Denosumab is a monoclonal antibody that binds receptor activator of nuclear factor kappa-β ligand (RANKL). Its development followed the discovery and description of the central role of the RANKL–RANK pathway in osteoclastogenesis (see also Chapter 7 , Osteocytes). RANKL, a member of the tumor necrosis factor (TNF) for family of growth regulators, is expressed by osteoblasts and osteocytes. The binding of RANKL to RANK, a cell surface receptor on preosteoclasts, is required for the differentiation, proliferation, maturation, activation, and survival of osteoclasts. Osteoprotegerin (OPG), a soluble RANK receptor, is secreted by osteoblasts and serves as a natural inhibitor of RANKL stimulation of osteoclasts by blocking the RANKL–RANK interaction.
Preclinical studies documented the pivotal role of this pathway in skeletal biology. Mice lacking RANKL or RANK develop high bone mass, whereas mice with OPG deficiency or RANKL excess develop an osteopenic phenotype . The few human cases with inactivating mutations of the RANKL gene have presented with high bone mass (osteopetrosis) and lack of osteoclasts . The crucial role of OPG in humans was demonstrated by a case of rapidly progressive osteoporosis with very high bone remodeling in a young man with celiac disease who developed autoantibodies that neutralized OPG . Since its discovery, upregulation of RANKL has been implicated in several metabolic bone disorders characterized by increased bone resorption, including estrogen deficiency, hyperparathyroidism, multiple myeloma, metastatic bone disease, and Paget disease of bone .
The elucidation of the RANKL pathway in skeletal health provided attractive targets for possible treatments for osteoporosis and other metabolic bone disorders. OPG administration to rats and monkeys inhibited bone resorption, prevented bone loss due to estrogen deficiency, and increased bone mass, indicating that RANKL inhibition had therapeutic potential . This possibility was first explored in humans when an OPG-Fc fragment construct was administered in single doses of 0.1–3 mg/kg to healthy volunteers . Rapid reduction in urinary levels of N-telopeptide of type I collagen (NTX) occurred. All doses reduced NTX levels by at least 40% at 12 hours after dosing. Duration of the antiresorptive effect was dose-dependent. With the largest dose, NTX values returned to baseline after about 2 months. Nonneutralizing antibodies to OPG were noted in one subject. While the study proved that inhibiting RANKL reduced osteoclastic activity, OPG was not evaluated further as a treatment option, in part because of concern about the possibility of inducing anti-OPG antibodies.
Denosumab is a fully human monoclonal IgG2 antibody that binds to RANKL with even greater affinity (Kd=3×10 −12 M) and specificity than does OPG . No in vitro cross-reactivity was observed with other TNF family members, including TRAIL, TNFβ, TNFα, or CD40L. Its effect of inhibiting the binding of RANKL to RANK mimics that of OPG. Denosumab has a bioavailability of 64% after subcutaneous dosing . It is thought that denosumab is cleared, as are other immunoglobulins, by the reticuloendothelial system with minimal renal excretion. Nonlinear pharmacokinetics (PKs) were observed in the first in-human studies of denosumab. Single doses from 0.01 to 3.0 mg/kg were given by subcutaneous injection to healthy postmenopausal women . Maximal serum concentrations of denosumab were reached at 5–21 days, earlier with the smaller doses. Clearance rates were slower with larger doses and were modestly affected by body weight but not by age or renal function . Urinary NTX values decreased rapidly, up to 80% by 12 hours. The nadir of NTX was reached by 1 month, and reversal of the reduced bone turnover was observed with all doses. The duration of the inhibition of bone resorption was dose-dependent with values remaining well below baseline at 6 months with the larger doses. Between-subject variability in pharmacodynamics (PDs) is modest, suggesting that fixed dosing is appropriate and that adjustment, based on patient demographics or renal function, is not necessary. Similar PK and PD results were observed in Japanese postmenopausal women .
These PK and PD properties allow for dosing at long intervals. In the Phase II dose-ranging study, subcutaneous doses of denosumab, 14 or 30 mg every 3 months or 14, 60, 100, or 210 mg every 6 months, were given to postmenopausal women with low bone mass . Other subjects were randomly assigned to receive placebo injections or open-label oral alendronate 70 mg/week. All doses of denosumab rapidly reduced serum levels of C-terminal telopeptide of type I collagen (CTX) by 68%–80% and urinary NTX values by 50%–60% by 3 days. Maximal decreases in the resorption markers occurred at 1 month after dosing. The duration of these effects was dose-dependent. Partial recovery of resorption markers occurred by 6 months. The average level of bone resorption markers 6 months following denosumab 60 mg was similar to that observed in the alendronate group. Very similar results were observed in a cohort of Japanese women with postmenopausal osteoporosis . Based on the PK–PD studies and the results of the Phase II trial, the subcutaneous dose of 60 mg every 6 months was chosen to evaluate the efficacy and safety of denosumab in the Phase III fracture trial.
The effectiveness of denosumab in reducing fracture risk was evaluated in an international, randomized, double-blind, multicenter trial Fracture REduction Evaluation of Denosumab in Osteoporosis every 6 Months (FREEDOM), in which 7808 postmenopausal women with osteoporosis received denosumab 60 mg or placebo every 6 months for 3 years . Subjects with prevalent grade 2 or 3 vertebral fractures or with multiple grade 1 deformities could not participate in the trial. The subjects were, on average, 71 years of age (range 60–90 years). Mean bone mineral density (BMD) T -score values in the lumbar spine and total hip regions were −2.8 and −1.9, respectively. Only 23% of subjects had grade 1 vertebral deformities at baseline. Because of the enrollment criteria, the incidence of osteoporosis-related fractures in this cohort was lower than in most previous pivotal fracture trials with other agents. After 3 years the incidences of new vertebral and hip fractures in the placebo group were 7.2% and 1.2%, respectively. Denosumab therapy reduced radiographically defined vertebral fracture risk by 68% (to 2.3%) at 3 years ( P <.001). This effect was evident after 12 months of treatment, the earliest time points at which repeat spine radiographs were obtained, when vertebral fracture risk was reduced by 61% ( P =.001). At 36 months, hip fracture risk was reduced by 40% ( P =.04), and the cumulative incidence of nonvertebral fracture was reduced by 20%, from 8.0% to 6.5% ( P =.01). Denosumab was similarly effective in reducing vertebral fracture risk in prespecified subgroups based upon age (<75 or ≥75 years), body mass index (BMI) (<25, 25–30, >30), femoral neck BMD (≤−2.5 vs >−2.5), and estimated creatinine clearance (<60 or ≥60 mL/min) . Nonvertebral fracture risk reduction was more clearly evident in subjects with lower femoral neck BMD, lower BMI, and without prevalent vertebral fracture.
Post hoc analyses evaluated the effectiveness of denosumab in high-risk subgroups based upon age, prevalent fracture history, and BMD . Vertebral fracture risk reduction was similar in higher and lower risk subgroups. However, denosumab was more effective in reducing the risk of hip fracture in the high-risk subgroups, while the effect was very modest and not statistically significant in subjects at low fracture risk.
At baseline in the FREEDOM trial, the median 10-year probability of a major osteoporotic fracture (calculated with BMD) was approximately 15% and for hip fracture was approximately 5% in both treatment groups . Adjusted for age and fracture probability, denosumab reduced the risk of clinical osteoporotic fractures to a greater extent in those at moderate-to-high risk. Denosumab decreased fracture risk by 11% ( P =.629) in women at 10% fracture probability while the reduction was 50% at 30% probability ( P =.001). Consistent with subgroup analyses, the reduction in fracture was independent of prior fracture, parental history of hip fracture, or secondary causes of osteoporosis.
The incidence of wrist fracture was low (2.9% in the placebo group) and not significantly affected by therapy in the entire FREEDOM cohort . In a post hoc analysis of the subjects with baseline femoral neck T -score −2.5 or less, wrist fracture incidence was decreased from 4.0% in the placebo group to 2.4% with denosumab (relative risk reduction 40%, P =.03).
At the end of the 3-year placebo-controlled FREEDOM study, about 4500 subjects agreed to participate in the FREEDOM Extension study . All subjects received open-label denosumab 60 mg every 6 months for an additional 7 years, and 2626 subjects completed the Extension study, including 1283 who received denosumab for the full 10 years (long-term group) and 1343 who received placebo during years 1–3 and denosumab for the subsequent 7 years (crossover group). During years 4 and 5 of the Extension study, the incidence of vertebral fractures in the cohort of 2207 women who took placebo for 3 years during FREEDOM was 0.9%/year . This was similar to the 0.7% incidence with denosumab and lower than the 2.5% yearly rate in the placebo group (from which these subjects came) in FREEDOM. This suggests that the initial response to denosumab in this crossover group was similar to the response in the group that received denosumab during FREEDOM.
The annualized incidence of vertebral fractures in the long-term group remained relatively stable and similar to the incidence observed during the FREEDOM trial through the 7 years of the Extension study . Nonvertebral fracture rates during the Extension study also remained below the rate observed in the placebo group of FREEDOM, which was 2.65 per 100 subject-years [95% confidence interval (CI) 2.23–3.15]. Among subjects receiving placebo during the first 3 years of FREEDOM, fracture rate ratios between Extension study years 4–7 (denosumab) and FREEDOM years 1–3 (placebo) were significantly lower for wrist (0.57), forearm (0.57), humerus (0.42), and upper limb (0.52) fractures ( P <.05 for all) .
Improved nonvertebral fracture protection with long-term exposure to denosumab was suggested in post hoc analyses. Nonvertebral fracture rates in both the long-term and crossover groups were significantly lower after 3 years of denosumab therapy when compared to the rates during the first 3 years of therapy . In the long-term group the nonvertebral fracture rate during years 4–10 of therapy (1.44 per 100 subject-years) was 26% lower than the rate (1.98 per 100 subject-years) during the first 3 years of treatment. The rate ratio of 0.74 (95% CI 0.60–0.93) was statistically significant ( P value=.008). Similar results were observed comparing fracture rates between years 4–7 and 1–3 of therapy in the crossover group. The lower rate of fracture after 3 years of therapy was observed in groups of subjects whose total hip BMD T -score after 3 years of denosumab therapy was ≤−2.5 and between −1.0 and −2.5. No such effect was observed in the group with T -score values >−1.0 in which the fracture rates were lower than in the other BMD subgroups.
Persistent or even improved fracture protection with long-term denosumab therapy was also suggested by comparing vertebral and nonvertebral fracture rates during the FREEDOM Extension study with rates predicted by a “virtual twin” model based on Monte Carlo simulations of placebo participants matched on baseline characteristics to the actual long-term treatment group participants using the actual placebo group data from FREEDOM . Based on this model, the estimated relative risk for new vertebral fractures on denosumab therapy during the Extension study was 0.62 (95% CI 0.47–0.80) and for nonvertebral fractures was 0.54 (95% CI 0.43–0.68).
Biochemical markers of bone turnover
The effects of denosumab therapy on markers of bone turnover have been observed consistently in several studies and in different patient populations. The dynamic effects on bone resorption markers in the Phase II study, described previously, remained evident over 8 years of continuous dosing with denosumab . Serum bone-specific alkaline phosphatase (BSAP) levels with denosumab therapy remained stable for 1 month after dosing but then decreased to reach a steady state at about 60% below baseline after about 6 months, paralleling values in patients given alendronate .
In a small substudy of FREEDOM trial, 160 women had bone marker values measured at month 1 and every 6 months for 3 years . Markers of bone resorption included serum CTX and tartrate-resistant acid phosphatase (TRAP 5b); bone formation markers were serum procollagen type I N-propeptide (P1NP) and BSAP. Decreases in CTX were more rapid and greater than that of P1NP and BSAP. One month after denosumab dosing, serum CTX values fell to below the premenopausal reference interval (PRI) in all subjects; there were no “nonresponders.” The predose values were below the PRI in 79% of subjects at month 6 and 51% at 36 months. CTX values at the end of the dosing period were influenced by baseline CTX values and the interval since the previous dose. Significant and sustained reductions in serum TRAP 5b, a marker of osteoclast number, were also noted with denosumab treatment. Bone formation markers BSAP and P1NP, followed the patterns observed in the Phase II study, decreasing later than 1 month after dosing and then remaining stable with continued treatment. CTX and P1NP remained below the PRI at all measured time points in 46% and 31% of denosumab-treated subjects, respectively, but in no placebo-treated subjects. There were significant correlations between reductions in bone turnover markers at 6 months in response to denosumab and increases in BMD at the lumbar spine and total hip at 36 months ( R values −0.24 to −0.47) . Very similar responses were observed in a Phase III study comparing denosumab 60 mg every 6 months with placebo in 336 postmenopausal women with low bone mass . In the FREEDOM Extension study, predose serum CTX and P1NP values remained stable at the lower end of the premenopausal reference range over the full 10 years of follow-up . These results demonstrate that long-term denosumab therapy was not associated with either progressive inhibition of bone remodeling or loss of effectiveness.
Bone mineral density
Increases in BMD in the spine and proximal femur have been consistently observed with denosumab in many studies. In the Phase II study, significant increases were observed as early as 1 month after dosing . At 1 year, increases from baseline of 4.6% in the lumbar spine and 3.6% in the hip region occurred with denosumab 60 mg every 6 months, while decreases of 0.8% and 0.6% were observed in the lumbar spine and total hip regions, respectively, in the placebo group. Larger doses were associated with no or minimally greater increments. In the FREEDOM Extension study, progressive increases occurred with continuous treatment. BMD values in the lumbar spine and total hip had increased by 13.7% and 7.0 %, respectively, after 5 years and by 21.7% in the lumbar spine and 9.2% in the total hip region after 10 years . The progressive increase in BMD at the hip over 10 years of therapy differs from that observed with long-term bisphosphonate therapy where a plateau in hip BMD is observed after 4–5 years . The mechanisms contributing to the continued increase in hip BMD with denosumab might include reductions in cortical porosity and persistence of modeling-based bone formation as reported in animal studies .
During the 3 years of FREEDOM, a modest (2.2%) but significant increase in BMD at the 1/3 radius was observed with denosumab while this value decreased by 1.2% in the placebo group . BMD at the 1/3 radius site then remained stable with long-term treatment, being 2.7% higher than baseline after 10 years of denosumab therapy in the BMD substudy of the FREEDOM Extension study .
In FREEDOM the percentage change in total hip BMD from baseline to 36 months accounted for a substantial portion of the effect of denosumab in reducing new or worsening vertebral fracture risk (35%, 95% CI 20%–61%) and of the reduction in nonvertebral fracture risk (87%, 95% CI 35%–>100%) . These analyses suggest a stronger correlation between changes in BMD and fracture risk reduction with denosumab than have been observed with other antiresorptive drugs. In addition, on-treatment total hip BMD, measured in each participant every year during FREEDOM and its Extension, correlated significantly with current nonvertebral fracture risk . Whether this is a unique effect of denosumab therapy, due to analytical differences between studies or is simply a matter of the larger increases with denosumab than with other drugs is not clear. These results are consistent, however, with the recent FNIH meta-regression demonstrating a strong association between changes in total hip BMD compared to placebo and reductions in fracture risk .
Comparisons with other osteoporosis drugs
When direct comparisons have been made, BMD gains with denosumab have consistently been equal to or greater than those with bisphosphonate therapy. In the Phase II study, similar responses at the lumbar spine and total hip were observed with denosumab 60 mg every 6 months and alendronate 70 mg each week . In the STAND study, a head-to-head comparison of denosumab 60 mg every 6 months and alendronate 70 mg once weekly in 594 postmenopausal women with low bone mass, Brown et al. reported greater gains with denosumab in the lumbar spine (5.3% vs 4.2%), total hip (3.0% vs 2.6%), and 1/3 radius at 12 months .
In studies of women with postmenopausal osteoporosis previously treated with bisphosphonates, increases in total hip BMD were greater after treatment with denosumab than with any of the 4 bisphosphonates studied with between-group differences at 12 months ranging from 0.85% to 1.6% . In those same studies, denosumab increased BMD at the lumbar spine by 1.2%–2.3% more than did a bisphosphonate after 12 months of treatment.
In the study of Brown et al. mentioned previously, the reduction in serum CTX was greater and much more rapid with denosumab 60 mg every 6 months than with alendronate 70 mg weekly . As was observed in the Phase II study, the mean serum CTX value rose from a nadir at 1 month after denosumab dosing to a level at 6 months that was similar to the value observed with alendronate. Reductions in P1NP were greater with denosumab but similar in timing, reaching steady-state levels by 3 months. In the three studies of women who had taken an oral bisphosphonate for an average of about 3 years, serum levels of CTX were reduced more at 1 month and over the course of 12 months in the groups that were randomly assigned to receive denosumab compared to the groups that received either weekly alendronate or risedronate or monthly oral ibandronate . Reductions in serum CTX and P1NP were similar at day 10 of treatment in groups of women previously treated with bisphosphonates who were randomized to receive either denosumab or intravenous zoledronate, but levels remained lower in the denosumab group over the ensuing 12 months of follow-up .
In women with postmenopausal osteoporosis randomly assigned to treatment with teriparatide or denosumab, 12 months of denosumab therapy increased BMD at the lumbar spine, total hip and 1/3 radius by 5.5%, 2.5%, and 1.7%, respectively, compared to increases of 6.2% at the lumbar spine, 0.7% at the total hip and a decrease of 1.7% at the 1/3 radius with teriparatide . Using high-resolution peripheral quantitative computed tomography (HR-pQCT), volumetric BMD (vBMD) in both the cortical and trabecular compartments of the distal tibia and radius increased more with 12 months of denosumab therapy than with teriparatide . Similarly, increases in bone strength and failure load, estimated by finite element analysis (FEA) of the HR-pQCT scans, were greater in the radius and tibia with denosumab compared to teriparatide. After 24 months of therapy, spine BMD was slightly higher in the teriparatide group while total hip BMD was slightly higher in the denosumab group, although neither difference was statistically significant . BMD at the 1/3 radius site had increased by 2.1% with denosumab and had decreased by 1.7% with teriparatide.
Other skeletal effects of denosumab
Denosumab therapy has consistently resulted in improvement in other indices of skeletal mass, structure, and strength. vBMD, assessed by quantitative computed tomography (QCT) measurements of the radius, was acquired in a cohort of 322 postmenopausal women with low BMD treated with denosumab 60 mg every 6 months or placebo for 2 years . Significant increases in vBMD were noted at all measurement sites with denosumab therapy versus placebo, including regions comprised mainly of cortical (mid-radius) or trabecular (ultradistal site) bone. Differences ranged from 2.4% in the mid-radius to 4.7% in the ultradistal region. Cortical thickness appeared to increase, and the polar moment of inertia (PMI), a measure of bone strength, was significantly increased with denosumab therapy.
In a subset of 209 postmenopausal women with osteoporosis in FREEDOM, vBMD in the lumbar spine and proximal femur was measured by QCT . Values in the trabecular compartment of the lumbar spine increased progressively, reaching 12.6% at 3 years while a decrease of 9.2% was observed in the placebo group. BMD by QCT in the total hip region increased by 5.2% with denosumab (primarily in the cortical compartment) and decreased by more than 3% with placebo.
The effects of denosumab therapy on BMD were further explored by QCT and HR-pQCT measurements of distal radius and distal tibia in a separate set of postmenopausal women with low bone mass . Compared to treatment with placebo and with alendronate, vBMD of the distal radius and PMI, measured by QCT, increased significantly with denosumab therapy over 12 months. With HR-pQCT, vBMD in both the radius and tibia was found to increase significantly with both alendronate and denosumab therapy compared to baseline and to the decline that occurred in subjects receiving placebo. The responses were significantly greater with denosumab. The increases in BMD with denosumab and alendronate occurred predominantly in the cortical bone compartment where an apparent increased cortical thickness was again noted with denosumab, likely due to infilling of endocortical surfaces. These observations are similar to those noted with denosumab therapy in monkeys, changes that were accompanied by decreased cortical porosity and substantial improvement of bone strength .
Trabecular bone score (TBS) is an index of the heterogeneity of density distribution on dual energy X-ray absorptiometry (DXA) scans of the lumbar spine that correlates with trabecular microarchitecture. TBS was assessed at baseline and months 12, 24, and 36 in a subset of 285 women (128 placebo, 157 denosumab) in FREEDOM who had TBS values at baseline and at least one postbaseline visit . In the denosumab group, progressive increases from baseline over 36 months were observed for BMD (9.8%) and TBS (2.4%), while values increased minimally or decreased in the placebo group. The changes in TBS were largely unrelated to baseline or change in BMD.
In a small subset of subjects in FREEDOM receiving placebo ( n =51) or denosumab ( n =48), estimates of hip and spine strength were assessed by a voxel-based FEA of computed tomography (CT) scans . After 36 months of therapy, estimated strength increased for the denosumab group compared with the placebo group by 14.3% and 22.4% in the hip and spine, respectively ( P <.0001 for both comparisons). This method of estimating bone strength was validated as an accurate reflection of strength assessed by direct mechanical testing in 52 cynomolgus monkeys randomly assigned to sham surgery or ovariectomy followed by placebo or denosumab therapy for 16 months . Vertebral strength by FEA was highly correlated ( R 2 =0.97) with mechanical testing, independent of treatment. In addition, the effect of denosumab on vertebral strength was similar between the two assessment methods [mechanical testing +57%; 95% CI (26%, 95%)] and FEA [+51% (20%, 88%)]. Analysis of the same set of CT scans from FREEDOM with an alternative smooth finite element methodology, vertebral body strength was increased, by 17.4% from baseline at 36 months . Femoral strength in the fall and stance configurations was increased by 7.2% and 5.2% from baseline, respectively, at 36 months.
In a subset of FREEDOM subjects, including 28 women who received denosumab and 22 from the placebo group, cortical porosity of the femoral shaft was assessed by a proprietary analysis of CT scans, and bone strength was estimated using FEA of those same scans . Cortical porosity in the baseline scans was inversely correlated with estimated bone strength. After 36 months of denosumab treatment, cortical porosity was reduced porosity by 3.6% relative to baseline. The increase of 7.9% from baseline in estimated hip integral strength during the 36 months of treatment was significantly correlated with the reduction in total porosity. In a separate study, cortical porosity of the distal radius was quantified from HR-pQCT scans . Compared to treatment with alendronate, denosumab therapy reduced bone resorption (reflected by changes in serum CTX) more quickly and more completely and significantly reduced cortical porosity. These findings were consistent with the effects of denosumab on bone remodeling and cortical porosity in cynomolgus monkeys .
Using a novel cortical mapping technique of hip CT scans, denosumab was shown to increase femoral cortical mass surface density and thickness as early as 12 months after therapy was begun . Over 3 years, treatment with denosumab increased femoral cortical mass surface density by 5.4% compared to placebo. One-third of the increase came from increasing cortical density, and two-thirds from increasing cortical thickness. Most of the femoral cortex displayed a statistically significant relative difference by 36 months, but this increase was as much as 12% compared to placebo at some locations such as the lateral femoral trochanter.
Bone histology and histomorphometry
Transiliac bone biopsies were obtained, after double tetracycline labeling, at 24 and/or 36 months in FREEDOM (47 women on denosumab, 46 on placebo) and at 12 months in STAND (21 subjects continuing alendronate and 15 switched to denosumab after alendronate) . In FREEDOM, micro-CT analyses of the biopsy specimens did not differ between placebo- and denosumab-treated subjects except for decreased cortical porosity noted with denosumab at 24 (but not at 36) months. Qualitative bone histology in all women treated with denosumab demonstrated normal lamellar bone, normal mineralization, and absence of marrow fibrosis. Quantitative histomorphometry in FREEDOM did not differ between months 24 and 36. Evidence of double tetracycline labeling was observed in trabecular or cortical bone in all biopsies from subjects on placebo. For subjects treated with denosumab, 40% had double tetracycline label, 25% had single label only, and 36% had no detectable label. Serum markers of bone turnover did not differ between subjects with or without tetracycline labels. Because of the infrequency of subjects with double label, an accurate assessment of the effects of denosumab on bone formation was not possible. In the subjects from STAND, measures of bone turnover tended to be lower in subjects who switched to denosumab compared to those who continued alendronate. Double tetracycline labeling was observed in all biopsies of subjects in the alendronate group, while for subjects switched to denosumab, double label, single label, and no label was observed in 60%, 20%, and 20% of biopsies, respectively.
Histomorphometry of cortical bone was assessed in more detail in a set of 112 iliac crest biopsy specimens, including 67 obtained at month 24 (37 placebo, 30 denosumab) and 45 at month 36 (25 placebo, 20 denosumab) in FREEDOM . Compared to the placebo group, endocortical osteoclast surface, eroded surface area, and mean and maximum erosion depth were significantly lower in the denosumab group at months 24 and 36. Cortical thickness, cortical porosity, and endocortical wall thickness did not differ between the two groups. These findings were consistent with the evidence of decreased cortical bone remodeling, reflected by the presence of tetracycline labels in cortical bone in 43% and 50% of biopsies in the denosumab group at months 24 and 36, respectively.
Normal bone histology and evidence of reduced bone resorption were also observed in a set of biopsies obtained at the end of year 2 of the FREEDOM Extension study in 28 subjects who had received denosumab for 5 years and 13 women who had received placebo during FREEDOM and then denosumab for 2 years in the Extension . Iliac crest bone biopsies obtained at the end of the FREEDOM Extension study were available for histology ( n =22) and histomorphometry ( n =21), including 11 women who had had a previous biopsy. After 10 years of denosumab therapy, normal lamellar bone structure and evidence of low remodeling activation frequency, similar to that observed after 2–3 and 5 years of therapy, was observed . Mineralization density was increased, and mineralization heterogeneity was reduced compared to values obtained earlier in the placebo group, but, importantly, these measures had remained stable between years 5 and 10 of therapy.
Use of denosumab with bone-forming agents
Denosumab reduces remodeling-based bone formation in addition to inhibiting bone resorption while teriparatide therapy for postmenopausal osteoporosis activates bone remodeling, increasing both bone formation and resorption . Adding denosumab to teriparatide to reduce resorption while activating bone formation has appeal, especially since the increase in bone resorption affected by teriparatide is mediated by RANKL. This paradigm was evaluated in the DATA Trial in which 94 women with postmenopausal osteoporosis were randomly assigned to receive open-label therapy with approved doses of denosumab, teriparatide, or both drugs for 2 years . After 12 months of therapy, BMD of the lumbar spine and total hip increased more (9.1% and 4.9%, respectively) in the combined therapy group than in either of the monotherapy groups (6.2% and 0.7% with teriparatide; 5.5% and 2.5% with denosumab). The advantage of combined therapy was maintained but did not increase during the second year of therapy. Denosumab prevented the increase in bone resorption induced by teriparatide as evidenced by the reduction in serum CTX being similar in the denosumab-only and the combination groups. The increase in markers of bone formation observed in the teriparatide-only group was also prevented by the combination with denosumab, although the decreases in bone formation markers were somewhat less in the combined group than in the denosumab only group. This raises the question of whether the improvements in trabecular microstructure expected with teriparatide therapy occur in the setting of reduced bone formation when coadministered with denosumab.
Cortical and trabecular bone mass and structure of the distal tibia and radius were assessed in the DATA Trial subjects with HR-pQCT scanning after 12 months of therapy . Increases in total, trabecular, and cortical vBMD as well as improvements in cortical and trabecular microarchitecture and in estimates of bone strength were greater with combination therapy than with teriparatide but not always greater than with denosumab alone. The addition of denosumab to teriparatide prevented the increase in cortical porosity induced by teriparatide alone at the radius and tibia. Increases in total and cortical vBMD and cortical thickness of the tibia were greater by about 50% with combined therapy compared to denosumab alone. In addition, tibial failure load was improved more with combined therapy than with monotherapy. Differences in indices of trabecular mass and structure in the tibia or of either cortical or trabecular status at the radius did not differ between the group receiving denosumab alone or in combination with teriparatide. The study was too small to determine whether these observed differences in BMD and structure connote better protection from fracture.
Information is clearer about the use of denosumab and teriparatide in sequence. Switching to denosumab after 24 months of teriparatide therapy resulted in sizable additional increases in both lumbar spine and proximal femur BMD . In contrast, significant decreases in BMD in the proximal femur and 1/3 radius were observed when subjects who had received denosumab for 24 months were transitioned to teriparatide.
No studies have yet evaluated the combination of denosumab with the antisclerostin inhibitor romosozumab. Following 1 or 2 years of romosozumab therapy with denosumab results in progressive increases in hip and spine BMD . The antifracture efficacy of romosozumab compared to placebo for 12 months is maintained when subjects receiving either romosozumab or placebo were treated with denosumab for an additional 24 months . In a very small group of patients who had received romosozumab for 2 years followed by 12 months of denosumab therapy, romosozumab retreatment for 12 months resulted in a small additional increase in lumbar spine BMD, while values remained stable at the total hip . Markers of both resorption and formation increased to above baseline values within 6 months of discontinuing denosumab and restarting romosozumab.
Safety and tolerability
Denosumab has been well tolerated in clinical trials for osteoporosis, and no major safety problems have been evident to date. The large placebo-controlled FREEDOM study and the long-term FREEDOM Extension study provide the best sources of information about safety. During 3 years of observation in FREEDOM, the overall frequency of adverse events was similar between the denosumab and placebo groups . Ninety subjects (2.3%) in the placebo group and 70 (1.8%) in the denosumab group died ( P =.08).
Skin-related problems and flatulence were the only adverse events to be statistically more common with denosumab treatment than with placebo . Skin rash or eczema occurred in 3% with denosumab treatment and 1.7% placebo ( P <.001). These events were generally mild and self-limited. In addition, skin infections (cellulitis or erysipelas) requiring hospitalization [and consequently defined as serious adverse events (SAEs)] occurred in 12 subjects received denosumab (0.3%) and in but one subject in the placebo group (<0.01%; P =.002). Many of the subjects who developed skin infections had risk factors, including skin wounds and venous stasis or ulceration. Infections responded to standard antimicrobial therapy. In only 1 of the 12 subjects with cellulitis treated with denosumab was a positive microbiological culture obtained ( Streptococcus pyogenes ) . No relationship was observed between the occurrence of skin rash or infection and the site or timing of the denosumab injection or the duration of treatment. There was no difference in incidence of injection site reaction between the denosumab and placebo groups. Skin rash and infection was not observed more frequently in the large clinical development program that evaluated the use of denosumab 120 mg given monthly in patients with cancer-related bone diseases . Pancreatitis (1 placebo, 3 denosumab) and endocarditis (1 placebo, 3 denosumab) occurred rarely but more commonly with denosumab than placebo in FREEDOM . During the FREEDOM Extension study, adverse events and SAEs, including pancreatitis, endocarditis, skin rash and cellulitis, occurred at similar or lower rates than were observed in FREEDOM . In the FREEDOM placebo group that began denosumab therapy at the beginning of the Extension study, pancreatitis, endocarditis, and cellulitis did not occur more frequently when they began denosumab than had occurred while they received placebo .
Adverse skeletal events
The potent inhibition of bone remodeling by denosumab raised concern about potential skeletal harm from long-term therapy due to too much inhibition of bone turnover for too long . These concerns include the risks of osteonecrosis of the jaw (ONJ), atypical femoral fractures and impaired fracture healing .
ONJ occurred with similar frequency with high doses of denosumab and intravenous zoledronic acid in patients with cancer-related bone diseases, patients who seem to be at high risk for this complication . In FREEDOM and its Extension, all oral adverse events were reviewed by a committee of experts, blinded to treatment allocation, to determine if criteria for diagnosing ONJ were met. No cases of ONJ were observed in either the denosumab or placebo groups during the 3 years of FREEDOM. Thirteen positively adjudicated cases of ONJ were identified in the FREEDOM Extension study, resulting in an exposure-adjusted ONJ rate of 5.2 per 10,000 subject-years . Some of these 13 cases had not been diagnosed as ONJ by the study site physicians, being categorized as ONJ only after the adjudication process. Treatment with denosumab before ONJ was diagnosed ranged from 1.1 to 9.6 years, with no clear relationship between duration of treatment and risk of ONJ. Twelve of the 13 cases reported having undergone an invasive oral event such as tooth extraction or dental implant . The incidence of ONJ was higher (0.68%) in women who reported an invasive oral event compared to those without an invasive event (0.05%). Among 212 patients reporting dental implant, one subject (who had two implants) experienced delayed osseointegration of the implant, thereby meeting the criteria for ONJ. This patient continued to receive denosumab; osseointegration ultimately occurred, and she retained the implants. The ONJ lesions resolved in 11 of the 13 cases, including 8 subjects who continued to receive denosumab after the ONJ event. In one case the ONJ lesion was recent and appeared to be healing at the end of the study, and another case was lost to follow-up before ascertainment of resolution was achieved.
No cases of atypical femoral fracture were reported during the FREEDOM study. In the third year of the Extension study, an adjudication process was initiated to review all femoral shaft fractures. Of 9 subtrochanteric or diaphyseal femoral fractures observed during the Extension study, 2 were adjudicated as atypical, resulting in an event rate of 0.8 per 10,000 participant-years . These two events occurred during the third and seventh years of therapy. From this information, no conclusion can be drawn about the relationship between duration of denosumab exposure and risk of atypical fracture. Several case reports of femoral shaft fractures with atypical radiographic features have been reported in patients receiving denosumab; in most cases, patients had received previous bisphosphonate therapy .
No impact of denosumab therapy on fracture healing has been observed. In a prespecified analysis of FREEDOM, two cases of delayed union were observed with denosumab therapy while four cases of delayed union and an additional case of nonunion occurred in patients receiving placebo .
As with other potent antiresorptive agents, a transient small decrease in serum calcium, accompanied by a compensatory rise in serum parathyroid hormone (PTH), occurred upon initiation of denosumab therapy . Hypocalcemia is a potential complication, especially in patients with impaired regulation of calcium homeostasis such as hypoparathyroidism or severe vitamin D deficiency. Subjects with hypocalcemia or vitamin D deficiency could not participate in the clinical trials with denosumab, and symptomatic hypocalcemia was not observed with denosumab therapy in FREEDOM . However, cases of severe hypocalcemia have been reported with the use of denosumab in clinical settings, especially in patients with severely impaired renal function (vide infra), those receiving calcimimetic drugs or with other metabolic bone disorders .
In a retrospective review of records from an Israeli health maintenance organization, 2505 patients (average age 76 years) were identified who received denosumab between 2010 and 2018 . Hypocalcemia, defined as an albumin-adjusted serum calcium concentration <8.5 mg/dL, was observed in 7.6% of patients, and 1% had serum calcium levels below 8.0 mg/dL. Hypocalcemia most often occurred after the initial dose. The two predictors of hypocalcemia were pretreatment levels of serum calcium and creatinine. It is prudent to ensure adequate intakes of calcium and vitamin D and to evaluate serum calcium and renal function before beginning denosumab therapy. Routine testing of serum calcium is not necessary before subsequent doses except in patients at risk for hypocalcemia.
Impaired immune function and infection
RANKL is expressed in dendritic and T-helper cells, but its exact role in modulating immune function is not known . Inhibition of RANKL in rodents does significantly alter inflammatory processes . Patients with genetic deficiency of RANKL do not have deficits in immune function except hypoglobulinemia . No laboratory evidence of immune deficiency has been observed with denosumab in clinical studies. The incidence of infection is not increased with denosumab compared to placebo or alendronate control groups . However, the frequency of infection or inflammatory disorders requiring hospitalization (thus an SAE) was numerically higher with denosumab therapy in some studies. In FREEDOM, SAEs related to infection occurred in 3.2% and 3.7% of patients given placebo or denosumab, respectively ( P =.14) . In a thorough analysis of these data, Watts et al. observed that (1) many of these SAEs were not infections but were inflammatory conditions such as diverticulitis or appendicitis that are grouped with infections in analyses of clinical trials; (2) the SAEs did not occur in a pattern suggesting a relationship to the time of injection or duration of therapy; and (3) there was no evidence of increased risk of opportunistic infections . An increasing risk of infection was not observed with long-term therapy in either treatment group in the FREEDOM Extension study . In a cohort study using claims data from a US commercial insurance plan database, the frequency of hospitalizations for infection was compared in patients treated for osteoporosis with denosumab or zoledronate . Denosumab was not associated with an increased risk of serious infection [hazard ratio (HR) 0.81; 95% CI, 0.55–1.21]. In a similar study in patients with rheumatoid arthritis receiving biological therapies who also were treated with either denosumab or zoledronate for osteoporosis, no increased risk of infections requiring hospitalization was observed with denosumab .
In FREEDOM, neoplasms occurred more frequently in patients receiving denosumab (4.8%) than in those who took placebo (4.1%), but this difference was not statistically significant ( P =.31) . No specific type of tumor accounted for this apparent difference in frequency, and immune-related tumors were not more common with denosumab. During the Extension study, rates of neoplasms remained stable and did not appear to increase with longer treatment .
As with all biological therapies, there is a risk of hypersensitivity with denosumab administration. Reactions, including throat tightness, facial and upper airway edema, pruritus, and urticaria, have been reported . Individual cases of hepatotoxicity, nephritis, skin reactions, and antineutrophil cytoplasmic antibody–associated vasculitis have been reported .
The RANKL/OPG pathway may play a role in the regulation of vascular calcification. OPG knockout mice develop extensive vascular calcifications . In human RANKL knocked-in mice with glucocorticoid-induced osteoporosis, RANKL inhibition decreased vascular calcifications . In FREEDOM, 12–13% of participants had a prior history of cardiovascular disease . The incidence of adverse events or SAEs related to cardiovascular disease did not differ between the denosumab and placebo groups . In a post hoc analysis in a subset of more than 1600 women in FREEDOM selected because of risk factors for cardiovascular disease, denosumab treatment had no effect on the progression of calcification of the abdominal aorta or the incidence of adverse cardiovascular events compared to placebo . In a large population-based cohort study, patients beginning denosumab and zoledronate had comparable risks of cardiovascular disease during the first year of therapy .
Falls, excluding those occurring on the same day as a fracture, occurred less commonly with denosumab (4.5%) versus placebo (5.7%) in FREEDOM ( p =.02) . It has subsequently been shown that RANKL inhibition improves muscle strength in a mouse model of Duchenne’s muscular dystrophy and in mice with sarcopenia . In women with osteoporosis, denosumab therapy, compared to no treatment, improved appendicular lean mass and handgrip strength while no improvement was observed with bisphosphonates . In a pooled metaanalysis of denosumab clinical trials, the risk of falls was lower with denosumab therapy than with placebo [HR (95% CI): 0.79 (0.66, 0.93); P =.0061] .
Withdrawal of therapy
There is no evidence in large clinical trials of loss of effectiveness or increasing safety concerns with long-term therapy with denosumab and, thus, no reason for a limit to the duration of therapy. Furthermore, consistent with the pharmacology of denosumab, its effects on bone remodeling dissipate quickly as the antibody is cleared from the circulation. This was observed in the Phase II study in subjects who received 210 mg denosumab every 6 months or 30 mg every 3 months for 2 years . Upon stopping treatment, markers of bone turnover returned to baseline within 3 months (9 months after the last dose), transiently rose to values above baseline, and then, in subjects who had taken 210 mg every 6 months, returned to baseline values within 2 years. All the gain in BMD that had occurred with 2 years of treatment (about 6%–8% in the lumbar spine, 4%–5% at the total hip) was lost within the first year after stopping treatment. Very similar responses were observed in postmenopausal women with low bone mass who had taken denosumab 60 mg every 6 months for 2 years and who were then followed for an additional 2 years after stopping treatment . After 8 years of denosumab therapy and BMD gains of 16.8% at the lumbar spine and 6.2% at the total hip, stopping therapy resulted in losses of 6.7% and 6.6% at the lumbar spine and total hip, respectively, over 12 months in women who did not receive follow-on therapy . These results suggest that the rate of bone loss may not be affected by the duration of therapy. The rate of bone loss or rise in serum CTX appeared to be blunted in the small number of women who took another osteoporosis drug for at least part of the year after stopping denosumab . Hypercalcemia has also been reported during the interval of high bone remodeling after denosumab discontinuation . One year after stopping treatment, denosumab was restarted in subjects who had taken 30 mg every 3 months for 2 years in the Phase II study . Rapid reduction in markers of bone turnover was noted and regain of BMD occurred within 1 year.
The complete reversibility of the effects of denosumab on remodeling activity was also demonstrated by quantitative histomorphometry. Transiliac bone biopsies were obtained in 15 subjects who had taken denosumab for 12 months and who had then discontinued treatment for 12–36 months . All biopsies demonstrated normal histology and remodeling parameters. Tetracycline labeling was noted in all subjects (double label in 14, single label in 1).
Appreciating that the rebound in bone remodeling and rapid bone loss following discontinuation denosumab on fracture risk could be of clinical importance, fracture rates were assessed in subjects in FREEDOM who discontinued treatment after receiving two to five doses of denosumab (327 subjects) or placebo (470 subjects) . Osteoporosis-related fractures occurring between 7 months after the last treatment dose and the end of observation (median time 0.8 years) were observed in 9% of subjects treated with placebo and in 7% who took denosumab. The HR of experiencing a fracture after discontinuing denosumab therapy, adjusted for age and total hip BMD at baseline, was 0.82 (95% CI 0.49–1.38).
Subsequently, several case reports and series were published of patients who experienced multiple and/or severe vertebral fractures within a few months of discontinuing denosumab therapy . This led to an additional specific assessment of the risk of new or worsening clinical or morphometric vertebral fractures, including the risk of multiple vertebral fractures, after discontinuation of denosumab in the FREEDOM and its Extension study . Participants who had received two to five doses of study drug and were followed for at least 7 months after the last dose of study drug were included in the analysis. The median follow-up time after discontinuing therapy was 0.5 years [interquartile range (IR) 0.3–1.4] in 470 patients who withdrew from placebo, 0.5 years (IR 0.2–1.4) in 327 participants who discontinued denosumab during FREEDOM and 0.2 years (IR 0.1–0.7) in 678 women who stopped denosumab during the FREEDOM Extension study. Vertebral fracture rates were 7.0 and 8.5 per 100 participant-years in the subjects receiving or having discontinued placebo, respectively. While receiving denosumab, vertebral fracture rate was 1.2% and increased to 7.1% during follow-up after denosumab discontinuation. The rate of multiple vertebral fractures increased from 0.4 per 100 participant-years while receiving denosumab to 4.2 after discontinuation, compared to 3.2 in the group that discontinued placebo. Age, BMI, baseline level of bone turnover (assessed by serum CTX), duration of treatment, and BMD at either baseline or at the beginning of the off-treatment interval did not differ between those who did or did not experience a posttreatment fracture, whereas prior vertebral fracture was a strong predictor of having multiple vertebral fractures upon discontinuing treatment. The value of these results is limited by the post hoc observational nature of the data collection; confounding of results by the use of other osteoporosis therapies in 14.5% and 42.8% of patients who stopped denosumab or placebo, respectively; and the short follow-up interval. That the latter is important is suggested by the observation that the median follow-up time was longer in participants who experienced vertebral fractures than in those who did not fracture, suggesting that these data may underestimate the risk of vertebral fractures after stopping denosumab therapy.
The resolution of the effects of denosumab (transient rebound in bone turnover, loss of BMD back to pretreatment levels, and rapid loss of fracture protection) is very reminiscent of that observed upon stopping estrogen therapy . In a 3-year observation of study participants after the Women’s Health Initiative (WHI) studies were halted, women who had stopped estrogen treatment, which had significantly reduced fracture risk, had vertebral and hip fracture rates that were similar to, but did not exceed, the rates in the group that had received placebo . Multiple vertebral fractures were not reported . Perhaps the younger women in WHI, whose average BMD was normal, had better trabecular architecture and strength compared to the older women with osteoporosis in FREEDOM who discontinued denosumab.
The optimal strategy to prevent the rebound in remodeling and its consequences upon stopping denosumab is not known. It seems reasonable that another potent antiremodeling agent such as a bisphosphonate would be effective. Alendronate therapy after estrogen, PTH receptor agonists and romosozumab, prevents the expected bone loss . Studies of the effectiveness of bisphosphonate therapy to prevent bone loss upon stopping denosumab have had variable results. In a study in which 106 postmenopausal women with osteoporosis received denosumab for 12 months, follow-on treatment with alendronate maintained spine and hip BMD for 12 months . In a study of 6 women who had been treated with denosumab for 7 years and who then received a dose of intravenous zoledronate begun 6 months after the last dose of denosumab, neither bone loss nor increase in serum P1NP was prevented . In a more recent study a single dose of zoledronate prevented bone loss and probably blunted the rebound in bone turnover for 2 years in 27 women who had received denosumab for an average of 2.2 years . Whether the difference between the two studies with zoledronate is due to small sample sizes or to the differences in duration of exposure to denosumab awaits further study. It may well be that potent bisphosphonates, begun 6 months after the last dose of denosumab, effectively prevent the rebound in turnover and rapid bone loss in most but not all patients. The latter group of patients would then require an additional dose of zoledronate 3–6 months later (9–12 months after stopping denosumab). Strategies to identify the patients who require this additional therapy, based on clinical risk factors or monitoring of turnover markers or BMD, are being evaluated.
Adherence to denosumab therapy
The infrequent dosing schedule of denosumab, coupled with therapy administered by a health-care provider, should provide better compliance with and persistence compared to oral osteoporosis therapies. In a blinded, controlled clinical trial specifically designed to evaluate adherence and patient satisfaction, adherence in the first 12 months of therapy was 87.3% and 76.6% to denosumab and alendronate, respectively, in women with osteoporosis randomly assigned to those therapies . While still blinded to treatment allocation, therapy was switched to the other drug during the second year of the study. Adherence was even more in favor of denosumab at the end of the second year (92.5% vs 63.5% with alendronate). Without knowing the sequence of their therapies, more than 92% of participants favored the denosumab subcutaneous injection dosing regimen over weekly oral therapy.
The real-world experience with persistence to denosumab therapy is less impressive. Using the national drug registry database, 2315 Swedish women were identified who began denosumab therapy between 2010 and 2012 . Persistence with therapy was 83% and 62% at 12 and 24 months, respectively. Among 945 women enrolled in a prospective study in the United States and Canada between 2011 and 2012, persistence with denosumab therapy was only 58% . Adherence for 24 months ranged between 75% and 86% in four European countries. In the Canadian province of Ontario, more than 46,000 patients who began denosumab therapy between early 2012 and March 2015 were followed for up to 4 years. Persistence rates at 2, 3, and 4 years were 59%, 48%, and 38%, respectively. Reasons for treatment cessation were not detailed, and evaluations in each of these studies were conducted before the clear recognition of the rapid loss of vertebral fracture protection upon stopping denosumab therapy. While these reported persistence rates may be higher than those reported with bisphosphonates, they are certainly not ideal.
Denosumab in other populations
Bone loss and increased fracture risk occur in men with aging, in adults receiving glucocorticoid therapy, men and women receiving hormone deprivation therapy for the treatment of prostate or breast cancer, and in patients with rheumatoid arthritis. Denosumab has been evaluated in each of these patient groups. The changes in markers of bone remodeling and in BMD in these groups were similar to those observed in the much larger studies in women with postmenopausal osteoporosis. The frequency of adverse events was generally similar between the denosumab and placebo groups, and no major additional safety concerns were noted.
Men with low bone mass
In a 12-month study of 242 men with low BMD, denosumab 60 mg every 6 months increased BMD by 5.7% at the lumbar spine in 2.4% at the total hip while the changes with placebo were 0.9% and 0.3%, respectively . The responses were similar in men with or without low baseline serum testosterone levels. The average serum CTX level was 60% below baseline after 12 months. The BMD and serum CTX changes were similar to those reported in postmenopausal women with osteoporosis.
The RANKL/OPG ratio is increased by glucocorticoids in vitro, suggesting that inhibiting RANKL with denosumab would have clinical benefit. Patients beginning glucocorticoid therapy or having taken glucocorticoids for more than 3 months were randomly assigned to receive denosumab 60 mg Q6 months or risedronate 5 mg po daily for 24 months . BMD responses and inhibition of markers of both bone resorption and formation were greater at 12 and 24 months with denosumab compared to risedronate. No difference in adverse events, including infections, was observed between the treatment groups.
Impaired renal function
The PKs of denosumab are not altered in patients with severe renal function . Post hoc analyses in the FREEDOM study assessed the effect of denosumab therapy on vertebral fracture risk and BMD in subjects stratified by baseline renal function . This included 2814 subjects with estimated glomerular filtration rate (GFR) of 30–59 mL/min and 73 subjects with estimated GFR of 15–29 mL/min. No difference in the effect of denosumab on the risk of vertebral fracture was observed across the range of renal function. Experience with denosumab in patients with end-stage renal disease on dialysis is very limited, consisting primarily of case reports of patients who experienced severe, often prolonged, hypocalcemia following therapy A metaanalysis of six observational studies with a total of 84 end-stage renal disease patients treated with denosumab demonstrated increased BMD but an estimated incidence of hypocalcemia of 42% . Hypocalcemia occurred most commonly during the first 1–3 weeks of therapy with the nadir of serum calcium levels occurring between 2 weeks and 2 months. While inadequate vitamin D replacement therapy appears to be a risk factor for hypocalcemia in these patients, hypocalcemia can occur in patients receiving active vitamin D metabolites. Whether denosumab increases the risk of the adynamic form of renal osteodystrophy is not known.
Hormone deprivation therapy
A 2-year study with a 1-year extension was performed in 1468 men receiving androgen deprivation therapy for nonmetastatic hormone-sensitive prostate cancer . Men aged 70 years and older were included irrespective of baseline BMD. Men younger than 70 years of age had BMD T -score values of −1.0 in the lumbar spine, total hip, or femoral neck or had a history of osteoporotic fracture. Mean age was 75 years; average T -score values in the lumbar spine and total hip regions were −0.4 and −0.9, respectively. After 24 months of treatment with denosumab 60 mg every 6 months, BMD at the lumbar spine increased by 6.7% and in a total hip region 4.8%. The BMD response was greater in men with higher baseline serum levels of CTX or TRAP 5b, markers of bone resorption and osteoclast number, respectively. As early as 12 months after beginning treatment, a significant effect of denosumab therapy on the risk of vertebral fracture was observed. The incidence of vertebral fracture was 3.9% placebo group and 1.5% with denosumab at 36 months (relative risk 0.38, P =.006). Fewer fractures at any site occurred with denosumab, but the 28% reduction was not statistically significant ( P =.10). Significant differences in nonvertebral fracture risk were not observed. Cataracts occurred more commonly in men receiving denosumab (4.7%) versus placebo (1.2%).
Women receiving aromatase inhibitor therapy for the treatment of nonmetastatic estrogen-receptor-positive breast cancer were treated with denosumab ( n =127) or placebo ( n =121) for 24 months . BMD increased by 7.6% in the lumbar spine and by 4.7% at the total hip compared to placebo. The frequency of fracture was too small in this small study to make meaningful comparisons between treatment groups. In a much larger study in Europe, 3420 postmenopausal women with nonmetastatic breast cancer receiving aromatase inhibitor therapy were randomly assigned to receive denosumab 60 mg or placebo every 6 months for up to 6 years . Median time on study was 38 months. Compared to placebo, denosumab therapy significantly delayed the time to first clinical fracture [HR 0.50 (95% CI 0.39–0.65), P <.0001], including a 47% reduction in radiographic vertebral fractures at 36 months. Fracture risk was reduced in all subgroups stratified by baseline BMD values, including women with normal (>−1.0) baseline BMD T -score values. Women receiving denosumab while on aromatase inhibitor therapy may also be at risk for vertebral fractures if denosumab is discontinued without follow-on bisphosphonate therapy .
RANKL has been implicated in both the skeletal and synovial pathology of rheumatoid arthritis. Patients with this disorder (218, about 1/3 men) were randomly assigned to receive placebo, denosumab 60 mg every 6 months or denosumab 180 mg every 6 months . All subjects were taking methotrexate, and about one-third were taking glucocorticoids. Very few patients received anti-TNF treatment during the study. After 12 months of denosumab the expected changes in BMD and markers of bone turnover were evident and were not affected by glucocorticoid therapy. The increase in magnetic resonance imaging (MRI) erosion score was significantly lower with denosumab 180 mg every 6 months than with placebo . The difference in MRI erosion score between placebo and denosumab 60 mg every 6 months was almost as large as with the 180 mg dose but was not statistically significant. A significant difference in the modified Sharp erosion score was observed at 6 months with the 180 mg dose and at 12 months with both doses. No differences were noted in joint space narrowing or measures of disease activity. No unexpected adverse events were reported. The frequency of infections, including those requiring hospitalization, was similar among treatment groups.
Denosumab is an effective treatment for osteoporosis in many clinical situations. Its mechanism of action is unique and elegant. Increments in treatment-induced BMD are at least as great with denosumab as with bisphosphonates, and no antiremodeling drug has been shown to be more effective in reducing fracture risk in patients with osteoporosis. BMD in the spine and hip increases progressively with treatment over 10 years, and nonvertebral fracture risk appears to decline with treatment beyond 3 years, suggesting that the fracture protecting effectiveness of denosumab improves with long-term treatment. These effects differ from the plateau of hip BMD and the persistent but not progressive decrease in fracture risk with long-term bisphosphonate therapy . There is no evidence that resistance to the skeletal effects of denosumab occurs due to counter-regulatory or immunologic mechanisms.
In my opinion, denosumab is appropriate for use as a first-line treatment of postmenopausal women and men with osteoporosis and in patients who remain at high risk of fracture after 5 years of bisphosphonate therapy. Recent studies have shown that therapy with romosozumab or teriparatide is more effective than a bisphosphonate and likely more effective than denosumab for patients at very high risk of fracture . Following either of these bone-forming drugs with denosumab results in additional gains in BMD and, in the case of romosozumab, persistence of the fracture protection benefits of the bone-forming agent . Parental dosing every 6 months is convenient and allows for sharing of the responsibility for treatment persistence between the patient and physician. Denosumab dosing also precludes concern about gastrointestinal intolerance or inadequate absorption and bioavailability due to intestinal pathology or poor compliance with dosing instructions. As a result, denosumab is an appealing option for the management of the many patients with real or perceived intolerance to oral bisphosphonates. Its effectiveness and lack of nephrotoxicity make it an attractive option in patients with moderately impaired renal function.
Since osteoporosis is a chronic condition not cured by any of our treatments, long-term if not lifelong management is required. Denosumab is, in my opinion, the best current option for long-term therapy, based on its effects on hip BMD, the decline in nonvertebral fracture rate and the clean safety profile in the FREEDOM Extension study. There is no limit on the duration of denosumab therapy and certainly no justification for a “drug holiday.” If therapy is discontinued, for whatever reason, transition to a bisphosphonate should be considered, particularly in patients with previous vertebral fractures and other risk factors for vertebral fracture. The rebound in bone remodeling, rapid loss of bone density, and the loss of vertebral fracture protection should not be viewed as a reason to not begin denosumab therapy. Rather, for every patient beginning osteoporosis therapy, especially those beginning denosumab, there needs to be a clear plan for sequential therapy. The rapid off effect of denosumab does make it a less attractive option than a long-acting bisphosphonate for the treatment of patients with low bone mass who take glucocorticoids or receive hormone deprivation therapy in whom treatment with the bone-offending drug may be stopped. Similarly, I would favor a long-acting bisphosphonate over denosumab in patients for whom short-term therapy is considered to prevent a transient interval of rapid bone loss in women in early menopause or stopping estrogen or patients with acute immobilization.
The excellent safety profile of denosumab over 10 years was demonstrated in the FREEDOM Extension study . There was no evidence of significant off-target effects, including impaired immune function or risks of infection, malignancy or cardiovascular events. The safety of denosumab in patients with immune dysfunction, including those taking biologic inhibitors of TNFα and other immune modulators, has not yet been adequately explored.
Despite these many attributes, clinically important limitations to and concerns about the use of denosumab exist. As a very rapidly acting inhibitor of bone resorption, the potential for inducing symptomatic hypocalcemia is greater than with oral bisphosphonates or raloxifene. Denosumab is contraindicated in patients with hypocalcemia, and therapy should not be started until the patient’s intake of calcium and vitamin D is adequate. The risk of hypocalcemia is especially high in patients with end-stage renal disease. While denosumab is not contraindicated in these patients, the effectiveness and safety of denosumab in patients with severely impaired renal function and renal osteodystrophy is uncertain. Denosumab should be used with great caution in these patients.
The major concern about long-term denosumab therapy is its effect on skeletal safety. Bone turnover markers and bone biopsy studies confirm that denosumab is a more potent inhibitor of bone resorption than is alendronate. Bone remodeling is reduced to very low levels, at least temporarily, in all patients. There is a gradual offset of the effect on bone resorption before the next dose in most patients. The clinical relevance and long-term safety of such a low level of remodeling is unknown. That patients in osteoporosis clinical trials experienced ONJ or atypical fractures raises the possibility of these untoward skeletal problems with long-term denosumab therapy. However, so few cases occurred that a relationship between denosumab treatment or the duration of treatment and the risk of ONJ or atypical fracture cannot be discerned.
Denosumab occupies a very important role in the long-term management of men and women with osteoporosis. As first-line therapy, it quickly and persistently reduces the risk of important fractures. As follow-on therapy after other drugs, it results in additional BMD gains, and it appears to be the safest and most effective drug for long-term use. Denosumab will be an important component of our emerging strategies for the sequential use of therapeutic agents to manage the risk of fracture over a patient’s remaining lifetime.