- Email: [email protected]
Clinical Efficacy and Safety of Denosumab in Postmenopausal Women with Low Bone Mineral Density and Osteoporosis: A Meta-Analysis
OSTEOPOROSIS
Clinical Efficacy and Safety of Denosumab in Postmenopausal Women with Low Bone Mineral Density and Osteoporosis: A Meta-Analysis Camilla von Keyserlingk, MA,* Robert Hopkins, MSc,*,† Athanasios Anastasilakis, MD,‡ Konstantinos Toulis, MD,§ Ron Goeree, MA,*,† Jean-Eric Tarride, PhD,*,† and Feng Xie, PhD*,†
Objective: Clinical trials indicate that denosumab could be a potential treatment for postmenopausal osteoporosis. The objective of this meta-analysis was to assess the clinical efficacy and safety of offering denosumab to postmenopausal women with low bone mass. Methods: Data sources included MEDLINE, EMBASE, and Cochrane Central Register of Controlled Trials (CENTRAL) from inception to February 3, 2010 and bibliographies of reviews. Randomized controlled trials comparing the efficacy and safety of denosumab to placebo for treatment of low bone mass (low bone mineral density or osteoporosis) in postmenopausal women were selected. Two reviewers independently abstracted data on study general characteristics and outcomes. Review Manager 5.0 software was used for data syntheses and meta-analysis. Results: The database search revealed 4 studies (comprising 8864 patients randomized) that met the inclusion criteria and contributed to some or all of the meta-analysis outcomes. Relative risk (95% CI) of fractures for the denosumab compared with placebo group was 0.58 (0.52 to 0.66); relative risk (95% CI) of serious adverse events was 1.33 (0.83 to 2.14); relative risk (95% CI) of serious adverse events related to infection was 2.10 (0.64 to 6.90); relative risk (95% CI) of neoplasm was 1.11 (0.91 to 1.36); relative risk (95% CI) of study discontinuation due to adverse events was 1.10 (0.83 to 1.47); and relative risk (95% CI) of death was 0.78 (0.57 to 1.06). Findings remained robust to sensitivity analyses. Conclusion: Our analysis found a significant reduction in relative fracture risk in the denosumab compared with the placebo group. © 2011 Elsevier Inc. All rights reserved. Semin Arthritis Rheum 41:178-186 Keywords: osteoporosis, low bone mineral density, bone mass, denosumab, postmenopausal, metaanalysis
O
steoporosis is a systemic skeletal disease characterized by low bone mass and microarchitectural deterioration of bone tissue, with a consequent increase in bone fragility and fracture risk (1). It is esti-
*Programs for Assessment of Technology in Health (PATH) Research Institute, St. Joseph’s Healthcare, Hamilton, Ontario, Canada. †Department of Clinical Epidemiology and Biostatistics, Faculty of Health Sciences, McMaster University, Hamilton, Ontario, Canada. ‡Department of Endocrinology, 424 Military Hospital, Thessaloniki, Greece. §Department of Endocrinology, Aristotle University, Thessaloniki, Greece. The authors have no conflicts of interest to disclose. Address reprint requests to Feng Xie, PhD, PATH Research Institute, 25 Main Street West, Suite 2000, Hamilton, Ontario, Canada L8P 1H1. E-mail: barboxie@ gmail.com.
178
mated that 75 million people in the United States, Europe, and Japan are affected by this disease (2). The incidence of fractures increases exponentially with age, resulting in high morbidity, mortality, and health care expenditure (3). For white women, the lifetime risk of sustaining a hip fracture is 1 in 6, compared with a 1-in-9 risk of developing breast cancer (4). Currently standard therapies for osteoporosis focus mainly on antiresorptive agents, with bisphosphonates as a first line of treatment (5). Antiresorptive therapy has proven successful in reducing bone loss and preventing fractures in postmenopausal women over multiple years (6). Concerns regarding potential fracture risk associated with long-term use of bisphosphonates (7-9), low adher-
0049-0172/11/$-see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.semarthrit.2011.03.005
C. von Keyserlingk et al.
ence rates with currently available agents (10), and advances in bone biology spur the development of new therapeutic agents (6). Recent discoveries in pathophysiology of bone metabolism have identified the nuclear factor kappa B ligand (RANKL) as a critical factor in osteoclastdriven bone resorption, and subsequent bone loss (11). Denosumab is a fully human monoclonal IgG2 antibody that specifically targets RANKL, thereby inhibiting osteoclast differentiation, activation, and survival and subsequently suppressing bone resorption (11). Denosumab is administered subcutaneously and could lead to better adherence rates than observed with oral bisphosphonates (12). Denosumab has been found to be effective in decreasing bone turnover markers and stabilizing or increasing bone mineral density (BMD). In the case of therapies for osteoporosis, it has been difficult to evaluate fracture outcomes through comparative randomized controlled trials (RCTs) due to the large sample sizes required for meaningful comparisons. RCTs with relatively small sample sizes (⬍150) recorded the incidence of fractures without conclusive results (13-15). To our best knowledge, the first large phase 3 RCT (16), powered to evaluate comparative fracture risk, was recently published and found a 42% risk reduction in relative fracture risk in the denosumab compared with placebo group at a dose of 60 mg subcutaneously every 6 months (16). While evidence on denosumab seems promising, some concerns have been raised. Since RANK activation by RANKL is also essential for T-cell growth and dendritic-cell function (17,18), inhibition of its action could simultaneously affect the immune system, leading to susceptibility of infections or malignancies. We conducted a systematic review and meta-analysis of RCTs to provide evidence-based knowledge on efficacy and safety of offering denosumab to postmenopausal women with low bone mass. METHODS Data Sources and Searches We performed a literature search for the purpose of identifying RCTs. To identify eligible studies, the main search was conducted in the electronic databases MEDLINE, EMBASE, and Cochrane Central Register of Controlled Trials (CENTRAL) from inception to February 3, 2010. The search strategy with controlled vocabulary, registry number of the drug, and keywords focused on the concepts of “denosumab” and “osteoporosis.” These concepts were combined with methodological filters (supplied by an information specialist) to limit retrieval to RCTs (Appendix 1). No language restrictions were applied. Study Selection Eligible studies for the meta-analysis were RCTs, which compared the efficacy and safety/tolerability profile of denosumab to placebo for the treatment of low bone min-
179
eral density or osteoporosis in postmenopausal women. Studies on individuals with low bone mineral density were included as it is well documented that more than 50% of fractures occur in women with low bone mineral density (19). Trials were excluded, if uncontrolled, and if they included men and/or subjects who had conditions that influenced bone metabolism (eg, metastatic breast cancer). Reviews, case series, abstracts, letters to the editor, phase 1, or observational studies were not eligible. However, reference sections of reviews were screened to identify possible relevant studies. Two reviewers (FX and CvK) independently screened the titles and abstracts of records retrieved from the searches for relevance using predefined checklists. Kappa statistics for interrater agreement were calculated. Discrepancies between reviewers were discussed until a consensus was reached. The full text of any relevant study passing title/abstract screening was retrieved and assessed for inclusion. Two reviewers (FX and CvK), using explicit predetermined criteria, made inclusion and exclusion decisions independently. These were checked for agreement, and discrepancies between reviewers were discussed until consensus was reached. Data Extraction Relevant information from the studies was extracted independently by 2 reviewers (FX and CvK) into predefined data extraction forms. Study characteristics and outcomes for both intervention and control groups were recorded and double-checked. Study quality was assessed independently by 2 reviewers (FX and CvK) using the assessment scale developed by Jadad and coworkers (20). The Jadad scale contains 3 items related directly to the reduction of bias. The items are presented in the form of 5 questions eliciting yes or no answers. Points awarded depended on the quality of the description of the methods to generate the sequence of randomization and of double blinding and the reporting of withdrawals and dropouts (20). Any discrepancies in data extraction were discussed until consensus was reached. Data Synthesis and Analysis Data were entered into the computer program Review Manager (Version 5.0. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2008). Efficacy was evaluated on the basis of the overall incidence of fractures in the treatment and placebo groups, involving a combination of clinical and radiological fractures reported in RCTs. The Mantel-Haenszel method was used to calculate risk ratios (RRs) and their 95% confidence intervals (CI) using either fixed or random effects model, depending on the amount of heterogeneity observed. The safety/tolerability profile was evaluated on the basis of the reported incidence of serious adverse events (SAEs), SAEs related to infection, neoplasms, and the percentage of study discontinuations due to adverse events. Medical conditions that required hospitalization were recorded as
180
Denosumab for women with low bone mass
SAEs. Adverse events (AEs) analyzed were chosen on the basis of biological plausibility. AE outcomes were examined according to the guidelines issued by the CONSORT group (21), stating that “intention-to-treat” analysis should not be used for examining adverse effects. Instead we examined AE outcomes on the basis of an “on-treatment” analysis whereby total patients completing the trial plus dropouts due to adverse events were included. Mantel-Haenszel method was used to calculate RRs and their 95% CI using either fixed or random effects model, whichever was appropriate. Mortality was expressed as RRs with 95% CI using Mantel-Haenszel method and fixed effects model. Heterogeneity between the results of different studies was assessed with the 2 test for significance (a cutoff of 10% for significance was used) and presented by I2 test (22), which can be interpreted as the percentage of total variation across several studies due
to heterogeneity (23). Due to the small number of studies, funnel plot analysis to check for publication bias was not performed. Sensitivity Analyses To check for possible biases introduced by heterogeneity, 2 one-way sensitivity analyses were conducted. First, there was variability in BMD T-scores across studies at patient enrollment (T-score ⱕ ⫺2.5 and ⱖ ⫺4.0 in Cummings and coworkers (16); T-score ⱕ ⫺1.0 and ⱖ ⫺2.5 in Bone and coworkers (13) and Ellis and coworkers (14); T-score ⱕ ⫺1.8 and ⱖ ⫺4.0 in Lewiecki and coworkers (15)). When we compared the actual mean T-scores across studies, it seemed that T-scores were least severe in Ellis and coworkers (14) In addition, the patient population in the study by Ellis and coworkers (14) referred to
Figure 1 QUORUM flowchart.
C. von Keyserlingk et al.
181
Table 1 Main Characteristics of the 4 Randomized Controlled Trials Included in the Meta-Analysis Author, year, reference Lewiecki 2007 [15]
Ellis 2008 [14]
Bone 2008 [13]
Design Randomized, plc-controlled, phase II dose-ranging study (8 doubleblind ⫹ 1 open-label treatment groups) Multicenter, randomized, doubleblind, plc-controlled, phase III study
Multicenter, randomized, doubleblind, plc-controlled, phase III study
Cummings 2009 [17]
International, multicenter, randomized, double-blind, plccontrolled, phase III study Secondary: Nonvertebral, 3/5 and hip fractures, BMD, BTM, safety
Total Study Size (plc) [60 mg/6 mo] 412 (46) [47]
252 (125) [127]
332 (166) [166]
7868 (3935) [3933]
Inclusion Criteria Postmenopausal women (mean age 62.5) with low bone mineral density or osteoporosis (T-score ⱕ ⫺1.8 and ⱖ ⫺4.0) Women (mean age 59.5; 73% postmenopausal [50% with more than 10 years and 23% with 5 to 10 years since their last menstrual period]) with nonmetastatic breast cancer under aromatase inhibitors and low bone mass (T-score ⱕ ⫺1.0 and ⱖ ⫺2.5) Postmenopausal women (mean age 59.4) with low bone mineral density or osteoporosis (T-score ⱕ ⫺1.0 and ⱖ ⫺2.5) Postmenopausal women (mean age 72.3) with osteoporosis (T-score ⱕ ⫺2.5 and ⱖ ⫺4.0)
plc, placebo; BMD, bone mineral density (measured by DXA); BTM, bone turnover marker; LS, lumbar spine; QCT, quantitative computed tomography.
women with nonmetastatic breast cancer on treatment with aromatase inhibitors, representing a population with a potentially different biological behavior than that of the target population (postmenopausal healthy women). A sensitivity analysis was conducted by excluding the study by Ellis and coworkers (14) from all outcomes. Second, there was variability in study duration. Bone and coworkers (13), Ellis and coworkers (14), and Lewiecki and coworkers (15) used 2-year interval data, while Cummings and coworkers (16) used 3-year data. To verify that results were similar during each year of the trial, as reported by Cummings and coworkers (16), we conducted a sensitivity analysis by substituting 2-year for 3-year data on fracture rates from Cummings and coworkers (16).
by combining different dosages of denosumab in this study. Results of the other included studies are based on a 60 mg/6 months dosage. The RCT by Ellis and coworkers (14) was not included in pooling the study discontinuation rate due to adverse event, as the rate was not reported by this study. All studies scored either 4 or 5 of 5 possible points on the Jadad scale, except the RCT by Cummings and coworkers (16), which scored only 3 of 5 points, as it seemed that the process of randomization was not appropriately described and reasons for withdrawals were not reported. Table 1 summarizes the main characteristics and the quality score for the Jadad scale of the 4 studies included in the present meta-analysis.
RESULTS
Meta-Analysis
Literature Review
Efficacy
The search identified 148 potentially relevant citations. A QUORUM (24) flowchart summarizing search results is shown in Figure 1. After screening titles and abstracts, 137 citations were excluded. Thus a total of 11 citations were reviewed in full text. According to our eligibility criteria, 4 studies (13-16) were included in the meta-analysis, contributing to some or all of the outcomes of interest (Fig. 1). Notably, the RCT by Lewiecki and coworkers (15) was excluded when pooling the outcomes of fracture risk, SAEs risk, SAEs related to infection risk, neoplasm risk, and mortality risk as these outcomes were reported
The risk ratio (95% CI) of fractures for postmenopausal women with low bone mass treated with denosumab compared with placebo was 0.58 (0.52 to 0.66) (Fig. 2). Safety The risk ratio (95% CI) of SAEs for postmenopausal women treated with denosumab compared with placebo was 1.33 (0.83 to 2.14), which was not statistically significant (Fig. 3A). Heterogeneity was noted for this outcome. The risk ratio (95% CI) of SAEs related to infec-
182
Denosumab for women with low bone mass
Table 1 Continued Intervention
Endpoint
Quality Score
Denosumab 6, 14, or 30 mg/3 mo or 14, 60, 100, 210 mg/6 mo, or open-label alendronate 70 mg/wk for 24 mo
Primary: LS BMD at 24 mo Secondary: Hip, distal one-third radius, total body BMD, BTM
4/5
Denosumab 60 mg/6 mo for 24 mo
Primary: LS BMD at 12 mo Secondary: LS BMD at 6 mo Hip and femoral neck BMD at 6 & 12 mo Exploratory: BMD at 1, 3, and 24 mo, BTM
5/5
Denosumab 60 mg/6 mo for 24 mo
Primary: LS BMD at 24 mo Secondary: Hip, femoral neck, one-third radius, and total body BMD at 24 mo, total volumetric BMD (QCT measure at distal radius), BTM Primary: new vertebral fractures
4/5
Denosumab 60 mg/6 mo for 36 mo
tion for postmenopausal women treated with denosumab compared with placebo was 2.10 (0.64 to 6.90), which was not statistically significant (Fig. 3B). Heterogeneity was noted for this outcome. The risk ratio (95% CI) of neoplasm for postmenopausal women treated with denosumab compared with placebo was 1.11 (0.91 to 1.36), which was not statistically significant (Fig. 3C). The risk ratio (95% CI) of study discontinuation due to AE for postmenopausal women treated with denosumab compared with placebo was 1.10 (0.83 to 1.47), which was not statistically significant (Fig. 3D). Mortality Rate The risk ratio (95% CI) of death for postmenopausal women treated with denosumab compared with placebo was 0.78 (0.57 to 1.06), which was not statistically signif-
Study or Subgroup Bone 2008 Cummings 2009 Ellis 2008 Total (95% CI)
Denosumab Placebo Events Total Events Total Weight 2 350 8
icant (Fig. 4). As there were very few events, heterogeneity measures were not applicable. Sensitivity Analysis A sensitivity analysis showed that by excluding the Ellis and coworkers (14) study from all outcomes, the risk ratios changed only very little (fracture risk: from 0.58 [0.52, 0.66] to 0.58 [0.51, 0.67]); (SAEs risk: from 1.33 [0.83, 2.14] to 1.29 [0.67, 2.49]; SAEs related to infection risk: from 2.10 [0.64, 6.90] to 2.32 [0.36, 14.85]; and neoplasm risk: from 1.11 [0.91, 1.36] to 1.12 [0.91, 1.37]) (mortality rate: stayed unchanged). Heterogeneity was still observed in the SAEs and SAEs related to infection risk outcomes. A sensitivity analysis, substituting 2-year for 3-year data on fracture rates from Cummings and coworkers (16), did not change the risk ratio for Risk Ratio M-H, Fixed, 95% CI
164 3702 125
8 165 600 3691 8 124
1.3% 97.4% 1.3%
0.25 [0.05, 1.17] 0.58 [0.51, 0.66] 0.99 [0.38, 2.56]
3991
3980
100.0%
0.58 [0.52, 0.66]
Total events 360 616 Heterogeneity: Chi² = 2.36, df = 2 (P = 0.31); I² = 15% Test for overall effect: Z = 8.67 (P < 0.00001)
Risk Ratio M-H, Fixed, 95% CI
0.01 0.1 1 10 100 Favors Denosumab Favors Placebo
Figure 2 Fracture risk for postmenopausal women assigned to denosumab compared with placebo. (Color version of figure is available online.)
C. von Keyserlingk et al.
A Study or Subgroup Bone 2008 Cummings 2009 Ellis 2008
183
Denosumab Placebo Events Total Events Total Weight 18 1004 19
Total (95% CI)
Risk Ratio M-H, Random, 95% CI
143 3430 106
9 146 972 3345 11 101
22.2% 52.8% 24.9%
2.04 [0.95, 4.39] 1.01 [0.94, 1.08] 1.65 [0.82, 3.28]
3679
3592
100.0%
1.33 [0.83, 2.14]
1041 992 Total events Heterogeneity: Tau² = 0.11; Chi² = 5.12, df = 2 (P = 0.08); I² = 61% Test for overall effect: Z = 1.19 (P = 0.23)
B Study or Subgroup Bone 2008 Cummings 2009 Ellis 2008
Denosumab Placebo Events Total Events Total Weight 8 159 3
Total (95% CI)
143 3430 106 3679
1 133 1
0.01 0.1 Favors Denosumab
Risk Ratio M-H, Random, 95% CI
146 3345 101
21.4% 59.4% 19.2%
8.17 [1.03, 64.47] 1.17 [0.93, 1.46] 2.86 [0.30, 27.03]
3592
100.0%
2.10 [0.64, 6.90]
Total events 170 135 Heterogeneity: Tau² = 0.61; Chi² = 3.98, df = 2 (P = 0.14); I² = 50% Test for overall effect: Z = 1.22 (P = 0.22)
C Study or Subgroup Bone 2008 Cummings 2009 Ellis 2008
Denosumab Placebo Events Total Events Total Weight 4 187 4
Total (95% CI)
D Study or Subgroup Bone 2008 Cummings 2009 Lewiecki 2007
146 1 166 3345 101 4
0.6% 97.1% 2.4%
4.08 [0.46, 36.10] 1.10 [0.90, 1.35] 0.95 [0.24, 3.71]
3679
3592
100.0%
1.11 [0.91, 1.36]
Total (95% CI)
1
Risk Ratio M-H, Fixed, 95% CI
143 3430 42
146 2 81 3345 39 1
2.3% 96.5% 1.2%
0.51 [0.05, 5.57] 1.12 [0.83, 1.50] 0.93 [0.06, 14.34]
3615
3530
100.0%
1.10 [0.83, 1.47]
84 95 Total events Heterogeneity: Chi² = 0.42, df = 2 (P = 0.81); I² = 0% Test for overall effect: Z = 0.66 (P = 0.51)
10 100 Favors Placebo
10 100 Favors Placebo
Risk Ratio M-H, Fixed, 95% CI
0.01 0.1 1 Favors Denosumab
Denosumab Placebo Events Total Events Total Weight 1 93 1
Risk Ratio M-H, Fixed, 95% CI
1
Risk Ratio M-H, Random, 95% CI
0.01 0.1 Favors Denosumab
143 3430 106
171 195 Total events Heterogeneity: Chi² = 1.43, df = 2 (P = 0.49); I² = 0% Test for overall effect: Z = 1.04 (P = 0.30)
Risk Ratio M-H, Random, 95% CI
10 100 Favors Placebo
Risk Ratio M-H, Fixed, 95% CI
0.01 0.1 Favors Denosumab
1
10 100 Favors Placebo
Figure 3 Risk of serious adverse events, serious adverse events related to infection, neoplasms, and study discontinuations due to adverse events for postmenopausal women assigned to denosumab compared with placebo. (A) Serious adverse event risk for postmenopausal women assigned to denosumab compared with placebo. (B) Serious adverse events related to infection risk for postmenopausal women assigned to denosumab compared with placebo. (C) Neoplasm risk for postmenopausal women assigned to denosumab compared with placebo. (D) Study discontinuation due to adverse event risk for postmenopausal women assigned to denosumab compared with placebo. (Color version of figure is available online.)
fractures. The risk ratio (95% CI) of fractures for the denosumab group compared with placebo remained at 0.58 (0.50 to 0.67). It was not possible to extract 2-year data for other outcomes. DISCUSSION This meta-analysis identified 4 RCTs contributing to some or all of the outcomes of interest. The relative risk of fractures was significantly lower in the denosumab group compared with the placebo group. Patients on denosumab experienced more AEs but this difference in the safety profile was not statistically significant between the denosumab and placebo groups. Mortality associated
with denosumab was lower than with placebo, but this was also not statistically significant. Our meta-analysis found 42% risk reduction in relative fracture risk compared with the insignificant 26% risk reduction reported in a previous meta-analysis by Anastasilakis and coworkers (25) The large RCT (16) by Cummings and coworkers (16) significantly affected the results in favor of denosumab and accounted for the increased risk reduction observed in our metaanalysis. This study had a substantially larger sample size, was longer in duration than previous trials, and was the first study powered to evaluate fracture risk as primary outcome.
184
Denosumab for women with low bone mass Study or Subgroup Bone 2008 Cummings 2009 Ellis 2008 Total (95% CI)
Denosumab Placebo Events Total Events Total Weight 0 70 1
164 3902 125
0 165 90 3906 1 124
4191 4195 Total events 71 91 Heterogeneity: Chi² = 0.03, df = 1 (P = 0.86); I² = 0% Test for overall effect: Z = 1.58 (P = 0.11)
Risk Ratio M-H, Fixed, 95% CI
98.9% 1.1%
Not estimable 0.78 [0.57, 1.06] 0.99 [0.06, 15.68]
100.0%
0.78 [0.57, 1.06]
Risk Ratio M-H, Fixed, 95% CI
0.01 0.1 Favors Denosumab
1
10 100 Favors Placebo
Figure 4 Mortality rate for postmenopausal women assigned to denosumab compared with placebo. (Color version of figure is available online.)
It should be noted that there was heterogeneity between the studies, as observed in the SAEs risk and SAEs related to infection risk outcomes (Fig. 3). First, the sample sizes varied across studies. Bone and coworkers (13), Ellis and coworkers (14), and Lewiecki and coworkers (15) had considerably small sample sizes (⬍150), while Cummings and coworkers (16) randomized more than 7000 patients. Second, there was variability in patient characteristics. The patient population in Ellis and coworkers (14) had less severe mean BMD T-scores compared with the other studies and referred to women with nonmetastatic breast cancer. Conducting a sensitivity analysis by excluding Ellis and coworkers (14) from all outcomes did not eliminate heterogeneity from the SAEs risk and SAEs related to infection risk outcomes. Third, there was variability in the study durations. The study by Cummings and coworkers (16) was based on 3-year data, while the other studies were of 2-year duration. Results from a sensitivity analysis showed that substituting 2-year for 3-year data on fracture rates from Cummings and coworkers (16) did not change the risk ratio of fractures for the denosumab group compared with the placebo group. Patients on denosumab experienced more SAEs, SAEs related to infections, neoplasms in absolute numbers, and more frequently discontinued the study due to AEs. However, the differences in the safety profile were not statistically significant between the denosumab and placebo groups. Studies have shown that there is a biological plausibility that RANKL inhibition affects susceptibility to infections and malignancies. RANKL inhibition has an effect on T-cell growth and dendritic cell function as well as the natural defense mechanism against tumorigenesis (26). Evidence of studies, which compare incidence of infections and neoplasms in denosumab-treated patients, compared with placebo show inconclusive results (26). However, as a negative effect of denosumab may likely be observed after a much longer exposure, than the duration of trials that have been conducted, it will be important to report any postmarketing occurrence of adverse events associated with denosumab. Similarly, while denosumab did not significantly decrease mortality, data on longterm continued use of denosumab are necessary to capture the full effect on mortality.
Clinical studies on the treatment of osteoporosis demonstrated the effect of denosumab on stabilizing BMD and decreasing bone turnover markers. However, efficacy evaluation on the basis of these variables was not possible, as relevant data were missing, or reported in figures, or without standard deviations (25). CONCLUSION Our analysis observed a significant reduction in relative fracture risk in the denosumab compared with the control group. There was no significant increase in relative serious adverse event risk in patients on denosumab. Fractures are the most serious consequence of osteoporosis (27) and, given that denosumab is administered subcutaneously, it could present an effective new treatment for osteoporosis with fewer adherence barriers than faced with present treatments (12). REFERENCES 1. Brown JP, Fortier M, Frame H, Lalonde A, Papaioannou A, Senikas V, et al. Canadian consensus conference on osteoporosis, 2006 update. J Obstet Gynaecol Can 2006;28:S95-112. 2. International Osteoporosis Foundation. (2009) Facts and statistics about osteoporosis and its impact. http://www.iofbonehealth. org/facts-and-statistics.html. Accessed 6 Jan 2010. 3. Stevenson M, Jones ML, De NE, Brewer N, Davis S, Oakley J. A systematic review and economic evaluation of alendronate, etidronate, risedronate, raloxifene and teriparatide for the prevention and treatment of postmenopausal osteoporosis. Health Technol Assess 2005;9:1-160. 4. Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. Lancet 2002;359:1761-7. 5. Adachi JD, Kennedy CC, Papaioannou A, Ioannidis G, Leslie WD, Walker V. Treating osteoporosis in Canada: what clinical efficacy data should be considered by policy decision makers? Osteoporos Int 2009;20:1785-93. 6. Clarke BL. New and emerging treatments for osteoporosis. Clin Endocrinol (Oxf) 2009;71:309-21. 7. Melo MD, Obeid G. Osteonecrosis of the jaws in patients with a history of receiving bisphosphonate therapy: strategies for prevention and early recognition. J Am Dent Assoc 2005;136:1675-81. 8. Odvina CV, Zerwekh JE, Rao DS, Maalouf N, Gottschalk FA, Pak CY. Severely suppressed bone turnover: a potential complication of alendronate therapy. J Clin Endocrinol Metab 2005;90: 1294-301. 9. Whyte MP, Wenkert D, Clements KL, McAlister WH, Mumm S. Bisphosphonate-induced osteopetrosis. N Engl J Med 2003; 349:457-63.
C. von Keyserlingk et al. 10. McClung MR, Lewiecki EM, Cohen SB, Bolognese MA, Woodson GC, Moffett AH, et al. Denosumab in postmenopausal women with low bone mineral density. N Engl J Med 2006; 354:821-31. 11. National Horizon Scanning Centre (NHSC). (2008) Denosumab (AMG 162) for the prevention of therapy-induced bone loss in non-metastatic breast and prostate cancer. National Horizon Centre. NHS National Institute for Health Research. University of Birmingham. http://www.haps.bham.ac.uk/publichealth/ horizon/outputs/documents/2008/sept-dec/Denosumab_boneloss_ breastandprostate_nonconfSep08.pdf. Accessed Jan 21 2010. 12. Burkiewicz JS, Scarpace SL, Bruce SP. Denosumab in osteoporosis and oncology. Ann Pharmacother 2009;43:1445-55. 13. Bone HG, Bolognese MA, Yuen CK, Kendler DL, Wang H, Liu Y, et al. Effects of denosumab on bone mineral density and bone turnover in postmenopausal women. J Clin Endocrinol Metab 2008;93:2149-57. 14. Ellis GK, Bone HG, Chlebowski R, Paul D, Spadafora S, Smith J, et al. Randomized trial of denosumab in patients receiving adjuvant aromatase inhibitors for nonmetastatic breast cancer. J Clin Oncol 2008;26:4875-82. 15. Lewiecki EM, Miller PD, McClung MR, Cohen SB, Bolognese MA, Liu Y, et al. Two-year treatment with denosumab (AMG 162) in a randomized phase 2 study of postmenopausal women with low BMD. J Bone Miner Res 2007;22:1832-41. 16. Cummings SR, San MJ, McClung MR, Siris ES, Eastell R, Reid IR, et al. Denosumab for prevention of fractures in postmenopausal women with osteoporosis. N Engl J Med 2009;361:756-65. 17. Kong YY, Yoshida H, Sarosi I, Tan HL, Timms E, Capparelli C, et al. OPGL is a key regulator of osteoclastogenesis, lymphocyte development and lymph-node organogenesis. Nature 1999;397: 315-23. 18. Wong BR, Josien R, Lee SY, Sauter B, Li HL, Steinman RM, et al. TRANCE (tumor necrosis factor [TNF]-related activation-induced cytokine), a new TNF family member predominantly ex-
185
19.
20.
21.
22.
23. 24.
25.
26.
27.
pressed in T cells, is a dendritic cell-specific survival factor. J Exp Med 1997;186:2075-80. Miller PD, Barlas S, Brenneman SK, Abbott TA, Chen YT, Barrett-Connor E, et al. An approach to identifying osteopenic women at increased short-term risk of fracture. Arch Intern Med 2004;164:1113-20. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17: 1-12. Altman DG, Schulz KF, Moher D, Egger M, Davidoff F, Elbourne D, et al. The revised CONSORT statement for reporting randomized trials: explanation and elaboration. Ann Intern Med 2001;134:663-94. Higgins JPT GS. The Cochrane Collaboration (2009) Cochrane Handbook for Systematic Reviews of Interventions Version 5.0.2. http://www.cochrane.org/resources/handbook/. Accessed 10 Jan 2010. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557-60. Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Quality of Reporting of Meta-analyses Lancet 1999;354:1896-900. Anastasilakis AD, Toulis KA, Goulis DG, Polyzos SA, Delaroudis S, Giomisi A, et al. Efficacy and safety of denosumab in postmenopausal women with low bone mineral density or osteoporosis: a systematic review and a meta-analysis. Horm Metab Res 2009;41: 721-9. Anastasilakis AD, Toulis KA, Polyzos SA, Terpos E. RANKL inhibition for the management of patients with benign metabolic bone disorders. Exp Opin Investig Drugs 2009;18:1085102. De Laet CE, Pols HA. Fractures in the elderly: epidemiology and demography. Baillieres Best Pract Res Clin Endocrinol Metab 2000;14:171-9.
186
Denosumab for women with low bone mass
Appendix 1 Literature Search Strategy Overview Interface: Databases:
Date of search: Study types: Limits: Syntax guide / .sh MeSH Fs exp * # ? ADJ ADJ# .ti .ab .hw .pt .rn No.
Ovid EMBASE ⬍1996 to 2009, week 53⬎ Ovid Medline ⬍1996 to December week 5, 2009⬎ Ovid Medline in-process and other nonindexed citations ⬍January 7, 2010⬎ Note: Subject headings have been customized for each database. Duplicates between databases were removed in Ovid. January 8, 2010 Randomized controlled trials Publication years 2009 to January 2010 At the end of a phrase, searches the phrase as a subject heading At the end of a phrase, searches the phrase as a subject heading Medical subject heading Floating subheading Explode a subject heading Before a word, indicates that the marked subject heading is a primary topic; or, after a word, a truncation symbol (wildcard) to retrieve plurals or varying endings Truncation symbol for 1 character Truncation symbol for 1 or no characters only Requires words are adjacent to each other (in any order) Adjacency within # number of words (in any order) Title Abstract Heading word; usually includes subject headings and controlled vocabulary Publication type CAS registry number Searches
Results
Clinical multidatabase strategy 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Denosumab exp denosumab/ denosumab.ti,ab. AMG 162 or 615258 to 40-7.ti,ab.rn. Osteoporosis exp osteoporosis/ osteoporo*.ti,ab. 1 or 2 or 3 4 or 5 6 and 7 Date limit limit 8 to yr ⫽ “2009-Current” Randomized controlled trial filter Randomized controlled trials as topic/ Randomized controlled trial/ (Randomized controlled trial or controlled clinical trial).pt. Randomization/ Double-blind procedure/ Double-blind studies/ Single-blind procedure/ Single-blind studies/ Placebos/or placebo/ (random* or sham or placebo*).ti,ab,hw. ((singl* or doubl*) adj (blind* or dumm* or mask*)).ti,ab,hw. ((tripl* or trebl*) adj (blind* or dumm* or mask*)).ti,ab,hw. or/10 to 21 8 and 22 Date limit Limit 23 to yr ⫽ “2009-Current” From 24 keep 1 to 53
470 265 162 58,185 53,123 637 73,830 353 110 203,369 332,000 217,635 56,627 56,107 114,704 7879 18,648 107,423 925,981 159,499 292 966,243 157 53 53
Other databases Cochrane Library, CENTRAL, Issue 4, 2009 Same MeSH, keywords, and date limits used as per Medline search, excluding study types and human restrictions. Syntax adjusted for Cochrane Library databases.
Clinical Efficacy and Safety of Denosumab in Postmenopausal Women with Low Bone Mineral Density and Osteoporosis: A Meta-Analysis Camilla von Keyserlingk, MA,* Robert Hopkins, MSc,*,† Athanasios Anastasilakis, MD,‡ Konstantinos Toulis, MD,§ Ron Goeree, MA,*,† Jean-Eric Tarride, PhD,*,† and Feng Xie, PhD*,†
Objective: Clinical trials indicate that denosumab could be a potential treatment for postmenopausal osteoporosis. The objective of this meta-analysis was to assess the clinical efficacy and safety of offering denosumab to postmenopausal women with low bone mass. Methods: Data sources included MEDLINE, EMBASE, and Cochrane Central Register of Controlled Trials (CENTRAL) from inception to February 3, 2010 and bibliographies of reviews. Randomized controlled trials comparing the efficacy and safety of denosumab to placebo for treatment of low bone mass (low bone mineral density or osteoporosis) in postmenopausal women were selected. Two reviewers independently abstracted data on study general characteristics and outcomes. Review Manager 5.0 software was used for data syntheses and meta-analysis. Results: The database search revealed 4 studies (comprising 8864 patients randomized) that met the inclusion criteria and contributed to some or all of the meta-analysis outcomes. Relative risk (95% CI) of fractures for the denosumab compared with placebo group was 0.58 (0.52 to 0.66); relative risk (95% CI) of serious adverse events was 1.33 (0.83 to 2.14); relative risk (95% CI) of serious adverse events related to infection was 2.10 (0.64 to 6.90); relative risk (95% CI) of neoplasm was 1.11 (0.91 to 1.36); relative risk (95% CI) of study discontinuation due to adverse events was 1.10 (0.83 to 1.47); and relative risk (95% CI) of death was 0.78 (0.57 to 1.06). Findings remained robust to sensitivity analyses. Conclusion: Our analysis found a significant reduction in relative fracture risk in the denosumab compared with the placebo group. © 2011 Elsevier Inc. All rights reserved. Semin Arthritis Rheum 41:178-186 Keywords: osteoporosis, low bone mineral density, bone mass, denosumab, postmenopausal, metaanalysis
O
steoporosis is a systemic skeletal disease characterized by low bone mass and microarchitectural deterioration of bone tissue, with a consequent increase in bone fragility and fracture risk (1). It is esti-
*Programs for Assessment of Technology in Health (PATH) Research Institute, St. Joseph’s Healthcare, Hamilton, Ontario, Canada. †Department of Clinical Epidemiology and Biostatistics, Faculty of Health Sciences, McMaster University, Hamilton, Ontario, Canada. ‡Department of Endocrinology, 424 Military Hospital, Thessaloniki, Greece. §Department of Endocrinology, Aristotle University, Thessaloniki, Greece. The authors have no conflicts of interest to disclose. Address reprint requests to Feng Xie, PhD, PATH Research Institute, 25 Main Street West, Suite 2000, Hamilton, Ontario, Canada L8P 1H1. E-mail: barboxie@ gmail.com.
178
mated that 75 million people in the United States, Europe, and Japan are affected by this disease (2). The incidence of fractures increases exponentially with age, resulting in high morbidity, mortality, and health care expenditure (3). For white women, the lifetime risk of sustaining a hip fracture is 1 in 6, compared with a 1-in-9 risk of developing breast cancer (4). Currently standard therapies for osteoporosis focus mainly on antiresorptive agents, with bisphosphonates as a first line of treatment (5). Antiresorptive therapy has proven successful in reducing bone loss and preventing fractures in postmenopausal women over multiple years (6). Concerns regarding potential fracture risk associated with long-term use of bisphosphonates (7-9), low adher-
0049-0172/11/$-see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.semarthrit.2011.03.005
C. von Keyserlingk et al.
ence rates with currently available agents (10), and advances in bone biology spur the development of new therapeutic agents (6). Recent discoveries in pathophysiology of bone metabolism have identified the nuclear factor kappa B ligand (RANKL) as a critical factor in osteoclastdriven bone resorption, and subsequent bone loss (11). Denosumab is a fully human monoclonal IgG2 antibody that specifically targets RANKL, thereby inhibiting osteoclast differentiation, activation, and survival and subsequently suppressing bone resorption (11). Denosumab is administered subcutaneously and could lead to better adherence rates than observed with oral bisphosphonates (12). Denosumab has been found to be effective in decreasing bone turnover markers and stabilizing or increasing bone mineral density (BMD). In the case of therapies for osteoporosis, it has been difficult to evaluate fracture outcomes through comparative randomized controlled trials (RCTs) due to the large sample sizes required for meaningful comparisons. RCTs with relatively small sample sizes (⬍150) recorded the incidence of fractures without conclusive results (13-15). To our best knowledge, the first large phase 3 RCT (16), powered to evaluate comparative fracture risk, was recently published and found a 42% risk reduction in relative fracture risk in the denosumab compared with placebo group at a dose of 60 mg subcutaneously every 6 months (16). While evidence on denosumab seems promising, some concerns have been raised. Since RANK activation by RANKL is also essential for T-cell growth and dendritic-cell function (17,18), inhibition of its action could simultaneously affect the immune system, leading to susceptibility of infections or malignancies. We conducted a systematic review and meta-analysis of RCTs to provide evidence-based knowledge on efficacy and safety of offering denosumab to postmenopausal women with low bone mass. METHODS Data Sources and Searches We performed a literature search for the purpose of identifying RCTs. To identify eligible studies, the main search was conducted in the electronic databases MEDLINE, EMBASE, and Cochrane Central Register of Controlled Trials (CENTRAL) from inception to February 3, 2010. The search strategy with controlled vocabulary, registry number of the drug, and keywords focused on the concepts of “denosumab” and “osteoporosis.” These concepts were combined with methodological filters (supplied by an information specialist) to limit retrieval to RCTs (Appendix 1). No language restrictions were applied. Study Selection Eligible studies for the meta-analysis were RCTs, which compared the efficacy and safety/tolerability profile of denosumab to placebo for the treatment of low bone min-
179
eral density or osteoporosis in postmenopausal women. Studies on individuals with low bone mineral density were included as it is well documented that more than 50% of fractures occur in women with low bone mineral density (19). Trials were excluded, if uncontrolled, and if they included men and/or subjects who had conditions that influenced bone metabolism (eg, metastatic breast cancer). Reviews, case series, abstracts, letters to the editor, phase 1, or observational studies were not eligible. However, reference sections of reviews were screened to identify possible relevant studies. Two reviewers (FX and CvK) independently screened the titles and abstracts of records retrieved from the searches for relevance using predefined checklists. Kappa statistics for interrater agreement were calculated. Discrepancies between reviewers were discussed until a consensus was reached. The full text of any relevant study passing title/abstract screening was retrieved and assessed for inclusion. Two reviewers (FX and CvK), using explicit predetermined criteria, made inclusion and exclusion decisions independently. These were checked for agreement, and discrepancies between reviewers were discussed until consensus was reached. Data Extraction Relevant information from the studies was extracted independently by 2 reviewers (FX and CvK) into predefined data extraction forms. Study characteristics and outcomes for both intervention and control groups were recorded and double-checked. Study quality was assessed independently by 2 reviewers (FX and CvK) using the assessment scale developed by Jadad and coworkers (20). The Jadad scale contains 3 items related directly to the reduction of bias. The items are presented in the form of 5 questions eliciting yes or no answers. Points awarded depended on the quality of the description of the methods to generate the sequence of randomization and of double blinding and the reporting of withdrawals and dropouts (20). Any discrepancies in data extraction were discussed until consensus was reached. Data Synthesis and Analysis Data were entered into the computer program Review Manager (Version 5.0. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2008). Efficacy was evaluated on the basis of the overall incidence of fractures in the treatment and placebo groups, involving a combination of clinical and radiological fractures reported in RCTs. The Mantel-Haenszel method was used to calculate risk ratios (RRs) and their 95% confidence intervals (CI) using either fixed or random effects model, depending on the amount of heterogeneity observed. The safety/tolerability profile was evaluated on the basis of the reported incidence of serious adverse events (SAEs), SAEs related to infection, neoplasms, and the percentage of study discontinuations due to adverse events. Medical conditions that required hospitalization were recorded as
180
Denosumab for women with low bone mass
SAEs. Adverse events (AEs) analyzed were chosen on the basis of biological plausibility. AE outcomes were examined according to the guidelines issued by the CONSORT group (21), stating that “intention-to-treat” analysis should not be used for examining adverse effects. Instead we examined AE outcomes on the basis of an “on-treatment” analysis whereby total patients completing the trial plus dropouts due to adverse events were included. Mantel-Haenszel method was used to calculate RRs and their 95% CI using either fixed or random effects model, whichever was appropriate. Mortality was expressed as RRs with 95% CI using Mantel-Haenszel method and fixed effects model. Heterogeneity between the results of different studies was assessed with the 2 test for significance (a cutoff of 10% for significance was used) and presented by I2 test (22), which can be interpreted as the percentage of total variation across several studies due
to heterogeneity (23). Due to the small number of studies, funnel plot analysis to check for publication bias was not performed. Sensitivity Analyses To check for possible biases introduced by heterogeneity, 2 one-way sensitivity analyses were conducted. First, there was variability in BMD T-scores across studies at patient enrollment (T-score ⱕ ⫺2.5 and ⱖ ⫺4.0 in Cummings and coworkers (16); T-score ⱕ ⫺1.0 and ⱖ ⫺2.5 in Bone and coworkers (13) and Ellis and coworkers (14); T-score ⱕ ⫺1.8 and ⱖ ⫺4.0 in Lewiecki and coworkers (15)). When we compared the actual mean T-scores across studies, it seemed that T-scores were least severe in Ellis and coworkers (14) In addition, the patient population in the study by Ellis and coworkers (14) referred to
Figure 1 QUORUM flowchart.
C. von Keyserlingk et al.
181
Table 1 Main Characteristics of the 4 Randomized Controlled Trials Included in the Meta-Analysis Author, year, reference Lewiecki 2007 [15]
Ellis 2008 [14]
Bone 2008 [13]
Design Randomized, plc-controlled, phase II dose-ranging study (8 doubleblind ⫹ 1 open-label treatment groups) Multicenter, randomized, doubleblind, plc-controlled, phase III study
Multicenter, randomized, doubleblind, plc-controlled, phase III study
Cummings 2009 [17]
International, multicenter, randomized, double-blind, plccontrolled, phase III study Secondary: Nonvertebral, 3/5 and hip fractures, BMD, BTM, safety
Total Study Size (plc) [60 mg/6 mo] 412 (46) [47]
252 (125) [127]
332 (166) [166]
7868 (3935) [3933]
Inclusion Criteria Postmenopausal women (mean age 62.5) with low bone mineral density or osteoporosis (T-score ⱕ ⫺1.8 and ⱖ ⫺4.0) Women (mean age 59.5; 73% postmenopausal [50% with more than 10 years and 23% with 5 to 10 years since their last menstrual period]) with nonmetastatic breast cancer under aromatase inhibitors and low bone mass (T-score ⱕ ⫺1.0 and ⱖ ⫺2.5) Postmenopausal women (mean age 59.4) with low bone mineral density or osteoporosis (T-score ⱕ ⫺1.0 and ⱖ ⫺2.5) Postmenopausal women (mean age 72.3) with osteoporosis (T-score ⱕ ⫺2.5 and ⱖ ⫺4.0)
plc, placebo; BMD, bone mineral density (measured by DXA); BTM, bone turnover marker; LS, lumbar spine; QCT, quantitative computed tomography.
women with nonmetastatic breast cancer on treatment with aromatase inhibitors, representing a population with a potentially different biological behavior than that of the target population (postmenopausal healthy women). A sensitivity analysis was conducted by excluding the study by Ellis and coworkers (14) from all outcomes. Second, there was variability in study duration. Bone and coworkers (13), Ellis and coworkers (14), and Lewiecki and coworkers (15) used 2-year interval data, while Cummings and coworkers (16) used 3-year data. To verify that results were similar during each year of the trial, as reported by Cummings and coworkers (16), we conducted a sensitivity analysis by substituting 2-year for 3-year data on fracture rates from Cummings and coworkers (16).
by combining different dosages of denosumab in this study. Results of the other included studies are based on a 60 mg/6 months dosage. The RCT by Ellis and coworkers (14) was not included in pooling the study discontinuation rate due to adverse event, as the rate was not reported by this study. All studies scored either 4 or 5 of 5 possible points on the Jadad scale, except the RCT by Cummings and coworkers (16), which scored only 3 of 5 points, as it seemed that the process of randomization was not appropriately described and reasons for withdrawals were not reported. Table 1 summarizes the main characteristics and the quality score for the Jadad scale of the 4 studies included in the present meta-analysis.
RESULTS
Meta-Analysis
Literature Review
Efficacy
The search identified 148 potentially relevant citations. A QUORUM (24) flowchart summarizing search results is shown in Figure 1. After screening titles and abstracts, 137 citations were excluded. Thus a total of 11 citations were reviewed in full text. According to our eligibility criteria, 4 studies (13-16) were included in the meta-analysis, contributing to some or all of the outcomes of interest (Fig. 1). Notably, the RCT by Lewiecki and coworkers (15) was excluded when pooling the outcomes of fracture risk, SAEs risk, SAEs related to infection risk, neoplasm risk, and mortality risk as these outcomes were reported
The risk ratio (95% CI) of fractures for postmenopausal women with low bone mass treated with denosumab compared with placebo was 0.58 (0.52 to 0.66) (Fig. 2). Safety The risk ratio (95% CI) of SAEs for postmenopausal women treated with denosumab compared with placebo was 1.33 (0.83 to 2.14), which was not statistically significant (Fig. 3A). Heterogeneity was noted for this outcome. The risk ratio (95% CI) of SAEs related to infec-
182
Denosumab for women with low bone mass
Table 1 Continued Intervention
Endpoint
Quality Score
Denosumab 6, 14, or 30 mg/3 mo or 14, 60, 100, 210 mg/6 mo, or open-label alendronate 70 mg/wk for 24 mo
Primary: LS BMD at 24 mo Secondary: Hip, distal one-third radius, total body BMD, BTM
4/5
Denosumab 60 mg/6 mo for 24 mo
Primary: LS BMD at 12 mo Secondary: LS BMD at 6 mo Hip and femoral neck BMD at 6 & 12 mo Exploratory: BMD at 1, 3, and 24 mo, BTM
5/5
Denosumab 60 mg/6 mo for 24 mo
Primary: LS BMD at 24 mo Secondary: Hip, femoral neck, one-third radius, and total body BMD at 24 mo, total volumetric BMD (QCT measure at distal radius), BTM Primary: new vertebral fractures
4/5
Denosumab 60 mg/6 mo for 36 mo
tion for postmenopausal women treated with denosumab compared with placebo was 2.10 (0.64 to 6.90), which was not statistically significant (Fig. 3B). Heterogeneity was noted for this outcome. The risk ratio (95% CI) of neoplasm for postmenopausal women treated with denosumab compared with placebo was 1.11 (0.91 to 1.36), which was not statistically significant (Fig. 3C). The risk ratio (95% CI) of study discontinuation due to AE for postmenopausal women treated with denosumab compared with placebo was 1.10 (0.83 to 1.47), which was not statistically significant (Fig. 3D). Mortality Rate The risk ratio (95% CI) of death for postmenopausal women treated with denosumab compared with placebo was 0.78 (0.57 to 1.06), which was not statistically signif-
Study or Subgroup Bone 2008 Cummings 2009 Ellis 2008 Total (95% CI)
Denosumab Placebo Events Total Events Total Weight 2 350 8
icant (Fig. 4). As there were very few events, heterogeneity measures were not applicable. Sensitivity Analysis A sensitivity analysis showed that by excluding the Ellis and coworkers (14) study from all outcomes, the risk ratios changed only very little (fracture risk: from 0.58 [0.52, 0.66] to 0.58 [0.51, 0.67]); (SAEs risk: from 1.33 [0.83, 2.14] to 1.29 [0.67, 2.49]; SAEs related to infection risk: from 2.10 [0.64, 6.90] to 2.32 [0.36, 14.85]; and neoplasm risk: from 1.11 [0.91, 1.36] to 1.12 [0.91, 1.37]) (mortality rate: stayed unchanged). Heterogeneity was still observed in the SAEs and SAEs related to infection risk outcomes. A sensitivity analysis, substituting 2-year for 3-year data on fracture rates from Cummings and coworkers (16), did not change the risk ratio for Risk Ratio M-H, Fixed, 95% CI
164 3702 125
8 165 600 3691 8 124
1.3% 97.4% 1.3%
0.25 [0.05, 1.17] 0.58 [0.51, 0.66] 0.99 [0.38, 2.56]
3991
3980
100.0%
0.58 [0.52, 0.66]
Total events 360 616 Heterogeneity: Chi² = 2.36, df = 2 (P = 0.31); I² = 15% Test for overall effect: Z = 8.67 (P < 0.00001)
Risk Ratio M-H, Fixed, 95% CI
0.01 0.1 1 10 100 Favors Denosumab Favors Placebo
Figure 2 Fracture risk for postmenopausal women assigned to denosumab compared with placebo. (Color version of figure is available online.)
C. von Keyserlingk et al.
A Study or Subgroup Bone 2008 Cummings 2009 Ellis 2008
183
Denosumab Placebo Events Total Events Total Weight 18 1004 19
Total (95% CI)
Risk Ratio M-H, Random, 95% CI
143 3430 106
9 146 972 3345 11 101
22.2% 52.8% 24.9%
2.04 [0.95, 4.39] 1.01 [0.94, 1.08] 1.65 [0.82, 3.28]
3679
3592
100.0%
1.33 [0.83, 2.14]
1041 992 Total events Heterogeneity: Tau² = 0.11; Chi² = 5.12, df = 2 (P = 0.08); I² = 61% Test for overall effect: Z = 1.19 (P = 0.23)
B Study or Subgroup Bone 2008 Cummings 2009 Ellis 2008
Denosumab Placebo Events Total Events Total Weight 8 159 3
Total (95% CI)
143 3430 106 3679
1 133 1
0.01 0.1 Favors Denosumab
Risk Ratio M-H, Random, 95% CI
146 3345 101
21.4% 59.4% 19.2%
8.17 [1.03, 64.47] 1.17 [0.93, 1.46] 2.86 [0.30, 27.03]
3592
100.0%
2.10 [0.64, 6.90]
Total events 170 135 Heterogeneity: Tau² = 0.61; Chi² = 3.98, df = 2 (P = 0.14); I² = 50% Test for overall effect: Z = 1.22 (P = 0.22)
C Study or Subgroup Bone 2008 Cummings 2009 Ellis 2008
Denosumab Placebo Events Total Events Total Weight 4 187 4
Total (95% CI)
D Study or Subgroup Bone 2008 Cummings 2009 Lewiecki 2007
146 1 166 3345 101 4
0.6% 97.1% 2.4%
4.08 [0.46, 36.10] 1.10 [0.90, 1.35] 0.95 [0.24, 3.71]
3679
3592
100.0%
1.11 [0.91, 1.36]
Total (95% CI)
1
Risk Ratio M-H, Fixed, 95% CI
143 3430 42
146 2 81 3345 39 1
2.3% 96.5% 1.2%
0.51 [0.05, 5.57] 1.12 [0.83, 1.50] 0.93 [0.06, 14.34]
3615
3530
100.0%
1.10 [0.83, 1.47]
84 95 Total events Heterogeneity: Chi² = 0.42, df = 2 (P = 0.81); I² = 0% Test for overall effect: Z = 0.66 (P = 0.51)
10 100 Favors Placebo
10 100 Favors Placebo
Risk Ratio M-H, Fixed, 95% CI
0.01 0.1 1 Favors Denosumab
Denosumab Placebo Events Total Events Total Weight 1 93 1
Risk Ratio M-H, Fixed, 95% CI
1
Risk Ratio M-H, Random, 95% CI
0.01 0.1 Favors Denosumab
143 3430 106
171 195 Total events Heterogeneity: Chi² = 1.43, df = 2 (P = 0.49); I² = 0% Test for overall effect: Z = 1.04 (P = 0.30)
Risk Ratio M-H, Random, 95% CI
10 100 Favors Placebo
Risk Ratio M-H, Fixed, 95% CI
0.01 0.1 Favors Denosumab
1
10 100 Favors Placebo
Figure 3 Risk of serious adverse events, serious adverse events related to infection, neoplasms, and study discontinuations due to adverse events for postmenopausal women assigned to denosumab compared with placebo. (A) Serious adverse event risk for postmenopausal women assigned to denosumab compared with placebo. (B) Serious adverse events related to infection risk for postmenopausal women assigned to denosumab compared with placebo. (C) Neoplasm risk for postmenopausal women assigned to denosumab compared with placebo. (D) Study discontinuation due to adverse event risk for postmenopausal women assigned to denosumab compared with placebo. (Color version of figure is available online.)
fractures. The risk ratio (95% CI) of fractures for the denosumab group compared with placebo remained at 0.58 (0.50 to 0.67). It was not possible to extract 2-year data for other outcomes. DISCUSSION This meta-analysis identified 4 RCTs contributing to some or all of the outcomes of interest. The relative risk of fractures was significantly lower in the denosumab group compared with the placebo group. Patients on denosumab experienced more AEs but this difference in the safety profile was not statistically significant between the denosumab and placebo groups. Mortality associated
with denosumab was lower than with placebo, but this was also not statistically significant. Our meta-analysis found 42% risk reduction in relative fracture risk compared with the insignificant 26% risk reduction reported in a previous meta-analysis by Anastasilakis and coworkers (25) The large RCT (16) by Cummings and coworkers (16) significantly affected the results in favor of denosumab and accounted for the increased risk reduction observed in our metaanalysis. This study had a substantially larger sample size, was longer in duration than previous trials, and was the first study powered to evaluate fracture risk as primary outcome.
184
Denosumab for women with low bone mass Study or Subgroup Bone 2008 Cummings 2009 Ellis 2008 Total (95% CI)
Denosumab Placebo Events Total Events Total Weight 0 70 1
164 3902 125
0 165 90 3906 1 124
4191 4195 Total events 71 91 Heterogeneity: Chi² = 0.03, df = 1 (P = 0.86); I² = 0% Test for overall effect: Z = 1.58 (P = 0.11)
Risk Ratio M-H, Fixed, 95% CI
98.9% 1.1%
Not estimable 0.78 [0.57, 1.06] 0.99 [0.06, 15.68]
100.0%
0.78 [0.57, 1.06]
Risk Ratio M-H, Fixed, 95% CI
0.01 0.1 Favors Denosumab
1
10 100 Favors Placebo
Figure 4 Mortality rate for postmenopausal women assigned to denosumab compared with placebo. (Color version of figure is available online.)
It should be noted that there was heterogeneity between the studies, as observed in the SAEs risk and SAEs related to infection risk outcomes (Fig. 3). First, the sample sizes varied across studies. Bone and coworkers (13), Ellis and coworkers (14), and Lewiecki and coworkers (15) had considerably small sample sizes (⬍150), while Cummings and coworkers (16) randomized more than 7000 patients. Second, there was variability in patient characteristics. The patient population in Ellis and coworkers (14) had less severe mean BMD T-scores compared with the other studies and referred to women with nonmetastatic breast cancer. Conducting a sensitivity analysis by excluding Ellis and coworkers (14) from all outcomes did not eliminate heterogeneity from the SAEs risk and SAEs related to infection risk outcomes. Third, there was variability in the study durations. The study by Cummings and coworkers (16) was based on 3-year data, while the other studies were of 2-year duration. Results from a sensitivity analysis showed that substituting 2-year for 3-year data on fracture rates from Cummings and coworkers (16) did not change the risk ratio of fractures for the denosumab group compared with the placebo group. Patients on denosumab experienced more SAEs, SAEs related to infections, neoplasms in absolute numbers, and more frequently discontinued the study due to AEs. However, the differences in the safety profile were not statistically significant between the denosumab and placebo groups. Studies have shown that there is a biological plausibility that RANKL inhibition affects susceptibility to infections and malignancies. RANKL inhibition has an effect on T-cell growth and dendritic cell function as well as the natural defense mechanism against tumorigenesis (26). Evidence of studies, which compare incidence of infections and neoplasms in denosumab-treated patients, compared with placebo show inconclusive results (26). However, as a negative effect of denosumab may likely be observed after a much longer exposure, than the duration of trials that have been conducted, it will be important to report any postmarketing occurrence of adverse events associated with denosumab. Similarly, while denosumab did not significantly decrease mortality, data on longterm continued use of denosumab are necessary to capture the full effect on mortality.
Clinical studies on the treatment of osteoporosis demonstrated the effect of denosumab on stabilizing BMD and decreasing bone turnover markers. However, efficacy evaluation on the basis of these variables was not possible, as relevant data were missing, or reported in figures, or without standard deviations (25). CONCLUSION Our analysis observed a significant reduction in relative fracture risk in the denosumab compared with the control group. There was no significant increase in relative serious adverse event risk in patients on denosumab. Fractures are the most serious consequence of osteoporosis (27) and, given that denosumab is administered subcutaneously, it could present an effective new treatment for osteoporosis with fewer adherence barriers than faced with present treatments (12). REFERENCES 1. Brown JP, Fortier M, Frame H, Lalonde A, Papaioannou A, Senikas V, et al. Canadian consensus conference on osteoporosis, 2006 update. J Obstet Gynaecol Can 2006;28:S95-112. 2. International Osteoporosis Foundation. (2009) Facts and statistics about osteoporosis and its impact. http://www.iofbonehealth. org/facts-and-statistics.html. Accessed 6 Jan 2010. 3. Stevenson M, Jones ML, De NE, Brewer N, Davis S, Oakley J. A systematic review and economic evaluation of alendronate, etidronate, risedronate, raloxifene and teriparatide for the prevention and treatment of postmenopausal osteoporosis. Health Technol Assess 2005;9:1-160. 4. Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. Lancet 2002;359:1761-7. 5. Adachi JD, Kennedy CC, Papaioannou A, Ioannidis G, Leslie WD, Walker V. Treating osteoporosis in Canada: what clinical efficacy data should be considered by policy decision makers? Osteoporos Int 2009;20:1785-93. 6. Clarke BL. New and emerging treatments for osteoporosis. Clin Endocrinol (Oxf) 2009;71:309-21. 7. Melo MD, Obeid G. Osteonecrosis of the jaws in patients with a history of receiving bisphosphonate therapy: strategies for prevention and early recognition. J Am Dent Assoc 2005;136:1675-81. 8. Odvina CV, Zerwekh JE, Rao DS, Maalouf N, Gottschalk FA, Pak CY. Severely suppressed bone turnover: a potential complication of alendronate therapy. J Clin Endocrinol Metab 2005;90: 1294-301. 9. Whyte MP, Wenkert D, Clements KL, McAlister WH, Mumm S. Bisphosphonate-induced osteopetrosis. N Engl J Med 2003; 349:457-63.
C. von Keyserlingk et al. 10. McClung MR, Lewiecki EM, Cohen SB, Bolognese MA, Woodson GC, Moffett AH, et al. Denosumab in postmenopausal women with low bone mineral density. N Engl J Med 2006; 354:821-31. 11. National Horizon Scanning Centre (NHSC). (2008) Denosumab (AMG 162) for the prevention of therapy-induced bone loss in non-metastatic breast and prostate cancer. National Horizon Centre. NHS National Institute for Health Research. University of Birmingham. http://www.haps.bham.ac.uk/publichealth/ horizon/outputs/documents/2008/sept-dec/Denosumab_boneloss_ breastandprostate_nonconfSep08.pdf. Accessed Jan 21 2010. 12. Burkiewicz JS, Scarpace SL, Bruce SP. Denosumab in osteoporosis and oncology. Ann Pharmacother 2009;43:1445-55. 13. Bone HG, Bolognese MA, Yuen CK, Kendler DL, Wang H, Liu Y, et al. Effects of denosumab on bone mineral density and bone turnover in postmenopausal women. J Clin Endocrinol Metab 2008;93:2149-57. 14. Ellis GK, Bone HG, Chlebowski R, Paul D, Spadafora S, Smith J, et al. Randomized trial of denosumab in patients receiving adjuvant aromatase inhibitors for nonmetastatic breast cancer. J Clin Oncol 2008;26:4875-82. 15. Lewiecki EM, Miller PD, McClung MR, Cohen SB, Bolognese MA, Liu Y, et al. Two-year treatment with denosumab (AMG 162) in a randomized phase 2 study of postmenopausal women with low BMD. J Bone Miner Res 2007;22:1832-41. 16. Cummings SR, San MJ, McClung MR, Siris ES, Eastell R, Reid IR, et al. Denosumab for prevention of fractures in postmenopausal women with osteoporosis. N Engl J Med 2009;361:756-65. 17. Kong YY, Yoshida H, Sarosi I, Tan HL, Timms E, Capparelli C, et al. OPGL is a key regulator of osteoclastogenesis, lymphocyte development and lymph-node organogenesis. Nature 1999;397: 315-23. 18. Wong BR, Josien R, Lee SY, Sauter B, Li HL, Steinman RM, et al. TRANCE (tumor necrosis factor [TNF]-related activation-induced cytokine), a new TNF family member predominantly ex-
185
19.
20.
21.
22.
23. 24.
25.
26.
27.
pressed in T cells, is a dendritic cell-specific survival factor. J Exp Med 1997;186:2075-80. Miller PD, Barlas S, Brenneman SK, Abbott TA, Chen YT, Barrett-Connor E, et al. An approach to identifying osteopenic women at increased short-term risk of fracture. Arch Intern Med 2004;164:1113-20. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17: 1-12. Altman DG, Schulz KF, Moher D, Egger M, Davidoff F, Elbourne D, et al. The revised CONSORT statement for reporting randomized trials: explanation and elaboration. Ann Intern Med 2001;134:663-94. Higgins JPT GS. The Cochrane Collaboration (2009) Cochrane Handbook for Systematic Reviews of Interventions Version 5.0.2. http://www.cochrane.org/resources/handbook/. Accessed 10 Jan 2010. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557-60. Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Quality of Reporting of Meta-analyses Lancet 1999;354:1896-900. Anastasilakis AD, Toulis KA, Goulis DG, Polyzos SA, Delaroudis S, Giomisi A, et al. Efficacy and safety of denosumab in postmenopausal women with low bone mineral density or osteoporosis: a systematic review and a meta-analysis. Horm Metab Res 2009;41: 721-9. Anastasilakis AD, Toulis KA, Polyzos SA, Terpos E. RANKL inhibition for the management of patients with benign metabolic bone disorders. Exp Opin Investig Drugs 2009;18:1085102. De Laet CE, Pols HA. Fractures in the elderly: epidemiology and demography. Baillieres Best Pract Res Clin Endocrinol Metab 2000;14:171-9.
186
Denosumab for women with low bone mass
Appendix 1 Literature Search Strategy Overview Interface: Databases:
Date of search: Study types: Limits: Syntax guide / .sh MeSH Fs exp * # ? ADJ ADJ# .ti .ab .hw .pt .rn No.
Ovid EMBASE ⬍1996 to 2009, week 53⬎ Ovid Medline ⬍1996 to December week 5, 2009⬎ Ovid Medline in-process and other nonindexed citations ⬍January 7, 2010⬎ Note: Subject headings have been customized for each database. Duplicates between databases were removed in Ovid. January 8, 2010 Randomized controlled trials Publication years 2009 to January 2010 At the end of a phrase, searches the phrase as a subject heading At the end of a phrase, searches the phrase as a subject heading Medical subject heading Floating subheading Explode a subject heading Before a word, indicates that the marked subject heading is a primary topic; or, after a word, a truncation symbol (wildcard) to retrieve plurals or varying endings Truncation symbol for 1 character Truncation symbol for 1 or no characters only Requires words are adjacent to each other (in any order) Adjacency within # number of words (in any order) Title Abstract Heading word; usually includes subject headings and controlled vocabulary Publication type CAS registry number Searches
Results
Clinical multidatabase strategy 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Denosumab exp denosumab/ denosumab.ti,ab. AMG 162 or 615258 to 40-7.ti,ab.rn. Osteoporosis exp osteoporosis/ osteoporo*.ti,ab. 1 or 2 or 3 4 or 5 6 and 7 Date limit limit 8 to yr ⫽ “2009-Current” Randomized controlled trial filter Randomized controlled trials as topic/ Randomized controlled trial/ (Randomized controlled trial or controlled clinical trial).pt. Randomization/ Double-blind procedure/ Double-blind studies/ Single-blind procedure/ Single-blind studies/ Placebos/or placebo/ (random* or sham or placebo*).ti,ab,hw. ((singl* or doubl*) adj (blind* or dumm* or mask*)).ti,ab,hw. ((tripl* or trebl*) adj (blind* or dumm* or mask*)).ti,ab,hw. or/10 to 21 8 and 22 Date limit Limit 23 to yr ⫽ “2009-Current” From 24 keep 1 to 53
470 265 162 58,185 53,123 637 73,830 353 110 203,369 332,000 217,635 56,627 56,107 114,704 7879 18,648 107,423 925,981 159,499 292 966,243 157 53 53
Other databases Cochrane Library, CENTRAL, Issue 4, 2009 Same MeSH, keywords, and date limits used as per Medline search, excluding study types and human restrictions. Syntax adjusted for Cochrane Library databases.