius in subjects transitioning to denosumab or continuing on alendronate therapy. Values are least-squares means; error bars are 95% confidence intervals. ^ap < .05; ^bp < .025; ^cp < .01 based on a repeated-measures model.

At the lumbar spine, denosumab increased BMD by 3.03% (95% CI 2.63%−3.44%) at month 12, and alendronate increased BMD by 1.85% (95% CI 1.44%−2.26%), for a difference of 1.18% (95% CI 0.63%−1.73%), with the lower limit of the confidence interval excluding the noninferiority margin of −0.22 and demonstrating the noninferiority of denosumab versus alendronate at this skeletal site (see Fig. 2). The increase in BMD with denosumab versus alendronate was statistically significant (p < .0001). Significantly greater increases in BMD with denosumab compared with alendronate also were observed at month 12 at the femoral neck and one-third radius (p ≤ .0121; see Fig. 2). Furthermore, significant BMD increases for denosumab compared with alendronate were observed as early as month 6 at the lumbar spine and all measured femoral sites (p < .05; see Fig. 2).

Least-significant change (LSC) in the BMD measurements was calculated from the duplicate DXA measurements obtained at baseline for the total hip and lumbar spine. The calculated LSC at the total hip was 3.24%, and the calculated LSC at the lumbar spine was 3.29%. At month 12, more subjects in the denosumab group than in the alendronate group had BMD gains at the total hip (23% denosumab versus 14% alendronate, p < .0001) and lumbar spine (41% denosumab versus 26% alendronate, p < .0001) that exceeded the calculated LSC for that skeletal site.

Additional analyses were conducted to assess factors that might influence BMD changes when transitioning to denosumab. When BMD changes were analyzed according to time-on-prior alendronate strata, the denosumab groups showed larger BMD gains than the alendronate groups, but the patterns varied somewhat among the hip, spine, femoral neck, and one-third radius sites (Fig. 3). At the total hip, there was a pattern of larger BMD increases with denosumab among subjects who had a shorter duration of prior alendronate therapy (see Fig. 3A). This pattern was not uniform for the other skeletal sites tested (see Fig. 3B–D). BMD changes were also analyzed by subgroups according to quartiles of CTX-I levels at baseline. With denosumab, BMD increases at month 12 at the total hip and lumbar spine were largest and differed most from the alendronate group among subjects in the two highest quartiles of baseline CTX-I and PINP levels. BMD changes with alendronate generally were similar across quartiles of baseline bone turnover marker levels.

Figure 3. Percentage change in bone mineral density at the (A) total hip, (B) lumbar spine, (C) femoral neck, and (D) one-third radius according to prior alendronate use strata. Values are least-squares means; error bars are 95% confidence intervals ^ap < .05; ^bp < .025; ^cp < .01 based on a repeated-measures model.

Biochemical markers of bone turnover

In both groups, median serum CTX-I levels at baseline were in the lower portion of the premenopausal reference range,38 with 47% of subjects randomized to alendronate and 51% of subjects randomized to denosumab having serum CTX-I values less than the lower limit of the referenced premenopausal range (0.2 ng/mL), as is expected in women receiving alendronate. In the continued alendronate group, CTX-I levels remained close to their study baseline level throughout the trial, as expected for ongoing alendronate therapy (Fig. 4A). In contrast, CTX-I levels decreased significantly by day 5 to 0.05 ng/mL in response to denosumab treatment. The reduction remained stable at months 1 and 3, followed by an attenuation of the reduction at month 6. A similar dynamic pattern of CTX-I reduction was observed following the second dose of denosumab (see Fig. 4A). At month 1, approximately half the denosumab group had CTX-I levels below the limit of quantification compared with 5% of the alendronate group. By month 6, CTX-I levels had returned to the quantifiable range for 84% of these denosumab subjects and all the alendronate subjects. At the end of the 12-month trial, 34% and 54% of subjects treated with denosumab and alendronate, respectively, had serum CTX-I values within or above the premenopausal range.

Figure 4. Serum levels of CTX-I and PINP at the indicated study visits. Values are medians; error bars show the interquartile range (Q1, Q3). ^ap < .0001 by Wilcoxon rank-sum test. Shaded area indicates the premenopausal reference range.38

Baseline PINP data indicated that 31% of patients randomized to either alendronate or denosumab had baseline values of less than the lower limit of the premenopausal reference range (17 µg/L). Median PINP levels in the alendronate group decreased modestly over months 3 to 9 and returned to baseline at month 12. In the denosumab group, PINP levels decreased more slowly than did CTX-I values, reaching a nadir at month 3, with an attenuation of the reduction at month 6. A similar pattern occurred following the second dose of denosumab (see Fig. 4B). After 1 year of therapy, 67% and 25% of patients treated with alendronate and denosumab, respectively, had PINP values within or above the premenopausal range.

Adverse events

Overall, a similar number of subjects in each treatment group reported adverse events during the study (78% denosumab, 79% alendronate; Table 2). The most frequent adverse events in the denosumab and alendronate groups, respectively, were nasopharyngitis (13.4% and 10.8%), back pain (10.7% and 11.6%), bronchitis (6.3% and 5.6%), arthralgia (5.9% and 10.4%), constipation (5.1% and 4.8%), and pain in an extremity (4.7% and 8.4%). Serious adverse events were reported in 5.9% of denosumab-treated subjects and 6.4% of alendronate-treated subjects (see Table 2). The incidence of serious adverse events of infections and neoplasms was similar between groups. One subject in the denosumab group died during the study from a stroke, which, in the investigators' opinion, was unrelated to study treatment.

Table 2. Summary of Adverse Events

	Alendronate (N = 249) n (%)	Denosumab (N = 253) n (%)	P Valuea
a p Values are based on Fisher's exact test. b On-study clinical fractures were as follows: Denosumab: 2 foot, 2 wrist, 1 radius, 1 fibula, 1 humerus, 1 pelvis, 1 rib, 1 tibia; alendronate: 1 foot, 1 wrist, 1 radius, 1 sacrum.
Any adverse event	196 (78.7)	197 (77.9)	0.8294
Serious adverse events	16 (6.4)	15 (5.9)	0.8546
Leading to study discontinuation	2 (0.8)	3 (1.2)	1.0000
Death	0 (0.0)	1 (0.4)	1.0000
Selected adverse events
Clinical fracturesb	4 (1.6)	8 (3.2)	0.3820
Gastrointestinal-related disorders	60 (24.1)	58 (22.9)	0.8333
Infections	93 (37.3)	111 (43.9)	0.1465
Neoplasms (benign or malignant)	9 (3.6)	9 (3.6)	1.0000
Selected serious adverse events
Infections	3 (1.2)	1 (0.4)	0.3695
Neoplasms (benign or malignant)	3 (1.2)	3 (1.2)	1.0000

Mean serum calcium concentrations remained in the normal range throughout the study in subjects transitioned to denosumab and in those continuing on alendronate therapy. One subject in the denosumab group experienced a single serum calcium measurement of 7.9 mg/dL. This decrease was transient and was not associated with symptoms. No other subjects in the denosumab group and no subjects in the alendronate group had serum calcium levels below 8.0 mg/dL. Serum phosphorus levels followed a pattern similar to serum calcium levels in the denosumab group and decreased transiently following each dose. No subject in either group experienced clinical adverse events related to hypophosphatemia. No notable changes in other serum chemistry or hematology assessments were observed in either treatment group. No subjects were positive for anti–denosumab-binding antibodies during the study.

This study was not designed to evaluate fracture endpoints. Clinical fractures reported as adverse events occurred in 8 denosumab and 4 alendronate subjects during the 12-month treatment period (P = .38; see Table 2). The distribution and types of fractures reported were typical for postmenopausal women with low bone mass. Of subjects who experienced fractures while on study, 2 of 4 alendronate subjects and 6 of 8 denosumab subjects had a history of previous osteoporosis-related fractures. During the 1-month run-in period while all subjects were receiving alendronate, 4 subjects experienced clinical fractures, including 2 hip fractures. Three of these subjects were not randomized, and 1 subject (rib fracture) was assigned to the denosumab group.

Discussion

In this trial in postmenopausal women with low BMD who were previously taking alendronate, transitioning to denosumab was found to increase BMD and reduce markers of bone turnover to a greater extent than continued alendronate therapy with no clinical hypocalcemic events and a similar adverse-event profile. Notably, denosumab treatment resulted in significantly greater gains in BMD than continued alendronate treatment at all skeletal sites evaluated. Our findings in this head-to-head study in women receiving prior long-term alendronate treatment are similar to results of a previous head-to-head study in women who were naive to osteoporosis therapy who demonstrated greater increases in BMD with denosumab compared with alendronate at the total hip and other sites.18

The modest increase in BMD and stable levels of bone turnover markers observed in the alendronate group in this study are consistent with what is expected for women who have been receiving alendronate therapy for several years.22, 23 A recent study showed that when subjects were transitioned from long-term alendronate to zoledronic acid, bone resorption markers exhibited an initial small decrease following zoledronic acid infusion and then gradually increased above baseline, whereas BMD remained essentially unchanged in both groups over the 12-month study.23 In contrast, transitioning from long-term alendronate treatment to denosumab produced a significant and dynamic reduction in bone turnover that was characterized by a rapid initial decrease followed by a gradual attenuation of effect toward the end of the dosing interval. This bone turnover marker profile differs from the stable, unchanging suppression seen with long-term alendronate treatment. Reductions in bone turnover with denosumab were accompanied by BMD increases that were nearly twice as large as with continued alendronate therapy.

Alendronate and other nitrogen-containing bisphosphonates bind with high affinity to bone and are subsequently taken up by osteoclasts, leading to disruption of bone-resorbing capability and osteoclast apoptosis (reviewed in ref. 24). Although denosumab also inhibits bone resorption, its mechanism of action is distinct from that of alendronate. Denosumab is a fully human monoclonal antibody that inhibits RANKL. Inhibition of RANKL reduces osteoclast activity and survival, similar to bisphosphonates, but unlike bisphosphonates, RANKL inhibition also targets osteoclasts at a more immature stage, preventing their maturation and activation before they adhere to the bone matrix.5–9 These differences in the mechanisms of action between agents may contribute to the larger increases in BMD and decreases in bone turnover seen with denosumab treatment compared with alendronate in this study.

As expected in view of their prior alendronate therapy, subjects in this trial had lower baseline levels of bone turnover markers than did subjects in previous denosumab studies.10, 12, 18 Owing to persistent effects of alendronate in bone, women in the denosumab group may well have been affected by two antiresorptives for part or all of the study. However, the median absolute values for serum CTX following denosumab transition were similar to values observed in denosumab-treated women without long-term exposure to bisphosphonates.10, 12

The 1-year percentage changes in BMD observed with denosumab were less than those seen with denosumab in studies of bisphosphonate treatment-naive women10, 12, 18; however, BMD gains were larger than have been reported when women switch treatment from alendronate to another bisphosphonate.23 The smaller change in BMD with denosumab in this study as compared with other studies with bisphosphonate treatment-naive women likely reflects a situation wherein the total remodeling space had already been reduced during treatment with alendronate, a hypothesis that also may explain the observation of larger BMD increases at the total hip and one-third radius with denosumab in women with the shortest previous use of alendronate therapy compared with those with a longer period of prior alendronate use. BMD gains with denosumab at the lumbar spine, however, were similar regardless of prior length of exposure to alendronate.

Denosumab and alendronate had similar safety profiles in this study. The initial decrease in serum calcium levels that occurred following transition to denosumab was similar to previous observations in treatment-naive subjects.12 The reductions were mild and transient, and no clinical consequences were observed. Adverse events and serious adverse events occurred with similar types and frequencies in the two groups. Although two of the previous studies in postmenopausal women reported a greater incidence in serious adverse events of infection for denosumab compared with placebo,10, 13 serious adverse events of infection were balanced between the groups in this trial and in two larger studies.14, 18

A strength of this study is that its design attempts to reflect a situation that is relevant to the clinical setting. Most clinical trials of osteoporosis medications are conducted in populations that have minimal previous exposure to osteoporosis therapies, but in clinical practice, approximately half of patients stop taking osteoporosis medication within the first year25, 26 and may require a switch to a different therapy. As with any chronic, multifactorial disease, having multiple treatment options available is important, especially therapies that have the potential to improve adherence. Several studies have shown that relative fracture risk is lower among patients who have high adherence with therapy.27–30 Subcutaneous denosumab given twice yearly would ensure adherence for 6 months after each dose.

Alendronate has been the most extensively used bisphosphonate for osteoporosis treatment and, as such, was chosen as a representative of the class. Our study did not compare fracture reduction efficacy, and a much larger study size would be needed to do so.31, 32 Nonetheless, the head-to-head design of the study allowed a direct comparison with alendronate of the effects of denosumab on BMD and bone turnover markers. These surrogates have been associated with the fracture risk reductions seen with different osteoporosis therapies, although the strength of the association varies as a function of treatment and with the type of analysis.33–37 In addition, this study provides further information on relative side effects between antiresorptive therapies.

In conclusion, these results demonstrate that postmenopausal women with low bone density may be safely transitioned from weekly oral alendronate to 6-monthly subcutaneous denosumab to achieve an incremental increase in bone mass.

Disclosures

This study was sponsored by Amgen Inc. Dr. Kendler is an investigator for Merck, Amgen, Eli Lilly, Novartis, Takeda, GlaxoSmithKline, Pfizer, Servier, Biosante, and Wyeth and has served as a speaker, consultant, or advisor for and/or received honoraria from Merck, Amgen, Eli Lilly, Novartis, Servier, Nycomed, and Wyeth. Professor Roux is an investigator for Amgen and has served as a consultant for and/or received honoraria or research funding from Amgen, Roche, Merck Sharp & Dohme, Alliance for Better Bone Health, Novartis, Servier, Lilly, and Wyeth. Professor Benhamou is an investigator for Amgen and has served as a consultant and/or investigator for Amgen, Lilly, Merck Sharp & Dohme, Novartis, Alliance for Better Bone Health, Pierre Fabre, Servier, and Wyeth. Dr. Brown is an investigator for Amgen and has served as a consultant for and/or received honoraria or research funding from Abbott, Amgen, Arthrolab, Bristol Myers Squibb, Eli Lilly, Genizon, GlaxoSmithKline, Merck Frosst, Nicox, Novartis, Pfizer, Procter & Gamble, Roche, Sanofi-aventis, Servier, Wyeth, and Zelos. Dr. Lillestol is an investigator for Amgen and reports financial disclosures for Alexion, Amgen, Astra/Zeneca, Bausch & Lomb, BioSante, Boehringer Ingelheim, Bristol Myers Squibb, CombinatoRx, Covance, Daiichi Sankyo, DP Clinical, Endo Pharmaceuticals, Forest, GlaxoSmithKline, Hisamitsu, i3 Research, Lilly, Novartis, Novo Nordisk, NPS Allelix, NPS Pharmaceuticals, Otsuka, Pfizer, PPD, Quintiles, Roche, Sanofi-Aventis, Schering-Plough, Sepacor, Smith Kline Beecham, Takeda, Viropharma, and Wyeth. Dr. Siddhanti, Ms. Man, and Dr. San Martin are full-time employees of Amgen Inc., and may own stock or stock options in Amgen Inc. Dr. Bone is an investigator for Amgen, Eli Lilly, Merck, Nordic Biosciences, Takeda, and Zelos; has served as a consultant for Amgen, Merck, Nordic Bioscience, Osteologix, Pfizer, Takeda, and Zelos; and has received speaker honoraria from Merck and Novartis.

Role in study conduct and manuscript preparation

Dr. Kendler participated in the acquisition and interpretation of data and the writing of the manuscript with the assistance of a medical writer and reviewed and approved the final version. Dr. Bone participated in the planning and design of the study, acquisition and interpretation of the data, and writing of the manuscript with the assistance of a medical writer and reviewed and approved the final version. Drs. Roux, Benhamou, Brown, and Lillestol participated in acquisition and interpretation of the data and critically reviewed and approved the outline and progressive drafts of the manuscript and reviewed and approved the final version. Drs. Siddhanti and San Martin participated in study implementation and data interpretation. They participated in writing the manuscript and critically reviewed and approved the progressive drafts of the manuscript, including the final version. Ms. Man participated in the study design and provided statistical expertise for data interpretation. She critically reviewed and approved the progressive drafts of the manuscript, including the final draft.

Acknowledgements

Amgen Inc., sponsored this study. Holly Brenza Zoog, PhD, provided medical writing assistance. Nate Mercaldo, MS, and Frank Harrell, PhD, of Vanderbilt University Department of Biostatistics performed the independent statistical analysis.

The following were principal investigators for the STAND trial: Michael Bolognese, Henry Bone, Robin Dore, Catherine Gerrish, John Hoekstra, Lynn Kohlmeier, Angelo Licata, Michael Lillestol, David Mandel, Steve Mattson, Harris McIlwain, Eric Orwall, Munro Peacock, Kent Rogers, Lisa Vendland, John Ruckle, Nathan Wei, Grattan Woodson, Chris Recknor, Jacques Brown, Anthony Hodsman, Terri Paul, David Kendler, Chui Kin Yuen, Ingrid Kull, Katre Maasalu, Ivo Valter, Claude Laurent Benhamou, Bernard Cortet, Phillipe Orcel, Christian Roux, Georges Weryha, Gerolamo Bianchi, Maria Luisa Brandi, Sergio Ortolani, Giancarlo Isaia, Slawomir Jeka, Grzegorz Kania, Justyna Klimkiewicz, Andrzej Sawicki, Anna Sidorowicz-Bialynicka, Jerzy Supronik, and Witold Tlustochowicz.

Portions of the data in this paper have been presented previously at the ASBMR and ACR 2008 and ECCEO 2009 annual congresses.