In men with prostate cancer, androgen deprivation therapy (ADT) decreases bone mineral density (BMD) and increases the risk of clinical fractures.1, 2 Treatment-related bone loss is accompanied by an increase in the markers of bone resorption (eg, serum concentration of type 1 C-telopeptide [sCTX]), tartrate-resistant alkaline phosphatase 5b [TRAP-5b]), and bone formation (eg, intact N-terminal propeptide of type I procollagen (P1NP).3, 4 Significant evidence of bone resorption, as indicated by increases in sCTX and TRAP-5b, was reported in men with prostate cancer receiving ADT for 35 months.5 In normal men receiving a gonadotropin-releasing hormone (GnRH) agonist, P1NP levels initially decreased, but then increased at 12 weeks of treatment.6 Historically, therapies for treatment-induced bone loss in men with early-stage prostate cancer have included bisphosphonates such as intravenous zoledronic acid and oral alendronate, although none are approved for that indication.7, 8
Denosumab is a fully human monoclonal antibody against receptor activator of NF-κB ligand (RANKL), a key activator of osteoclast formation, function, and survival. By binding and inactivating RANKL, denosumab inhibits osteoclast function and bone resorption. Denosumab (Prolia, Amgen Inc., Thousand Oaks, CA, USA) is approved in the United States for the treatment of bone loss in early-stage prostate and breast cancer patients receiving hormone ablation therapy. In two phase 2 studies of patients with breast cancer, prostate cancer, or other solid tumors and bone metastases, patients treated with denosumab every 4 weeks experienced a rapid reduction of the bone turnover markers (BTMs) sCTX, TRAP-5b, and P1NP, which was sustained over the 25 weeks of the studies.9 Denosumab also suppressed bone turnover, indicated by decreased levels of sCTX and P1NP over 36 months, in a phase 3 study of denosumab administered every 6 months to postmenopausal women with osteoporosis.10
Although elevated BTMs are associated with increased risk for skeletal complications in men with advanced prostate cancer,11 a prognostic role for BTMs in men with early-stage prostate cancer treated with ADT has yet to be established.12 Among these men, those who are older and those with a longer duration of hormonal treatment have been shown to have a greater risk for skeletal events.2, 13 Whereas early decreases in BTMs have been shown to predict long-term BMD responses to bisphosphonate therapy among women with osteoporosis,14 similar data are lacking in men with early-stage prostate cancer.
In a phase 3, randomized, double-blind study of men who were receiving ADT for nonmetastatic prostate cancer, denosumab significantly increased BMD at all measured sites and significantly reduced new vertebral fractures by 62% over 3 years.15 Adverse events in this study were representative of the study population and balanced between the study arms.15 In this report, we describe results from a protocol-specified exploratory analysis of longitudinal changes in BTMs in the phase 3 clinical trial. In order to assess therapy outcomes using potentially predictive baseline demographic and clinical variables, the analysis includes results for the overall study population as well as subgroup analyses based on baseline age, duration of prior ADT, and baseline BTM values. Additionally, we performed a correlative analysis of lumbar spine and total hip BMD responses at 36 months with BTM measurements at 6 months.
Materials and Methods
This was an international, multicenter, randomized, double-blind, placebo-controlled study. The study was conducted in accord with the Declaration of Helsinki and the International Conference on Harmonization Tripartite Guideline on Good Clinical Practice. Complete eligibility criteria have been reported previously.15 Briefly, eligible patients were men with histologically confirmed prostate cancer. They were aged ≥ 70 years old, or if <70 years of age, had a history of osteoporotic fracture or a BMD T-score < −1.0 at the lumbar spine, total hip, or femoral neck. Patients had an Eastern Cooperative Oncology Group (ECOG) performance status of 0, 1, or 2; had undergone bilateral orchiectomy or had initiated luteinizing hormone-releasing hormone (LHRH) treatment and were expected to continue therapy for ≥ 12 months. Patients with other concurrent anti-neoplastic therapy or radiotherapy; prostate specific antigen [PSA] >5 mg/mL after being on ADT for >1 month; or BMD T-score < −4.0 at the lumbar spine, total hip, or femoral neck were excluded.
We stratified patients at baseline by age group (< 70 vs ≥ 70 years old) and duration of ADT (≤ 6 vs > 6 months) and randomly assigned them to receive placebo or denosumab 60 mg subcutaneously every 6 months for 36 months. We advised all patients to take ≥ 1000 mg supplemental calcium and ≥ 400 IU vitamin D daily. The primary endpoint of the study was the percentage change in lumbar spine BMD from baseline to month 24; results have been reported.15 In this report we describe results for a prespecified exploratory endpoint: the percentage change from baseline in the bone turnover markers sCTX, TRAP-5b, and P1NP at months 1, 6, 12, 24, and 36. We also describe the results of an ad hoc analysis evaluating the relationship between changes in these bone turnover markers at 6 months and changes in BMD at 36 months.
Bone turnover markers were assessed by a central laboratory at Amgen, Inc. (sCTX) or a specialty laboratory (P1NP, TRAP-5b; Covance Laboratories, Chantilly, VA, USA). We evaluated sCTX using the Serum CrossLaps® enzyme-linked immunosorbent assay (ELISA; Nordic Bioscience Diagnostics A/S, Herlev, Denmark; interassay coefficient of variation [CV] = 4% to 12%). For P1NP assessments we used the P1NP RIA kit (Orion Diagnostica, Espoo, Finland; CV = 9% to 18%). We used the BoneTRAP® Assay Kit (IDS, Inc., Boldon, UK; CV = 9% to 13%) for TRAP-5b assessments.
In the analyses of bone turnover markers, we included all patients who were randomized and for whom baseline and at least one postbaseline measurement of the endpoint of interest were available. We measured fasting values for sCTX, TRAP-5b, and P1NP at baseline and 1 month postdose, and predose (trough level) at months 6, 12, 24, and 36. The limits of quantification for sCTX, TRAP-5b, and P1NP were 0.049 ng/mL, 1.3 U/L, and 10 µg/L, respectively. Subgroup analyses evaluated bone turnover marker response at 1, 6, 12, 24, and 36 months according to age (< 70 years versus ≥ 70 years), prior duration of ADT (≤ 6 months versus >6 months), and baseline marker values (≤ median versus > median). We used the van Elteren stratified rank test16 to compare treatment groups for percent change in bone turnover markers, using the stratification variables of age group and duration of ADT or time since orchiectomy at study entry. We calculated Spearman correlation coefficients to evaluate the relationship between changes in BTM values at 6 months (sCTX, TRAP-5b, and P1NP) with changes in BMD at 36 months for the lumbar spine and total hip.
A total of 1468 men receiving ADT for nonmetastatic prostate cancer were assigned randomly to denosumab (n = 734) or placebo (n = 734).15 Sixty-two percent of patients (64% denosumab, 61% placebo) completed the 36-month study. The most frequently cited reason for discontinuation was withdrawal of consent (17% denosumab, 19% placebo), most of which occurred when patients declined to consent to an additional year in the study when it was extended from 24 to 36 months in a protocol amendment.
As described previously,15 baseline demographic and disease characteristics were well balanced between the two groups. Mean age was 75 years; 83% in each treatment group were aged ≥ 70 years. Most men were white and had received ADT for >6 months at study entry (median 20.4 months denosumab and 20.8 months placebo). Baseline BTMs for both treatment groups are summarized in Table 1.
Table 1. Baseline Bone Turnover Markers
Bone turnover markers at baseline
Placebo (N = 734)
Denosumab (N = 734)
Values are median (Q1, Q3).
N = number of subjects randomized; n = number of subjects with observed data; sCTX = serum type 1 C-telopeptide; TRAP-5b = tartrate-resistant alkaline phosphatase 5b; P1NP = procollagen-1 N-terminal telopeptide; Q1 = first quartile; Q3 = third quartile.
Normal human serum range in men: below the lower limit of quantification (0.049 ng/mL) to 0.475 ng/mL.
Normal human serum range in men: 2.804 U/L to 5.196 U/L.
Normal human serum range in men: 36.305 µg/L to 97.531 µg/L.
Changes in bone turnover markers in the overall study population
In this population of men receiving ADT for nonmetastatic prostate cancer, denosumab treatment resulted in a rapid, sustained decrease of BTMs over time that was statistically significant compared with placebo for all markers. The median (first quartile [Q1], third quartile [Q3]) change in sCTX at month 1 was −90% (−93%, −85%) in the denosumab group, compared with −3% (−21%, +19%) in the placebo group (p < 0.0001; Fig. 1). At the end of the first dosing interval at 6 months, statistically significant suppression of sCTx continued, with median (Q1, Q3) changes in sCTX of −65% (−80%, −42%) for the denosumab group and −7% (−27%, +31%) for the placebo group. Significantly greater suppression of sCTX in the denosumab group than in the placebo group continued through the end of the last dosing interval at month 36 (p < 0.0001; Fig. 1).
The median (Q1, Q3) change in TRAP-5b at month 1 was −55% (−63%, −45%) in the denosumab group and −3% (−12%, +6%) in the placebo group (p < 0.0001; Fig. 2). As with sCTX, suppression of TRAP-5b continued through month 36 to a statistically significantly greater degree in the denosumab group than in the placebo group (p < 0.0001; Fig. 2.)
Median (Q1, Q3) change in P1NP after 1 month of treatment was −27% (−36%, −15%) in the denosumab group, compared with a median change of 1% (−9%, +14%) in the placebo group (p < 0.0001; Fig. 3). Maximal median suppression of P1NP was observed at month 12 and maintained to a statistically significant degree through month 36 (Fig. 3).
Changes in bone turnover markers in patient subgroups
At 1 month following the first dose, denosumab consistently reduced bone turnover markers across all subgroups studied, including subgroups based on age (Fig. 4), prior duration of ADT (Fig. 5), and bone turnover marker levels at baseline (Fig. 6). Compared with placebo, these differences were statistically significant for all bone turnover markers and all subgroups (p < 0.0001). Similar statistically significant consistency of treatment effect was observed across all subgroups at 6, 12, 24, and 36 months (data not shown).
Changes in bone turnover markers related to changes in BMD
Changes in BTMs at 6 months were associated with changes in BMD at 36 months. In general, correlations between BTM decreases and gains in BMD were stronger in the denosumab group compared with the placebo group (Table 2).
Table 2. Spearman Correlation Between BMD Change from Baseline at Month 36 and Change in sCTX, P1NP, and TRAP-5b at Month 6
Placebo r (n)
Denosumab r (n)
Placebo r (n)
Denosumab r (n)
r = Spearman correlation; n = number of randomized subjects; BMD = bone mineral density; sCTX = serum type 1 C-telopeptide; TRAP-5b = tartrate-resistant alkaline phosphatase 5b; P1NP = procollagen-1 N-terminal telopeptide.
In this recently reported phase 3, randomized, double-blind study of men who were receiving ADT for nonmetastatic prostate cancer, denosumab significantly increased BMD at all measured sites and significantly reduced new vertebral fractures.15 In the current prespecified analysis of data from that study, denosumab induced rapid inhibition of bone turnover over time, with significant decreases in sCTX, TRAP-5b, and PINP evident at 1 month and thereafter. The consistency of the antiresorptive effect is reflected by the narrow interquartile range around the 1-month bone resorption marker levels and indicates high pharmacodynamic response rates across patients (Figs. 1 and 2). The rapid and substantial inhibition of bone turnover is consistent with the significant increases in BMD at all skeletal sites observed after 1 month of therapy—the earliest measured time point in the study. Reductions in BTMs at 6 months were also associated with BMD gains at the lumbar spine and total hip at 36 months (Table 2). Suppression of bone turnover was sustained throughout the 36 months study; this finding is also consistent with the continued gains in BMD at all measured time points.
sCTX and TRAP-5b decreased sharply from baseline to month 1 and then increased slightly at later time points. Peak marker inhibition at this early time point may reflect timing of measurements relative to drug dosing. The month 1 measurement was taken approximately 30 days after the first denosumab dose in an every-6-month dosing regimen, while the measurements at months 6, 12, 24, and 36 reflected trough levels of denosumab, measured at the end of the 6-month dosing cycle. Because BTMs were not evaluated between subsequent treatment cycles, we cannot determine whether each denosumab treatment resulted in similar sharp decreases in BTMs. Nevertheless, these findings suggest that the effects of denosumab on bone are reversible.17
In men receiving ADT for prostate cancer, two prior studies reported older age and longer ADT duration are associated with greater risk of clinical fractures.2, 13 However, the effects on BTMs for these subgroups were not reported. In the current analyses, denosumab significantly decreased BTMs in all subgroups including men aged 70 years or older, men with an ADT duration at baseline longer than 6 months, and men with higher baseline BTM values (Figs. 4–6). The significant treatment effects observed in each of these subgroups suggests that denosumab is effective in men at greatest risk for fracture. These results are congruent with previous findings that denosumab consistently increased BMD at all measured skeletal sites and all subgroups regardless of age, ADT duration, or baseline levels of BTMs.15
RANKL is essential for osteoclast formation, function, and survival. RANKL expression is increased in high bone turnover states. By binding to RANKL and thereby inhibiting osteoclast activity, denosumab significantly decreased bone turnover, as indicated by decreased levels of sCTX, TRAP-5b, and P1NP. Although marked inhibition was observed regardless of baseline levels of bone turnover and the differences between subgroups were small, the median decreases in BTM were greatest for subjects with the highest BTM values at baseline. These observations highlight the important role of RANKL in regulating osteoclast activity and the efficacy of RANKL inhibition in decreasing bone turnover, even in high bone turnover states.
Because patients were excluded from this study if they had received either 3 years continuous treatment with oral bisphosphonates or intravenous (IV) bisphosphonates within 5 years prior to enrollment, we were unable to assess the effect of prior antiresorptive therapy on subsequent BTM responses to denosumab. Although both bisphosphonates and denosumab have been shown to decrease BTMs and increase BMD in prostate cancer patients,8, 18, 19 denosumab has a distinct mechanism of action. In contrast to bisphosphonates, which accumulate in bone where they inhibit osteoclast recruitment, function, and survival,20 denosumab prevents the formation, maturation, and survival of osteoclasts.21 The recent comparison of results from two phase 2 studies suggested a potential treatment difference based on previous bisphosphonate exposure.9 Responses to treatment among women with metastatic breast cancer who had no previous IV bisphosphonate treatment were compared with patients with advanced solid tumors or multiple myeloma who had elevated levels of the bone turnover marker urinary N-telopeptide at study entry despite at least 8 weeks of previous IV bisphosphonate treatment. Although both denosumab and zoledronic acid produced reductions in bone turnover, in patients who had been previously treated with bisphosphonates, suppression of TRAP-5b at study week 25 was substantially greater with denosumab than with zoledronic acid (73% versus 11%). This difference suggests that in some patients, osteoclast function can persist despite bisphosphonate treatment, and that identification of these patients through BTM assessment early in therapy could lead to improved treatment outcomes.
In men receiving ADT for nonmetastatic prostate cancer, denosumab resulted in a rapid and sustained decrease of BTMs, including the bone resorption markers sCTX and TRAP-5b and the bone formation marker P1NP, compared with placebo. The rapid and sustained suppression of BTMs was consistent with previously reported increases in BMD and reduction in new fractures in this population. Denosumab provides a new therapeutic option to prevent fractures in men receiving ADT for prostate cancer.
MRS has served as a consultant to Amgen. FS is an investigator in this study and is a consultant and advisory board member, for Amgen, Novartis, and Sanofi Aventis. BE is an investigator in this study and is a consultant, advisory board member, and speaker for Amgen, Bayer, and Pfizer. PS is an investigator in this study and a consultant and advisory board member for Amgen. TLJT is an investigator in this study and is a consultant, advisory board member, and/or speaker for Amgen, Orion Pharma, AstraZeneca, Sanofi-Aventis, Astellas, and GSK. BZL is a consultant for Amgen and has received research funding from Amgen and Lilly. CK and CG are employees of Amgen Inc. and have received Amgen stock/stock options.
We thank the patients and investigators at all participating sites who made important contributions to this study. We also thank Sue Hudson and Geoffrey Smith, PhD, supported by Amgen, for medical writing assistance in preparing this manuscript. Funding for the study and financial support for the preparation of the manuscript was provided by Amgen Inc. This study is registered with ClinicalTrials.gov, with the identifier NCT00089674. Dr. Smith is supported by an NIH Midcareer Investigator Award (5K24CA121990) and competitive research awards from the Prostate Cancer Foundation.
Authors' roles: Study design: MRS, CG, CK. Study conduct: MRS, FS, BE, PS, TLJT, and CG. Data collection: MRS, FS, BE, PS, TLJT, and CG. Data analysis: MRS, FS, BE, PS, TLJT, BZL, CK, and CG. Data interpretation: MRS, FS, BE, PS, TLJT, BZL, and CG. Drafting manuscript: MRS, FS, BE, PS, TLJT, BZL, CK, and CG. Revising the manuscript content: MRS, FS, BE, PS, TLJT, BZL, CK, and CG. Approving final version of manuscript: MRS, FS, BE, PS, TLJT, BZL, CK, and CG. MRS takes responsibility for the integrity of the data analysis.