Anabolic Therapies for Osteoporosis in Postmenopausal Women: Effectiveness and Value Final Evidence Report July 17, 2017 Prepared for Note: When our process began, ICER expected FDA approval of two new anabolic agents for osteoporosis in the first half of 2017. On May 21, 2017, Amgen and UCB issued a press release with topline results from the ARCH trial of romosozumab. Among the findings summarized was a new safety signal regarding serious cardiovascular adverse events. Amgen has agreed with the FDA that the ARCH data should be considered in the regulatory review prior to the initial marketing authorization, and as a result the company does not expect approval of romosozumab in the US to occur in 2017. Due to this delay, we have removed romosozumab from our network meta-analysis and our economic modeling and did not consider any voting questions that included romosozumab. However, we have elected to retain the summary of the romosozumab trial results as well as the newly available summary results of the ARCH trial because they provide important contextual information to frame the larger discussion of the role of anabolic therapies in preventing osteoporotic fractures. ©Institute for Clinical and Economic Review, 2017 University of Washington School of Pharmacy ICER Staff and Consultants Modeling Group Jeffrey A. Tice, MD Lotte Steuten, MSc, PhD Professor of Medicine Associate Professor University of California, San Francisco Pharmaceutical Outcomes Research and Policy Program, Department of Pharmacy Rick Chapman, PhD, MS University of Washington Director of Health Economics Institute for Clinical and Economic Review Gregory F. Guzauskas, MSPH, PhD Senior Research Scientist Varun Kumar, MBBS, MPH, MSc Pharmaceutical Outcomes Research and Policy Health Economist Program, Department of Pharmacy Institute for Clinical and Economic Review University of Washington Patricia Synnott, MALD, MS David L. Veenstra, PharmD, PhD Senior Research Associate Professor and Associate Director Institute for Clinical and Economic Review Pharmaceutical Outcomes Research and Policy Program, Department of Pharmacy Matt Seidner, BS University of Washington Program Manager Institute for Clinical and Economic Review The role of the University of Washington (UW) School of Daniel A. Ollendorf, PhD Pharmacy Modeling Group is limited to the Chief Scientific Officer development of the cost-effectiveness model, and the Institute for Clinical and Economic Review resulting ICER reports do not necessarily represent the views of the UW. David M. Rind, MD, MSc Chief Medical Officer Institute for Clinical and Economic Review Steven D. Pearson, MD, MSc President Institute for Clinical and Economic Review DATE OF PUBLICATION: July 17, 2017 We would also like to thank Sonya Khan, Molly Morgan, and Noah Mwandha for their contributions to this report. ©Institute for Clinical and Economic Review, 2017 Page i Final Evidence Report – Anabolic Therapies for Osteoporosis About ICER The Institute for Clinical and Economic Review (ICER) is an independent non-profit research organization that evaluates medical evidence and convenes public deliberative bodies to help stakeholders interpret and apply evidence to improve patient outcomes and control costs. ICER receives funding from government grants, non-profit foundations, health plans, provider groups, and health industry manufacturers. For a complete list of funders, visit http://www.icer- review.org/about/support/. Through all its work, ICER seeks to help create a future in which collaborative efforts to move evidence into action provide the foundation for a more effective, efficient, and just health care system. More information about ICER is available at http://www.icer- review.org About CTAF The California Technology Assessment Forum (CTAF) – a core program of ICER – provides a public venue in which the evidence on the effectiveness and value of health care services can be discussed with the input of all stakeholders. CTAF seeks to help patients, clinicians, insurers, and policymakers interpret and use evidence to improve the quality and value of health care. The CTAF Panel is an independent committee of medical evidence experts from across California, with a mix of practicing clinicians, methodologists, and leaders in patient engagement and advocacy. All Panel members meet strict conflict of interest guidelines and are convened to discuss the evidence summarized in ICER reports and vote on the comparative clinical effectiveness and value of medical interventions. More information about CTAF is available at https://icer- review.org/programs/ctaf/. ©Institute for Clinical and Economic Review, 2017 Page ii Final Evidence Report – Anabolic Therapies for Osteoporosis In the development of this report, ICER’s researchers consulted with several clinical experts, patients, manufacturers and other stakeholders. The following clinical experts provided input that helped guide the ICER team as we shaped our scope and report. None of these individuals is responsible for the final contents of this report or should be assumed to support any part of this report, which is solely the work of the ICER team and its affiliated researchers. For a complete list of stakeholders from whom we requested input, please visit: https://icer-review.org/material/osteoporosis-stakeholder-list/ Expert Reviewers Douglas Bauer, MD Professor of Medicine University of California, San Francisco School of Medicine No relevant conflicts of interest to disclose, defined as more than $10,000 in healthcare company stock or more than $5,000 in honoraria or consultancies during the previous year from health care manufacturers or insurers. Teresa Fama, MD Rheumatologist University of Vermont Health Network, Central Vermont Medical Center No relevant conflicts of interest to disclose, defined as more than $10,000 in healthcare company stock or more than $5,000 in honoraria or consultancies during the previous year from health care manufacturers or insurers. Anna N. A. Tosteson, ScD James J Carroll Professor Professor of Medicine and of The Dartmouth Institute for Health Policy and Clinical Practice Director, Multidisciplinary Clinical Research Program in Musculoskeletal Diseases Director, Comparative Effectiveness Research Program. The Dartmouth Institute for Health Policy and Clinical Practice Geisel School of Medicine at Dartmouth No relevant conflicts of interest to disclose, defined as more than $10,000 in healthcare company stock or more than $5,000 in honoraria or consultancies during the previous year from health care manufacturers or insurers. ©Institute for Clinical and Economic Review, 2017 Page iii Final Evidence Report – Anabolic Therapies for Osteoporosis Table of Contents Executive Summary................................................................................................................................... ES1 1. Background ............................................................................................................................................... 1 1.1 Introduction ........................................................................................................................................ 1 2. The Topic in Context ................................................................................................................................. 4 3. Summary of Coverage Policies and Clinical Guidelines ............................................................................ 9 3.1 Coverage Policies ................................................................................................................................ 9 3.2 Clinical Guidelines ............................................................................................................................. 12 4. Comparative Clinical Effectiveness ......................................................................................................... 16 4.1 Overview ........................................................................................................................................... 16 4.2 Methods ............................................................................................................................................ 16 4.3 Results ............................................................................................................................................... 19 5. Other Benefits or Disadvantages ............................................................................................................ 34 6. Economic Analyses .................................................................................................................................. 36 6.1 Long-Term Cost-Effectiveness .......................................................................................................... 36 6.2 Value-Based Benchmark Prices......................................................................................................... 54 6.3 Potential Budget Impact ................................................................................................................... 55 6.4 Summary and Comment: Long-Term Cost Effectiveness and Potential Budget Impact .................. 58 7. Summary of the Votes and Considerations for Policy ............................................................................ 60 7.1 About the CTAF Process .................................................................................................................... 60 7.2 Voting Results ................................................................................................................................... 62 7.3 Roundtable Discussion and Key Policy Implications ......................................................................... 64 References .................................................................................................................................................. 69 Appendix A. Search Strategies and Results................................................................................................. 77 Appendix B. California Health Exchange and Medicaid Coverage Policies ................................................. 82 Appendix C. Previous Systematic Reviews and Technology Assessments ................................................. 83 Appendix D. Ongoing Studies...................................................................................................................... 85 Appendix E. Comparative Clinical Effectiveness Supplemental Information.............................................. 88 Appendix F. Comparative Value Supplemental Information .................................................................... 100 Appendix G. Public Comments.................................................................................................................. 109 Appendix H. Conflict of Interest Disclosures............................................................................................. 112 ©Institute for Clinical and Economic Review, 2017 Page iv Final Evidence Report – Anabolic Therapies for Osteoporosis List of Acronyms Used in this Report ACE American College of Endocrinology ACP American College of Physicians AACE American Academy of Clinical Endocrinologists AE Adverse event AHRQ Agency for Healthcare Research and Quality ASBMR American Society for Bone and Mineral Research BSCA Blue Shield of California BMD Bone mineral density CI Confidence interval CMS Centers for Medicare and Medicaid Services CrI Credible interval DCHS Department of Health Care Services FDA United States Food and Drug Administration FRAX Fracture Risk Assessment Tool GDP Gross domestic product HEDIS Health Effectiveness Data and Information Set HR Hazard ratio IV Intravenous NAMS North American Menopause Society NMA Network meta-analysis NOF National Osteoporosis Foundation PSA Probabilistic sensitivity analysis PTH Parathyroid hormone PTHrP Parathyroid hormone-related protein QALY Quality-adjusted life-year QM Quantitative morphometry RR Relative risk SAE Serious adverse event SC Subcutaneous SQ Semiquantitative UHC United Healthcare US United States USPSTF United States Preventive Services Task Force WAC Wholesale acquisition cost ©Institute for Clinical and Economic Review, 2017 Page v Final Evidence Report – Anabolic Therapies for Osteoporosis Executive Summary Background Osteoporosis, the weakening of the bones through loss of bone mineral content and a decrease in bone quality, is a common disease of aging that is estimated to affect approximately 10 million Americans (based on bone mineral density [BMD] measurements; this does not take into account additional people who have demonstrated osteoporosis as a result of having a fragility fracture).1 Approximately half of women and one quarter of men will experience at least one fracture due to osteoporosis during their lifetimes.2 Experts estimate that there are approximately two million osteoporotic fractures each year, which results in $19 billion in related costs.3 By 2025, these figures are predicted to grow to approximately three million fractures and $25 billion in costs annually as the population of older Americans increases.3 The goal of treatment is to prevent the fragility fractures associated with osteoporosis: most commonly hip, spine, and wrist fractures. There are two emerging anabolic (i.e., bone-building) therapies for osteoporosis: abaloparatide (Tymlos™, Radius Health, Inc.) and romosozumab (Amgen, Inc. and UCB, Inc.). The only other FDA-approved anabolic agent is teriparatide (Forteo®, Eli Lilly and Co.), which acts through a similar mechanism to abaloparatide. All other agents approved by the United States Food and Drug Administration (FDA) approved agents for osteoporosis are anti-resorptive (i.e., they decrease the breakdown of bone). This assessment will focus on abaloparatide and teriparatide, because the FDA is no longer expected to issue a decision on romosozumab in 2017.4 The Topic in Context Osteoporosis is diagnosed primarily through measurement of bone density at the hip and lumbar spine. Bone density is reported as the number of standard deviations from the bone mass of a young, healthy woman. This is called the T-score. Since humans achieve peak bone mass around the age of 30, the T-score is usually negative. A T-score of -1 or higher is considered normal; a T- score between -1 and -2.5 is considered low bone mass or osteopenia; and a T-score less than -2.5 is considered osteoporosis. The average T-score for a 75-year old white woman is -2.5, so approximately half of white women ages 75 and older have osteoporosis. Osteoporosis is also diagnosed when an individual experiences a fragility fracture in a location associated with osteoporosis (i.e., vertebra, hip). A fragility fracture is a fracture from a low-energy injury that would not normally be expected to result in a broken bone, such as a fall from standing height or less. The most common fractures associated with osteoporosis are vertebral (27%), wrist (19%), hip (14%), and pelvic (7%).3 ©Institute for Clinical and Economic Review, 2017 Page ES1 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Many organizations have treatment guidelines for osteoporosis.5-9 There is general agreement that treatment is indicated for patients over age 50 who have experienced a hip or vertebral fracture or have a bone density T-score less than or equal to -2.5. Treatment may also be indicated for patients with a T-score from -1 to -2.5 and a 10-year probability of hip fracture ≥ 3% or a 10-year probability of a major osteoporotic fracture ≥ 20%. For most patients, first-line therapy is to ensure adequate vitamin D and calcium intake, weight bearing exercise, and an oral medication from the bisphosphonate class of drugs. If patients are unable to tolerate oral bisphosphonates or compliance cannot be ascertained, then IV bisphosphonates are generally recommended. Bone is constantly broken down (resorption) and rebuilt; bisphosphonates work by decreasing bone resorption. There are several other drugs approved for osteoporosis that also decrease bone resorption (estrogen, calcitonin, raloxifene, denosumab). They are not considered first-line therapies because of side effects, less evidence of efficacy, route of administration, and/or cost. Osteoporotic fractures can lead to pain, disability, and death. Even vertebral fractures that do not come to clinical attention may result in loss of height and pronounced curving of the spine (kyphosis) that interferes with activities and make breathing difficult. Patients have become increasingly concerned about two adverse events associated with use of bisphosphonate therapy: osteonecrosis of the jaw and atypical femoral fractures. These concerns may partially explain the 50% decrease in the use of bisphosphonate therapy from 2008 to 2012 in the US.10 Practitioners and clinical societies have noted that rates of osteonecrosis of the jaw and atypical femoral fractures in treated patients are much lower than rates of hip fractures in untreated individuals, and that the overall benefit of treatment is far greater than the harm.6 Adherence with bisphosphonate therapy is a major concern. The oral bisphosphonates must be taken with water on an empty stomach in the morning and then the patient needs to remain upright for at least 30 minutes without consuming any additional food or medications. Observational studies in the real world estimate that only 45% of patients remain adherent with oral bisphosphonate therapy one year after the initial prescription and only 30% after two years.11 The long-acting bisphosphonate, zoledronic acid, which requires only one IV infusion each year may have greater adherence, but some studies report greater than 50% discontinuation of therapy with zoledronic acid by two years.12 This appears to be a problem across classes of parenteral agents for osteoporosis with discontinuation rates at one year of 49% for denosumab, 59% for zoledronic acid, and 67% for teriparatide.12 Given the poor adherence to currently available therapies, new treatments are needed. Individuals on currently-approved drugs continue to experience fragility fractures, so many may benefit from drugs with greater efficacy and acceptable side-effect profiles. ©Institute for Clinical and Economic Review, 2017 Page ES2 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Anabolic or Bone-Building Agents Parathyroid Hormone (PTH) and PTH-related Protein (PTHrP) Analog Drugs Teriparatide was the first drug approved by the FDA for the treatment of osteoporosis that works primarily by increasing bone formation rather than decreasing bone resorption. It is indicated for the treatment of postmenopausal women with osteoporosis who are at high risk for fracture, defined as a history of osteoporotic fracture, or multiple risk factors for fracture, or prior unsuccessful treatment with or intolerance to previous osteoporosis therapy, based upon physician assessment.13 Teriparatide requires a daily injection of 20 mcg under the skin and the drug must be kept refrigerated. Patients are supplied with a pen injector that contains 28 daily doses, which translates to approximately 13 pens per year. In rat studies, teriparatide caused bone tumors (osteosarcomas); however, these have not been observed in humans. Due to concerns that prolonged use could cause osteosarcomas, teriparatide is only used for two years. Abaloparatide is a new PTHrP analog, approved by the FDA on 4/28/17.14 It is indicated for the treatment of postmenopausal women with osteoporosis at high risk for fracture; high fracture risk is defined using the same terms as in the teriparatide label.15 Abaloparatide requires a daily injection of 80 mcg under the skin, but does not require refrigeration after the first dose. Abaloparatide is administered by a pen injector containing 30 daily doses, or approximately 12 pens per year. Anti-Sclerostin Antibodies Romosozumab is a monoclonal antibody directed at the protein sclerostin. Sclerostin decreases bone formation, and by blocking sclerostin function, romosozumab increases bone formation and thus builds bone. Romosozumab also appears to have anti-resorptive effects. It is given by subcutaneous injection once monthly and requires refrigeration. It has not yet been approved by the FDA, and a decision in 2017 is no longer anticipated while the FDA reviews data from the ARCH trial that includes an unexpected safety signal regarding serious cardiovascular adverse events.4 Insights Gained from Discussions with Patients and Patient Groups In the NOF’s Bone Health Index Survey in 2016, patients ranked loss of independence (42%) and lost mobility (25%) as their top two concerns.16 The primary concern among caregivers of patients with osteoporosis was that they would not be able to manage the care of their loved one (50%). Other notable findings included 60% of patients who had sustained a fracture reported not being referred for a bone density test, and fewer than half (47%) were prescribed a medication for osteoporosis. Among those prescribed a medication, 38% said that they never took it, primarily because of fears about side effects (79%). More than half of patients (51%) who started a medication stopped taking it because of side effects (53%) or concerns about the risk for side effects (38%). ©Institute for Clinical and Economic Review, 2017 Page ES3 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Patient groups told us that clinical trials rarely report the outcomes that are most meaningful to patients. These include living independently, the ability to perform the activities of daily living, social engagement, quality of life, reduced fear and anxiety about the disease and treatment, and safety from adverse drug effects. Other outcomes include pain, mobility, depression, and caregiver burden. The details of taking the medication are also important. Medications that require refrigeration (teriparatide, romosozumab) may be particularly burdensome. Many patients have a fear of needles, which is another barrier to adherence with all of the anabolic therapies. There are also insurance barriers to treatment. One patient noted that “health care today is so confusing with copay and coinsurance that I never know what is the right way to go.” Patients also note that insurance often requires that they fail an oral therapy before authorizing an injectable therapy. This adds administrative burden on clinicians, and extra office visits for patients. Comparative Clinical Effectiveness For each of the three anabolic drugs, there is only one pivotal trial.17-19 Each of the trials is good quality and is described in detail in the full report. The pivotal trial of abaloparatide also included open-label teriparatide. We performed a network meta-analysis to compare each drug to the others and included zoledronic acid in the NMA based on feedback from multiple stakeholders. Guidelines recommend zoledronic acid for patients at high risk for fracture and it is a commonly used parenteral therapy for patients with osteoporosis who are unable to tolerate oral therapy. We performed NMAs for morphometric vertebral fractures (i.e., fractures identified by radiographic assessment of paired x-rays of the spine obtained before randomization and at the end of the trial) and non-vertebral fractures. There were insufficient data to evaluate hip fractures in an NMA. As noted above, we excluded romosozumab from all comparative effectiveness analyses including the NMAs. The studies all enrolled postmenopausal women at high risk of fracture. However, the inclusion criteria were different and there were baseline differences in the percentage of participants within each trial with vertebral fractures at baseline (Table ES1 below). If there was significant effect modification for any of the drugs by vertebral fracture status (for example, if one of the drugs was more effective among women with prior fracture than no prior fracture), then it would be inappropriate to combine the studies in a network meta-analysis. However, investigators have specifically analyzed the clinical trials for each of the drugs for effect modification by prior vertebral fracture and other baseline measure and have not identified significant effect modification. ©Institute for Clinical and Economic Review, 2017 Page ES4 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table ES1. Summary of the Randomized Trials of Anabolic Agents and Zoledronic Acid for Osteoporosis F/U, Age, BMI, Prior Reference Study Group N months years kg/m2 Fracture Neer 200119 Fracture Teriparatide 541 21 69 26.8 100% V Prevention Trial Placebo 544 21 69 26.7 Miller 201618 ACTIVE Abaloparatide 824 18 69 25.0 24% V Teriparatide 818 18 69 25.2 63% any Placebo 821 25.1 Cosman FRAME Romosozumab 3589 12 71 24.7 18% V 201617 Placebo 3591 12 71 24.7 22% non-V Black 200720 HORIZON Zoledronic acid 3889 36 73 25.1 63% V Placebo 3876 36 73 25.4 Non-V: non-vertebral fracture, V: vertebral fracture Clinical Benefits Morphometric Vertebral Fractures The pivotal trials of teriparatide, abaloparatide, and zoledronic acid all reported a significant reduction in vertebral fractures, though the definition of incident vertebral fractures differs somewhat between trials.18-20 The results of the NMA confirmed this finding (Table ES2 below). For teriparatide, we used data from a re-analysis conducted by Prevhral et al in 2009, 21 eight years after the original trial. In the Prevhral analysis, the investigators assessed vertebral fractures using an approach that was similar to the one taken in the trials of abaloparatide and zoledronic acid (see full report for additional details); this definition resulted in a lower incidence of new vertebral fractures than in the original paper. All three drugs were significantly better than placebo at reducing morphometric vertebral fractures. Neither of the two anabolic agents was significantly different from one another, nor were they significantly different from zoledronic acid; the credible intervals for the three comparisons of the active drugs all contain 1. ©Institute for Clinical and Economic Review, 2017 Page ES5 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table ES2. Network Meta-Analysis Results for the Relative Risk of Morphometric Vertebral Fractures* Abaloparatide (80 mcg) 0.76 Teriparatide (0.20 – 2.26) (20 mcg) 0.44 0.57 Zoledronic Acid (0.12 – 1.15) (0.30 – 1.02) (5 mg) 0.13 0.17 0.30 Placebo (0.03 – 0.33) (0.09 – 0.29) (0.24 – 0.37) Legend: Each box represents the estimated rate ratio and 95% credible interval for the combined direct and indirect comparisons between two drugs: the drug at the top of the column compared to the drug at the right of the row. Estimates in bold signify that the 95% credible interval does not contain 1. As expected, the NMA estimates for the reduction in vertebral fractures for each drug versus placebo are similar to the direct estimates versus placebo in the randomized trials (Table ES3 below). Table ES3. Comparison of the Relative Risk Versus Placebo for Morphometric Vertebral Fractures Between the Network Meta-Analysis and the Randomized Controlled Trials Drug NMA Estimate RCT Estimates Abaloparatide 0.13 (0.03-0.33) 0.14 (0.05-0.39) Teriparatide 0.17 (0.09-0.29) 0.16 (0.08-0.33) 0.20 (0.08-0.47) Zoledronic acid 0.30 (0.24-0.37) 0.30 (0.24-0.38) Non-Vertebral Fragility Fractures In the key randomized trials, both teriparatide and abaloparatide significantly reduced non- vertebral fractures. The results of the NMA confirmed this finding (Table ES4). Again, neither of the anabolic agents were significantly different from one another, nor were they significantly different from zoledronic acid. ©Institute for Clinical and Economic Review, 2017 Page ES6 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table ES4. Network Meta-Analysis Results for the Relative Risk of Non-Vertebral Fragility Fractures* Abaloparatide (80 mcg) 0.83 Teriparatide (0.46 – 1.46) (20 mcg) 0.69 0.82 Zoledronic Acid (0.38 – 1.16) (0.54 – 1.22) (5 mg) 0.51 0.61 0.75 Placebo (0.28 – 0.85) (0.41 – 0.88) (0.64 – 0.87) Legend: Each box represents the estimated rate ratio and 95% credible interval for the combined direct and indirect comparisons between two drugs: the drug at the top of the column compared to the drug at the right of the row. Estimates in bold signify that the 95% credible interval does not contain 1. As with vertebral fractures, the NMA estimates for the reduction in non-vertebral fractures for each drug versus placebo are similar to the direct estimates versus placebo in the randomized trials (Table ES5 below). Table ES5. Comparison of the Relative Risk Versus Placebo for Non-Vertebral Fractures Between the Network Meta-Analysis and the Randomized Controlled Trials Drug NMA Estimate RCT Estimates Abaloparatide 0.51 (0.28-0.85) 0.57 (0.32-1.00) Teriparatide 0.61 (0.41-0.88) 0.47 (0.25-0.88) 0.72 (0.42-1.22) Zoledronic acid 0.75 (0.64-0.87) 0.75 (0.64-0.87) Hip Fractures Hip fractures are an important sub-type of non-vertebral fractures because they are associated with loss of independence and mortality in additional to significant short-term morbidity and costs. The randomized trials of abaloparatide and teriparatide did not have sufficient power to demonstrate a reduction in hip fractures, but one observational data reported a 45% reduction in hip fractures for teriparatide.22 The randomized trial of zoledronic acid found a 41% reduction in hip fractures compared to placebo (RR 0.59, 95% CI 0.42-0.83). Harms In the pivotal trials, there were no significant differences in serious adverse events or discontinuation due to adverse events between the active treatments and placebo. Both abaloparatide and teriparatide are associated with injection site reactions and hypercalcemia. Rats developed osteosarcomas during treatment with abaloparatide and teriparatide, but this has not ©Institute for Clinical and Economic Review, 2017 Page ES7 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents been observed in humans. However, the treatment duration for the two drugs is limited to two years due to concerns that prolonged use could cause osteosarcomas. Zoledronic acid is associated with flu-like infusion reactions in up to 30% of patients following the first treatment. Zoledronic acid has also been associated with rare, but serious atypical femoral fractures and osteonecrosis of the jaw. It is estimated that treatment of 10,000 women with zoledronic acid for 3 years would prevent approximately 710 vertebral fractures, 110 hip fractures, while causing 1 atypical femoral fracture and less than 1 case of osteonecrosis of the jaw. Controversies and Uncertainties The primary controversy is whether it was appropriate to combine the data from the different study populations of the three trials in a NMA. As noted earlier, there is evidence that there is no effect modification for any of the drugs by patient characteristics including prior vertebral fractures and other risk factors for fracture. Thus, it is appropriate to compare the relative effects of the drugs in a network meta-analysis. None of the previously published NMAs of drug therapy for osteoporosis included abaloparatide (see Appendix C).23-27 Similar to our findings, the NMAs concluded that both teriparatide and zoledronic acid reduce the risk of vertebral and non-vertebral fractures compared to placebo. They found no significant differences between the drugs, though teriparatide ranked higher than zoledronic acid. They also concluded that zoledronic acid reduced hip fractures, but there was insufficient evidence for teriparatide. A major area of uncertainty reflects the relative paucity of evidence for each of the anabolic agents, particularly for the hip fracture outcome. The trials were relatively small given the large number of women with osteoporosis. In addition, active treatment continued for only one to two years. We could not model stable estimates for hip fracture reduction because of the low number of events. Indeed, the recent ACP clinical guideline did not recommend any of the anabolic agents as first line therapy for osteoporosis because of the lack of randomized trial evidence on hip fracture prevention.9 Some have suggested that anabolic therapy may have more rapid onset of fracture prevention than antiresorptive therapy. Given the paucity of head-to-head trials, it is difficult to evaluate this hypothesis. However, in the HORIZON trial of zoledronic acid, the reductions in hip fractures, non- vertebral fractures, and any clinical fractures, as assessed by the Kaplan-Meier curves, appeared to begin at randomization. In the ACTIVE trial, abaloparatide appeared to have a more rapid reduction in non-vertebral fractures, clinical fractures, and major osteoporotic fractures than teriparatide, but the differences were not statistically significant except for major osteoporotic fractures (p=0.03). There are insufficient data to assess the relative efficacy of the anabolic agents compared to ©Institute for Clinical and Economic Review, 2017 Page ES8 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents zoledronic acid in the first three to six months of therapy. There are no significant differences in fracture reduction between anabolic therapy and zoledronic acid over longer time periods. Another important area of uncertainty is sequencing of therapies. Studies suggest that the bone density gains from anabolic agents are quickly lost if no follow-up therapy is used.28 Since anabolic agents are only used for one to two years, they will need to be followed by some form of anti- resorptive therapy to maintain the reduction in fracture risk. Other studies have found that the beneficial effects of PTH-related therapies on bone mass are blunted among individuals previously treated with anti-resorptive drugs.29 This suggests that anabolic agents may be most effective if used prior to anti-resorptive therapy. The best agent to use and the optimal length of follow-up treatment is uncertain and awaits additional fracture endpoint studies. The outcomes of greatest interest to patients are maintenance of independence and prevention of disability. These and other patient-centered outcomes were not reported in the pivotal trials. Summary The evidence to date demonstrates with high certainty that the two anabolic agents reduce vertebral fractures compared to no therapy. However, there is insufficient evidence to distinguish the anabolic agents from each other and from zoledronic acid for vertebral fractures. The differences in fracture reduction are small and the credible intervals all contain 1, so the therapies may be comparable. The evidence is even less certain for non-vertebral fragility fractures and, in particular, hip fractures. Including the unpublished data from the VERO study, which compared teriparatide to the oral bisphosphonate risedronate, did not change these conclusions. The harms of therapy are relatively small and have little influence on the net benefit for each therapy compared to the others. Adherence to both initial anabolic therapy and subsequent anti-resorptive therapy is essential to preserve the fracture reduction benefit. However, there are minimal real- world data available to compare adherence to therapy between the two anabolic agents. For the two anabolic agents, we judged the evidence to be promising, but inconclusive (P/I) for the net health benefit when compared to zoledronic acid in postmenopausal women with osteoporosis at high risk for fracture. When compared to no treatment, we judged with moderate certainty that the anabolic agents provided a small or substantial net health benefit compared to no therapy, with high certainty of at least a small net health benefit when compared to no therapy (B+). There is a substantial reduction in vertebral fractures, a small to moderate reduction in non-vertebral fractures, and uncertain benefits for hip fractures, though observational data do support a benefit for teriparatide. When abaloparatide is compared to teriparatide, we judged that there is insufficient evidence to assess the comparative clinical effectiveness of the two drugs because of low certainty in the ©Institute for Clinical and Economic Review, 2017 Page ES9 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents evidence. The extensive real world clinical experience with teriparatide without identification of new adverse events and observational evidence confirming benefits is reassuring. However, in the ACTIVE trial, there was a non-significant trend towards greater reduction in both vertebral and non- vertebral fractures with abaloparatide compared with teriparatide. Other Benefits or Disadvantages There are important differences in the treatments that may be important for some patients and preferences will differ among patients. Abaloparatide and teriparatide require daily injections, which is a barrier to adherence for some patients. The comparator, zoledronic acid requires an annual visit for a 15-minute infusion that can be associated with systemic symptoms, particularly following the first dose. The once-a-year dosing may be an advantage, but the requirement for an intravenous infusion may decrease acceptability. In addition, some patients may have concerns about a drug that remains in the body for a long time. There are no clear differences among the anabolic drugs in terms of impact on caregiver burden, although daily injections may be burdensome if a caregiver is required to perform the injection. Abaloparatide acts through a similar mechanism as teriparatide. However, both anabolic drugs work through a fundamentally different mechanism from the other available agents, including zoledronic acid. There is evidence that starting with an anabolic agent followed by an antiresorptive agent may result in greater long-term fracture prevention than treating with an antiresorptive agent for the same length of time. However, to date, there are no published randomized trials demonstrating that this is the optimal approach. Comparative Value We conducted a cost-effectiveness analysis using a simulation model comparing the anabolic drugs abaloparatide and teriparatide, each followed by treatment with a bisphosphonate (zoledronic acid), versus treatment with zoledronic acid alone in a representative cohort of postmenopausal women who are at high risk for osteoporotic fractures. The target population was 70-year-old postmenopausal women with a fracture incidence similar to that observed in the clinical trials of the anabolic drugs. Annual relative risks of vertebral and other non-vertebral fractures for each drug were obtained from our evidence review’s NMA. Relative risk of hip fractures for the anabolic drugs were estimated using the ratio of hip to non-vertebral fracture relative risks reported in the HORIZON trial (zoledronic acid vs. placebo). Briefly, the HORIZON-derived ratio was 0.59 (hip) / 0.75 (non-vertebral) = 0.79, which was multiplied by the NMA-derived relative risks for non-vertebral fractures (abaloparatide = 0.51, teriparatide = 0.61) to obtain base case estimates. The baseline utility estimates for patients with no new fracture were from a study of the non- institutionalized US adult population for seven health-related quality-of-life scores; we used the ©Institute for Clinical and Economic Review, 2017 Page ES10 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents EuroQoL-5D (EQ-5D) age- stratified estimates for US women.30 We applied utility multipliers to baseline estimates for each fracture health state, with utilities for vertebral fractures applied only to 35% of patients with vertebral fractures, reflecting the proportion of these fractures that were clinical fractures in a retrospective cohort analysis.31 More details on our methods for utility estimation can be found in Section 6 of the report. Other estimates of relevant clinical parameters, and drug and health care costs were obtained from published literature. Several key assumptions were made in the model, and a comprehensive list of model assumptions along with the rationale for each is available in Section 6 of the report: • Patients can only transition to a worse fracture state or death from a post-fracture state, thus preventing patients who experience more serious fractures from forfeiting costs and utilities associated with these states • Patients can have an unlimited number of fractures while in the model • As described above, hip fracture relative risk estimates for anabolic drugs were based on the ratio of hip fracture relative risk versus non-vertebral fracture relative risk reported in the HORIZON trial. • Serious adverse events were not modeled in the base-case analysis, as event rates were similar between each agent and placebo in their respective trials. • Both anabolics were assumed to maintain 100% of their efficacy throughout their duration of administration, the six years of zoledronic acid administration following the anabolics, plus an additional three years, before efficacy declined linearly to a relative risk of one over a ten-year period. Similarly, zoledronic acid was assumed to maintain 100% efficacy of its efficacy during the six years of administration plus an additional three years post-therapy, before this efficacy declined linearly to a relative risk of one over a ten-year period. • We assumed 100% adherence for all agents included in the model. The model tracked vertebral fractures (both morphometric and clinical), hip fractures, other non- vertebral fractures, and death. The model’s base-case analysis adopted a health care system perspective with outcomes modeled over a life-time horizon using 3% discount rates for costs and outcomes. Each anabolic agent was associated with a net price per pen (Table ES6) based on the wholesale acquisition cost (WAC) of $2,998 for teriparatide and $1,625 for abaloparatide. We applied a discount of 38% to the teriparatide WAC (resulting in a price of $1,866 per pen), based on data provided by SSR Health combining data on net US dollar sales with information on unit sales to derive net pricing at the unit level across all payers.32 Net pricing data for abaloparatide was unavailable due to the agent’s recent approval, so we assumed a discount of 27% from WAC (resulting in a price of $1,186 per pen), representing the average industry-wide discount for branded drugs.33 Other non-drug costs such as fracture treatment costs and administration costs for ©Institute for Clinical and Economic Review, 2017 Page ES11 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents zoledronic acid were derived from the published literature. All costs were converted to 2016 US dollars. Table ES6. Drug Cost Inputs Base- Acquisition Drug Name, Labeled Dose, Strength WAC/Pen Net Price* Case Tx Cost Per Tx Administration Route (Pen Size) Duration Course† Teriparatide 20 mcg SC QD 250 mcg/ml $2,997.90 $1,866.34‡ 2 years $48,691 (2.4 ml) Abaloparatide 80 mcg SC QD 3,120 $1,625 $1,186.25§ 2 years $29,312 mcg/1.56 ml Zoledronic Acid 5 mg IV Q year 5 mg/100 ml $306 # $306# 6 years $1,837 IV: intravenous, SC: subcutaneous, QD: once daily, Q mo: once monthly, Q year: once yearly, Tx: treatment, WAC: wholesale acquisition cost *Net price is the estimated price after discounts and rebates from WAC. No discounts have been applied to generic zoledronic acid. †Acquisition cost of initial drug using net price (or average generic WAC for zoledronic acid) and assuming full course of treatment; costs would be lower if a modeled patient died before completing a course of therapy. Costs do not include the additional costs of post-anabolic zoledronic acid therapy. ‡Price per pen including 38% discount §Price per pen based on announced list price and assumed 27% discount #Annual dose cost based on average generic WAC Model outputs include total costs, quality-adjusted life years (QALYs), life years for the interventions and comparators, and incremental costs per additional QALY and life year gained for the two interventions versus zoledronic acid. In addition to a base case analysis, sensitivity analyses using ranges of values for model inputs were conducted. Further details of the model structure and assumptions are provided in Chapter 6 of the full report. Base-Case Results The anabolic therapies resulted in increased costs, QALYs, and life years compared to zoledronic acid (Table ES7). QALYs gained versus zoledronic acid ranged from 0.066 for abaloparatide to 0.046 for teriparatide over the lifetime horizon. Incremental costs versus zoledronic acid ranged from a low of $22,061 for abaloparatide to $43,440 for teriparatide. The base case incremental cost- effectiveness ratios (ICERs) for each anabolic drug compared to zoledronic acid far exceeded the commonly-cited cost-effectiveness threshold of $150,000 per QALY (Table ES8). ©Institute for Clinical and Economic Review, 2017 Page ES12 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table ES7. Base-Case Results Regimen Cost QALYs Life Years Zoledronic acid $25,465 8.933 12.188 Teriparatide $68,905 8.979 12.193 Abaloparatide $47,525 8.999 12.195 QALY: quality-adjusted life year Table ES8. Pairwise Results for Anabolic Therapies Compared to Zoledronic Acid Regimen Incr. Cost Incr. QALYs Incr. LYs ICER vs. Zoledronic Acid Teriparatide $43,440 0.046 0.005 $941,537 Abaloparatide $22,061 0.066 0.007 $333,892 ICER: incremental cost-effectiveness ratio, Incr.: incremental, LY: life year, QALY: quality-adjusted life year Sensitivity Analysis Results Detailed findings from the one-way sensitivity analyses varying the model inputs for anabolic agents versus zoledronic acid can be found in Figures ES1 and ES2. Parameters associated with hip fractures were the largest contributors to uncertainty for abaloparatide and teriparatide versus zoledronic acid, particularly the anabolics’ relative risks for hip fracture (the most expensive and severe of the fracture types) as they approached 1.0 (i.e., no efficacy vs. untreated patients). Results were also sensitive to uncertainty in the long-term utility multipliers and drug costs. None of the modeled parameters’ range values resulted in ICERs less than $150,000 per QALY gained. (Negative ICERs shown below result from negative incremental QALYs vs. zoledronic acid.) ©Institute for Clinical and Economic Review, 2017 Page ES13 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Figure ES1. Incremental Cost-Effectiveness Results of One-Way Sensitivity Analyses for Teriparatide Versus Zoledronic Acid -$18M -$15M -$12M -$9M -$6M -$3M $0 $3M Parameter Low Value High Value Low Result High Result Spread Teriparatide Relative Risk: Hip Fracture 0.280 0.750 $479,590 -$16,532,705 $17,012,296 Zoledronic Acid Relative Risk: Hip Fracture 0.420 0.830 $2,384,332 $482,093 $1,902,239 Utility Multiplier: Hip Fracture Year 2+ 0.640 0.960 $700,502 $1,435,465 $734,963 Cost: Teriparatide 600 mcg/2.4mL pen $1,493 $2,240 $737,109 $1,145,966 $408,857 Teriparatide Relative Risk: Vertebral Fracture 0.090 0.290 $833,277 $1,165,710 $332,433 Utility Multiplier: Other Fracture Year 2+ 0.800 1.000 $637,548 $941,537 $303,990 Teriparatide Relative Risk: Other Fracture 0.410 0.880 $840,319 $1,110,995 $270,675 General Population Utility: Age 80+ 0.576 0.864 $1,081,405 $833,707 $247,698 Utility Multiplier: Vertebral Fracture Year 2+ 0.745 1.000 $781,722 $1,018,714 $236,992 Utility Multiplier: Other Fracture Year 1 0.722 1.000 $838,241 $1,009,089 $170,848 Figure ES2. Incremental Cost-Effectiveness Results of One-Way Sensitivity Analyses for Abaloparatide Versus Zoledronic Acid $0 $2M $4M $6M $8M Parameter Low Value High Value Low Result High Result Spread Abaloparatide Relative Risk: Hip Fracture 0.171 0.741 $165,144 $7,807,367 $7,642,223 Zoledronic Acid Relative Risk: Hip Fracture 0.420 0.830 $624,585 $181,708 $442,877 Utility Multiplier: Hip Fracture Year 2+ 0.640 0.960 $243,159 $532,644 $289,485 Cost/pen: Abaloparatide $949 $1,424 $250,425 $417,358 $166,933 Abaloparatide Relative Risk: Vertebral Fracture 0.030 0.330 $298,026 $435,752 $137,726 Utility Multiplier: Other Fracture Year 2+ 0.800 1.000 $197,820 $333,892 $136,072 Abaloparatide Relative Risk: Other Fracture 0.280 0.850 $293,843 $405,122 $111,279 General Population Utility: Age 80+ 0.576 0.864 $381,325 $296,954 $84,371 Utility Multiplier: Other Fracture Year 1 0.722 1.000 $289,681 $364,076 $74,395 Utility Multiplier: Vertebral Fracture Year 2+ 0.745 1.000 $284,412 $356,900 $72,488 ©Institute for Clinical and Economic Review, 2017 Page ES14 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Probabilistic sensitivity analysis indicated that our ICER results are highly uncertain, but the probability that the ICERs for the anabolic therapies were below $150,000 per QALY gained were either low (abaloparatide: 7.1%) or zero (teriparatide) (Figure ES3). This was primarily due to the small QALY gains and higher prices of anabolics versus zoledronic acid. Figure ES3. Cost-Effectiveness Acceptability Curve for Anabolic Agents Compared to Zoledronic Acid In a scenario analysis of patient populations with a higher risk of fracture than in the base-case, fracture risks would need to be approximately 118% higher for abaloparatide to approach the $150,000 per QALY threshold, or the approximate risk of an 85-year-old woman with a T-score of -4. Teriparatide did not approach commonly-cited cost-effectiveness thresholds until a >1000% increased risk of fracture was applied. We also considered a scenario in which patients may not be able to take zoledronic acid. In this scenario, incremental QALYs decreased due to the shortened efficacy time window for the anabolics, and none of the treatments reached the $150,000 per QALY threshold. Increasing refracture risks resulted in modest improvements in incremental QALYs and cost; however, none of these improvements were sufficient to make the incremental cost- effectiveness ratios for anabolic agents fall below $150,000 per QALY. ©Institute for Clinical and Economic Review, 2017 Page ES15 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Threshold Analysis Results Prices for each drug that would achieve commonly-cited cost-effectiveness thresholds ranging from $50,000 to $150,000 per QALY gained are presented in Table ES9, along with net price per pen (i.e., base-case cost). Table ES9. Resulting Pen Prices for Each Anabolic Therapy to Reach Cost per QALY Thresholds Drug Base-Case Cost $50,000/QALY $100,000/QALY $150,000/QALY Teriparatide $1,866.34 $238.47 $329.77 $421.07 (cost per pen) Abaloparatide $1,186.25 $379.30 $521.42 $663.55 (cost per pen) QALY: quality-adjusted life year Model Validation Model validation followed standard practices in the field. Three independent modelers stress- tested the model using a range of inputs, including null inputs, to check for model stability. As part of the validation process, we also compared our model to previously published osteoporosis models. One model by Tosteson et al. was structurally similar to our own.34 However, incremental cost-effectiveness ratios were greater in the ICER model compared to the Tosteson model due to differences in several model characteristics and inputs. For example, the annual cost of teriparatide in the 2008 Tosteson paper was approximately $6,300, versus approximately $24,350 in the ICER analysis. The ICER analysis also used a lifetime time horizon, while Tosteson used a 10-year time horizon. When using 10-year time horizon, the ICER model produced similar QALY gains to those seen in the Tosteson model. Additional details pertaining to this comparison can be found in Section 6 of the report. Value-based Benchmark Prices Our value-based benchmark prices for abaloparatide and teriparatide are presented in Table ES10. As noted in the initial ICER methods document (http://icer- review.org/wpcontent/uploads/2016/02/Value-Assessment-Framework-slides-for-July-29-webinar- FINALcorrected-8-22-1.pdf), the value-based benchmark price for a drug is defined as the price range that would achieve cost-effectiveness ratios between $100,000 and $150,000 per QALY gained. For both abaloparatide and teriparatide, the discounts required to meet both threshold prices are greater than the current discounts from WAC (assumed 27% for abaloparatide, 38% for teriparatide). ©Institute for Clinical and Economic Review, 2017 Page ES16 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table ES10. Value-based Benchmark Prices for Abaloparatide and Teriparatide for Osteoporosis Treatment Discount from WAC Average Net Annual Net Price* Cost to Achieve Cost to Achieve to reach $100,000 Price Within Drug Name WAC per per Pen $100,000/QALY $150,000/QALY and $150,000/QALY Benchmark Pen Threshold Range? Teriparatide $2,997.90 $1,866.34 $329.77 ‡ $421.07 86% to 89% No Abaloparatide $1,625.00 $1,186.25† $521.42 $663.55 59% to 68% No QALY: quality-adjusted life year, WAC: wholesale acquisition cost, WTP: willingness to pay *Net price is the estimated price after discounts and rebates from WAC. † Price per pen based on announced list price and assumed 27% discount ‡ Price per pen including 38% discount Potential Budget Impact We used results from the same model employed for the cost-effectiveness analyses to estimate total potential budget impact of abaloparatide, calculating incremental health care costs (including drug costs) minus any offsets in these costs from averted health care events. We did not estimate the budget impact of teriparatide, given its established presence in the market. The potential budget impact was defined as the total incremental net cost of using abaloparatide, taking market share from teriparatide and zoledronic acid in the ratio 80:20 over a five-year time horizon. The potential budget impact analysis included the entire candidate population for treatment, which consisted of postmenopausal women (assumed to be women over 50 years of age) diagnosed with osteoporosis and with a high risk of fractures. To estimate the size of the potential candidate population for treatment with abaloparatide, we first determined the number of women over 50 years of age in the US, approximately 62.6 million. Of those women, we assumed that 13% currently receive treatments for osteoporosis, based on a claims database analysis by Parthan et al., conducted to identify this percentage for a published budgetary impact analysis of denosumab in a hypothetical health plan.35 Of those receiving treatment, 66% were diagnosed with osteoporosis while the remaining were treated for osteopenia.35 We assumed that 46% of those women diagnosed and treated for osteoporosis had a high risk of osteoporotic fractures, based on occurrence of previous fractures and/or intolerance to previous osteoporosis treatment.35 This high-risk population was assumed to be eligible to receive treatment with abaloparatide. Applying these estimates to the projected 2017 US population resulted in an estimate of approximately 2.47 million eligible patients in the US. Table ES11 below illustrates the per-patient budget impact results in more detail. Costs for abaloparatide were calculated using the WAC, discounted WAC, and three threshold prices. The discounted WAC price of teriparatide was derived from the SSR database, and average WAC price for generic zoledronic acid was used to calculate costs for those treatments. ©Institute for Clinical and Economic Review, 2017 Page ES17 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents When treating the eligible cohort with abaloparatide, the average potential budgetary impact (adjusted for differing periods of drug utilization and associated cost-offsets over the five-year period) resulted in cost-savings using the WAC, discounted WAC and across all three cost- effectiveness thresholds, ranging from approximately -$120 per patient using the WAC price ($1,625), to approximately -$10,500 per patient using the price to achieve $50,000 per QALY ($379). Table ES11. Per-Patient Budget Impact Calculation Over a Five-year Time Horizon Average Annual Per-Patient Budget Impact Discounted WAC $150,000/QALY $100,000/QALY $50,000/QALY WAC Abaloparatide $13,952 $10,290 $5,928 $4,742 $3,556 Teriparatide + Zoledronic acid* $14,072 (Discounted WAC Only) Difference -$120† -$3,782† -$8,144† -$9,330† -$10,516† *Weighted in the ratio 80:20 for teriparatide:zoledronic acid †Indicates cost-saving N/A: not available, QALY: quality-adjusted life year, WAC: wholesale acquisition cost Summary and Comment We estimated the cost-effectiveness of anabolic treatments compared to zoledronic acid in patients with osteoporosis at high risk for fragility fractures. The cost per additional QALY was estimated to be above $150,000 per QALY for each anabolic agent, assuming a 38% and 27% discount on list prices of teriparatide and abaloparatide, respectively. This finding remained robust over a wide range of sensitivity and scenario analyses. These included analyses of patients at even higher risk for fracture, assuming that the benefits of zoledronic acid are delayed, and varying the rate of decline in benefit after treatment is stopped. The results were most sensitive to uncertainty in relative risk estimates for hip fracture, long-term fracture utility multipliers, and drug costs. When the anabolic agents are compared to no treatment, the results suggest that anabolic treatments would not produce incremental cost-effectiveness ratios of less than $150,000 per QALY. Our study has some limitations that are worth noting. First, our model assumes a fracture hierarchy that prevents patients from having a fracture classified as less severe than their last fracture. This likely underestimates the number of less severe fractures, and potentially overestimates impacts of hip fractures, which were the most severe fractures in the hierarchy. We attempted to mitigate the influence of hip fracture by calibrating our base-case hip fracture estimates to reflect those predicted by the FRAX Fracture Assessment Tool. Second, we did not consider adverse events, given that anabolic regimens and zoledronic acid exhibited similar serious adverse event rates compared to placebo and to each other in their respective trials. These small event rate differences would have minimal impact on the results. Third, we assumed 100% adherence to all treatments, ©Institute for Clinical and Economic Review, 2017 Page ES18 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents which would not occur in actual practice. Finally, our base-case cost and cost-effectiveness results for anabolics reflect our current assumptions about drug price. Despite this, one-way sensitivity analysis showed that drug prices were much less influential on results than differences in fracture prevention efficacy, and we provided threshold analysis results to offer insight into the drug prices that would make each agent cost-effective under traditional thresholds. Finally, budget impact analysis for abaloparatide indicates that use in place of teriparatide and zoledronic acid is not likely to generate access or affordability alerts when using WAC, discounted WAC, or the prices to achieve cost-effectiveness thresholds of $150,000 per QALY or lower. California Technology Assessment Forum Votes The California Technology Assessment Forum (CTAF) deliberated on key questions raised by ICER’s report at a public meeting on June 30 in Los Angeles, California. The results of these votes are presented below, and additional information on the deliberation surrounding the votes can be found in the full report. 1) For postmenopausal women with osteoporosis and a high risk of fracture, is the evidence adequate to demonstrate that the net health benefit of treatment with teriparatide (Forteo®, Eli Lilly and Co.), is greater than that of treatment with zoledronic acid? Yes: 2 votes No: 13 votes 2) For postmenopausal women with osteoporosis and a high risk of fracture, is the evidence adequate to demonstrate that the net health benefit of treatment with abaloparatide (Tymlos™, Radius Health Inc.), is greater than that of treatment with zoledronic acid? Yes: 2 votes No: 13 votes 3) For postmenopausal women with osteoporosis and a high risk of fracture, is the evidence adequate to distinguish between the net health benefit of teriparatide and abaloparatide? Yes: 2 votes No: 13 votes 4) Given the available evidence on comparative effectiveness and incremental cost-effectiveness, and considering other benefits, disadvantages, and contextual considerations, what is the long-term value for money of treatment with teriparatide followed by zoledronic acid versus treatment with zoledronic acid alone for postmenopausal women with osteoporosis at high risk for fracture? Low: 13 votes Intermediate: 2 votes High: 0 votes ©Institute for Clinical and Economic Review, 2017 Page ES19 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 5) Given the available evidence on comparative effectiveness and incremental cost-effectiveness, and considering other benefits, disadvantages, and contextual considerations, what is the long-term value for money of treatment with abaloparatide followed by zoledronic acid versus treatment with zoledronic acid alone for postmenopausal women with osteoporosis at high risk for fracture? Low: 13 votes Intermediate: 2 votes High: 0 votes Key Policy Implications Following its deliberation on the evidence, the CTAF Panel engaged in a moderated discussion with a policy roundtable about how best to apply the evidence on anabolic therapies for osteoporosis in postmenopausal women to policy and practice. The policy roundtable members included a patient, three clinical experts, two payers, and two representatives from pharmaceutical manufacturers. The discussion reflected multiple perspectives and opinions, and therefore, none of the statements below should be taken as a consensus view held by all participants. The top-line policy implications are presented below, and additional information can be found in the full report. Manufacturers • Reduce the prices of anabolic agents to align with the clinical benefits they bring to patients • Abstain from direct to consumer advertising and detailing to primary care providers • Include broader patient groups in randomized trials Payers • Given the lack of clinical expert consensus on how to identify patients who would benefit most from consideration of anabolic therapy, design coverage policies with a broad set of criteria by which to determine whether the risk of fracture and the underlying bone pathology would make anabolic therapy a more appropriate first choice than intensive anti- resorptive therapy. • Create a prior authorization process for anabolic therapies that is clear and efficient for providers • If the prices of anabolic agents are reduced, ease access restrictions Patient Advocacy Organizations • Demand the inclusion of patient-centered outcomes in clinical trials • Continue to promote lifestyle changes that protect against osteoporosis ©Institute for Clinical and Economic Review, 2017 Page ES20 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Specialty Societies • Develop clear guidelines for use of anabolic agents Regulators • Promote hip fracture as the most important outcome in pivotal clinical trials • Require that pivotal trials include an active comparator Researchers • Develop better risk assessment tools to identify patients at extreme risk for fracture ©Institute for Clinical and Economic Review, 2017 Page ES21 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 1. Background 1.1 Introduction Background Osteoporosis, the weakening of the bones through loss of bone mineral content and a decrease in bone quality, is a common disease of aging that is estimated to affect approximately 10 million Americans (based on bone mineral density [BMD] measurements; this does not take into account additional people who have demonstrated osteoporosis as a result of having a fragility fracture).1 Approximately half of women and one quarter of men will experience at least one fracture due to osteoporosis during their lifetimes.2 Experts estimate that there are approximately two million osteoporotic fractures each year, which results in $19 billion in related costs.3 By 2025, these figures are predicted to grow to approximately three million fractures and $25 billion in costs annually as the population of older Americans increases.3 The goal of treatment is to prevent the fragility fractures associated with osteoporosis: most commonly hip, spine, and wrist fractures. There are two emerging anabolic (i.e., bone-building) therapies for osteoporosis: abaloparatide (Tymlos™, Radius Health, Inc.) and romosozumab (Amgen, Inc. and UCB, Inc.); romosozumab also decreases bone resorption.36 The only other FDA- approved anabolic agent is teriparatide (Forteo®, Eli Lilly and Co.), which acts through a similar mechanism to abaloparatide. All other agents approved by the United States Food and Drug Administration (FDA) approved agents for osteoporosis are anti-resorptive (i.e., they decrease the breakdown of bone). All three anabolic drugs are delivered via subcutaneous injection. This assessment will focus on abaloparatide and teriparatide, because the FDA is no longer expected to issue a decision on romosozumab in 2017.4 Scope of the Assessment The scope for this assessment is described on the following pages using the PICOTS (Population, Intervention, Comparators, Outcomes, Timing, and Settings) framework. Evidence was abstracted from randomized controlled trials. There was only one head-to-head study of these interventions with fracture outcomes,18 so we included placebo-controlled studies and derived indirect comparisons from a network meta- analysis. Analytic Framework The general analytic framework for assessment of therapies for osteoporosis is depicted in Figure 1. ©Institute for Clinical and Economic Review, 2017 Page 1 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Figure 1. Analytic Framework: Anabolic Therapies for Osteoporosis AE: adverse event, SAE: serious adverse event Populations The population for the review was postmenopausal women with an indication for treatment to prevent osteoporotic fractures, with a focus on high-risk individuals such as those with a prior fragility fracture and a T-score less than -2.5. The primary focus is on women who have not received prior treatment for osteoporosis. Interventions The list of interventions was developed with input from patient organizations, clinicians, manufacturers, and payers on which drugs to include. The full list of interventions is as follows: • Abaloparatide (Tymlos™, Radius Health, Inc.) • Teriparatide (Forteo®, Eli Lilly and Co.) • Romosozumab (Amgen, Inc. and UCB, Inc.) Due to the delay in FDA consideration of romosozumab announced after the release of our draft report, we have limited the evaluation of the drug to the presentation of clinical trial results. We have not included romosozumab in the network meta-analysis or in the cost-effectiveness model, nor have we made any judgements regarding the comparative effectiveness of romosozumab to the other agents. Comparators We compared abaloparatide and teriparatide to each other, to no therapy, and to the intravenous (IV) bisphosphonate zoledronic acid. We selected zoledronic acid as the key bisphosphonate comparator because several osteoporosis guidelines recommend it for individuals at high risk for ©Institute for Clinical and Economic Review, 2017 Page 2 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents fracture and because multiple stakeholders recommended it as the most appropriate comparator. Comparing the agents to zoledronic acid allowed us to evaluate the relative incremental benefits and harms of these agents when used first line in patients at high risk for fragility fractures. Outcomes The primary goal of treatment is to prevent fractures. The most important fracture to prevent is hip fracture because of the associated morbidity and mortality, but these fractures are relatively uncommon. Next in importance are clinical vertebral fractures, which are compression fractures of the spine that cause pain. Finally, non-vertebral fragility fractures were assessed. Changes in BMD, bone turnover markers, and radiographic vertebral fractures will be considered as surrogate outcomes. Where possible we reported the absolute risk reduction and number needed to treat in addition to the relative risk reduction for the treatment comparisons. Clinical Outcomes Key harms Hip fractures Atypical femoral fractures All fragility fractures Osteonecrosis of the jaw Clinical vertebral fractures Osteosarcoma Living independently Significant adverse events Mobility Adverse events leading to discontinuation Pain Injection site reactions Ability to attend to activities of daily living Hypocalcemia/Hypercalcemia Quality of life Non-clinical Outcomes Bone mineral density Bone turnover markers Radiographic vertebral fractures Timing Evidence on intervention effectiveness was derived from studies of at least one year’s duration. Settings All relevant settings were considered, including community dwelling and institutionalized populations. ©Institute for Clinical and Economic Review, 2017 Page 3 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 2. The Topic in Context Osteoporotic fractures occur most commonly in older, white women. Of the estimated 2 million fractures occurring in 2005 in the United States, 71% occurred in women and only 14% occurred in non-white Americans.3 For example, the age-standardized rates of hip fracture in 2008-2009 were 58% lower in black women than white women, 49% lower in Asian women, and 39% lower in Hispanic women.37 Osteoporosis is diagnosed primarily through measurement of bone density at the hip and lumbar spine. Bone density is reported as the number of standard deviations from the bone mass of a young, healthy woman. This is called the T-score. Since humans achieve peak bone mass around the age of 30, the T-score is usually negative. A T-score of -1 or higher is considered normal; a T- score between -1 and -2.5 is considered low bone mass or osteopenia; and a T-score less than -2.5 is considered osteoporosis. The average T-score for a 75-year old white woman is -2.5, so approximately half of white women ages 75 and older have osteoporosis. Osteoporosis is also diagnosed when an individual experiences a fragility fracture in a location associated with osteoporosis. A fragility fracture is a fracture from a low-energy injury that would not normally be expected to result in a broken bone, such as a fall from standing height or less. The most common fractures associated with osteoporosis are vertebral (27%), wrist (19%), hip (14%), and pelvic (7%).3 The US Preventive Services Task Force (USPSTF) recommends screening average-risk women with a bone density measurement at age 65, and screening younger women who have risk factors that give them the risk of a 65-year old woman.38 However, screening rates are only about 26% for women 65 to 74 years of age.39 The assessment of bone mineral density or treatment for osteoporosis is also a Health Effectiveness Data and Information Set (HEDIS) quality measure for older women.40 Common risk factors for osteoporosis include older age, female sex, prior fractures, smoking, low body mass index, hyperthyroidism, excessive alcohol intake, malabsorption, and some medications (corticosteroids, seizure medications). Many other less common medical conditions and medications impact the risk of fracture. Several organizations have treatment guidelines for osteoporosis including the National Osteoporosis Foundation (NOF), the American Society for Bone and Mineral Research (ASBMR), the American Association of Clinical Endocrinologists (AACE), the American College of Endocrinology (ACE), the North American Menopause Society (NAMS), and the American College of Physicians (ACP).5-9 There is general agreement that treatment is indicated for patients over age 50 who have experienced a hip or vertebral fracture or have a bone density T-score less than or equal to -2.5. Treatment may also be indicated for patients with a T-score from -1 to -2.5 and a 10-year probability of hip fracture ≥ 3% or a 10-year probability of a major osteoporotic fracture ≥ 20%. For ©Institute for Clinical and Economic Review, 2017 Page 4 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents most patients, first-line therapy is to ensure adequate vitamin D and calcium intake, weight bearing exercise, and an oral medication from the bisphosphonate class of drugs. If patients are unable to tolerate oral bisphosphonates or compliance cannot be ascertained, then IV bisphosphonates are generally recommended. Bone is constantly broken down (resorption) and rebuilt; bisphosphonates work by decreasing bone resorption. There are several other drugs approved for osteoporosis that also decrease bone resorption (estrogen, calcitonin, raloxifene, denosumab). They are not considered first-line therapies because of side effects, less evidence of efficacy, route of administration, and/or cost. Osteoporotic fractures can lead to pain, disability, and death. Even vertebral fractures that do not come to clinical attention may result in loss of height and pronounced curving of the spine (kyphosis) that interferes with activities and make breathing difficult. Patients have become increasingly concerned about two adverse events associated with use of bisphosphonate therapy: osteonecrosis of the jaw and atypical femoral fractures. These concerns may partially explain the 50% decrease in the use of bisphosphonate therapy from 2008 to 2012 in the US.10 Practitioners and clinical societies have noted that rates of osteonecrosis of the jaw and atypical femoral fractures in treated patients are much lower than rates of hip fractures in untreated individuals, and that the overall benefit of treatment is far greater than the harm.6 Adherence with bisphosphonate therapy is a major concern. The oral bisphosphonates must be taken with water on an empty stomach in the morning and then the patient needs to remain upright for at least 30 minutes without consuming any additional food or medications. Observational studies in the real world estimate that only 45% of patients remain adherent with oral bisphosphonate therapy one year after the initial prescription and only 30% after two years.11 The long-acting bisphosphonate, zoledronic acid, which requires only one IV infusion each year may have greater adherence, but some studies report greater than 50% discontinuation of therapy with zoledronic acid by two years.12 This appears to be a problem across classes of parenteral agents for osteoporosis with discontinuation rates at one year of 49% for denosumab (a fully humanized monoclonal antibody against the RANKL cytokine with anti-resorptive effects), 59% for zoledronic acid, and 67% for teriparatide.12 Given the poor adherence to currently available therapies, new therapies are needed. Individuals on currently-approved therapy continue to experience fragility fractures, so many may benefit from drugs with greater efficacy and acceptable side-effect profiles. Anabolic or Bone-Building Agents Parathyroid Hormone (PTH) and PTH-related Protein (PTHrP) Analog Drugs Teriparatide was the first drug approved by the FDA for the treatment of osteoporosis that works primarily by increasing bone formation rather than decreasing bone resorption. It is indicated for ©Institute for Clinical and Economic Review, 2017 Page 5 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents the treatment of postmenopausal women with osteoporosis who are at high risk for fracture. In the label, high risk for fracture is defined as a history of osteoporotic fracture, or multiple risk factors for fracture, or prior unsuccessful treatment with or intolerance to previous osteoporosis therapy, based upon physician assessment.13 Teriparatide requires a daily injection of 20 mcg under the skin and the drug must be kept refrigerated. Patients are supplied with a pen injector that contains 28 daily doses, which translates to approximately 13 pens per year. In rat studies, teriparatide caused bone tumors (osteosarcomas); however, these have not been observed in humans. Due to concerns that prolonged use could cause osteosarcomas, teriparatide is only used for two years. Abaloparatide is a new PTHrP analog, approved by the FDA on 4/28/17.14 It is indicated for the treatment of postmenopausal women with osteoporosis at high risk for fracture; high fracture risk is defined using the same terms as in the teriparatide label.15 Abaloparatide requires a daily injection of 80 mcg under the skin, but does not require refrigeration after the first dose from each 30-day supply of injector pens. Abaloparatide is administered by a pen injector containing 30 daily doses, or approximately 12 pens per year. Anti-Sclerostin Antibodies Romosozumab is a monoclonal antibody directed at the protein sclerostin. Sclerostin decreases bone formation, and by blocking sclerostin function, romosozumab increases bone formation and thus builds bone. Romosozumab also appears to have anti-resorptive effects. It is given by subcutaneous injection once monthly and requires refrigeration. It has not yet been approved by the FDA, and a decision in 2017 is no longer anticipated while the FDA reviews data from the ARCH trial that includes an unexpected safety signal regarding serious cardiovascular adverse events.4 Definitions Table 1. Categories of Bone Density T-score Category 0 to -1.0 Normal bone mass -1 to -2.5 Low bone mass <-2.5 Osteoporosis Fragility fractures: Fractures caused by forces that would not normally cause a fracture, usually defined as a fall from a standing height or less. Vertebral fractures: The majority of vertebral fractures do not come to clinical attention. As required for FDA approval, the primary outcome in most of the pivotal trials is new vertebral fractures identified by a radiographic assessment of paired x-rays of the spine obtained before randomization and at the end of the trial. These are known as morphometric vertebral fractures. ©Institute for Clinical and Economic Review, 2017 Page 6 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Morphometric vertebral fractures are typically assessed in two ways: semiquantitative (SQ) and quantitative morphometry (QM). The SQ assessment is typically performed by a radiologist who grades each vertebra on lateral radiographs of the spine according to a standard scale (Table 2). For QM assessment, six points are placed on a digital image of the vertebral body at the four corners and the midpoints of the endplates. The height of the vertebral body is measured in three places: anterior, middle, and posterior. Three ratios are typically calculated (anterior height/posterior height; middle height/posterior height, and posterior height/posterior height of adjacent vertebrae. A fracture is typically diagnosed when any of the three ratios are more than three standard deviations or greater from the mean of a reference population for that vertebra (prevalent fractures) or a decrease of at least 20% in height compared to prior imaging of the same vertebra (incident fractures). Each of the approaches has some subjectivity, so they are often combined with one approach used to confirm the other or requiring agreement from more than one radiologist. The optimal definition remains controversial. Table 2. Grading of Vertebral Fractures Decrease in Height Category < 20% Normal (Grade 0) 20% to 25% Mild (Grade 1) 26% to 40% Moderate (Grade 2) >40% Severe (Grade 3) The subset of morphometric fractures that come to clinical attention are called clinical vertebral fractures. Major osteoporotic fracture: A major osteoporotic fracture is a fracture of the proximal humerus, the wrist, the hip, or a clinical vertebral fracture. Non-vertebral fractures: Non-vertebral fractures exclude fractures of the skull, face, fingers, toes, metacarpals, and vertebrae as well as pathologic fractures and fractures associated with severe trauma. Insights Gained from Discussions with Patients and Patient Groups In the NOF’s Bone Health Index Survey in 2016, patients ranked loss of independence (42%) and lost mobility (25%) as their top two concerns.16 The primary concern among caregivers of patients with osteoporosis was that they would not be able to manage the care of their loved one (50%). Other notable findings included 60% of patients who had sustained a fracture reported not being referred for a bone density test, and fewer than half (47%) were prescribed a medication for osteoporosis. Among those prescribed a medication, 38% said that they never took it, primarily because of fears about side effects (79%). More than half of patients (51%) who started a medication stopped taking it because of side effects (53%) or concerns about the risk for side effects (38%). ©Institute for Clinical and Economic Review, 2017 Page 7 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Patient groups told us that clinical trials rarely report the outcomes that are most meaningful to patients. These include living independently, the ability to perform the activities of daily living, social engagement, quality of life, reduced fear and anxiety about the disease and treatment, and safety from adverse drug effects. Other outcomes include pain, mobility, depression, and caregiver burden. The details of taking the medication are also important. Medications that require refrigeration (teriparatide, romosozumab) may be particularly burdensome. Many patients have a fear of needles, which is another barrier to adherence with all of the anabolic therapies. There are also insurance barriers to treatment. One patient noted that “health care today is so confusing with copay and coinsurance that I never know what is the right way to go.” Patients also note that insurance often requires that they fail an oral therapy before authorizing an injectable therapy. This adds administrative burden on clinicians, and extra office visits for patients. ©Institute for Clinical and Economic Review, 2017 Page 8 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 3. Summary of Coverage Policies and Clinical Guidelines 3.1 Coverage Policies To understand the insurance landscape for osteoporosis treatments, we reviewed publicly-available coverage policies from the Centers for Medicare and Medicaid Services (CMS); California Department of Health Care Services (DHCS); Medicare Part D Plans offered by major private national and California-based insurers (Aetna, Anthem, Cigna, Humana, United Healthcare [UHC], Health Net and Blue Shield of California [BSCA]); and silver-tier Covered California plans offered in both Northern and Southern California (Anthem, Health Net, Kaiser Permanente, and BSCA). We focused on policies pertaining to teriparatide, abaloparatide, oral alendronate, and zoledronic acid. We were unable to identify any CMS National Coverage Determinations (NCDs) or Local Coverage Determinations (LCDs) relevant to California related to the use of bisphosphonates, teriparatide, or abaloparatide. California DHCS listed both alendronate and zoledronic acid, but not teriparatide and abaloparatide, on its contract drug list.41 Teriparatide may be covered under Medicare Part B (when administered in a hospital setting or by a home health aide) or Part D (when self-administered). When covered under Part D, each of the surveyed Medicare Part D plans listed teriparatide at the specialty formulary tier, indicating that patients would be subject to higher out-of-pocket costs (Table 3).42-48 Four of seven plans (Anthem, Aetna, UHC, BSCA) required T-scores of -2.5 or lower. However, each of these payers also covered the drug for patients with prior fragility fractures and/or prior treatment failure, contraindication, or intolerance to another osteoporosis therapy, most frequently an oral bisphosphonate. Only two payers defined treatment failure in their policies; BSCA listed a T-score that remains ≤ -2.5 with or without a low-impact fracture while on bisphosphonate treatment, while Cigna listed a “significant” decrease in BMD after one year of treatment or a new fracture while on bisphosphonate treatment. Cigna and BSCA also required prior therapy with denosumab.49-52 Two payers, Cigna and UHC, covered teriparatide with no additional requirements for patients with T-scores of -3.5 or lower.53 Humana did not require a T-score, only that patients demonstrate the failure of or contraindication/intolerance to one oral bisphosphonate.54 Only one payer, Health Net, did not utilize step therapy or prior authorization requirements for teriparatide. As an illustrative example, Anthem’s prior authorization policy covered teriparatide for individuals with a T-score of -2.5 or lower; or a history of one or more fragility fractures at high risk for fracture; or more than three months of systemic corticosteroid use. Risk factors for fracture include a history of fracture, sustained glucocorticoid use, advanced age, family history of osteoporosis, cigarette ©Institute for Clinical and Economic Review, 2017 Page 9 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents smoking, three or more alcoholic drinks per day, etc.; or prior unsuccessful treatment with or intolerance to at least one other osteoporosis therapy. As of June 14, 2017, only one payer (Humana) had updated their publicly available Part D Formulary to mention abaloparatide; the payer does not cover the drug under its Part D Plans.47 All of the surveyed Part D Plans covered alendronate at the lowest or second-lowest formulary tier. One payer, Anthem, included zoledronic acid at the lowest tier; five payers (Aetna, Cigna, UHC, Health Net, BSCA) listed the drug at the highest non-specialty tier (i.e., patients would be subject to greater out-of-pocket costs for zoledronic acid as compared to alendronate); and one payer (Humana) listed the drug at the specialty tier. For zoledronic acid, UHC required patients to meet one of the following criteria: 1) a T-score of lower than -2.5, 2) a recent vertebral compression fracture or fragility fracture of the hip or distal radius, or 3) a T-score from -1 to -2.5 and a 10-year probability of hip fracture greater than 3% or a 10-year probability of major osteoporotic fracture greater than 20%. Humana required patients to attempt treatment with one oral bisphosphonate prior to zoledronic acid, and BSCA required an attempt at oral bisphosphonate therapy or a recent fragility fracture of the hip.55-57 Although Anthem lists a prior authorization requirement for zoledronic acid, we were unable to locate any publicly-available information about its policy. Each of the surveyed silver-tier exchange plans covered teriparatide at the specialty tier (Appendix Table B1).46,58-61 Only Health Net required prior authorization for teriparatide, and their policy required prior unsuccessful treatment with alendronate and a diagnosis of osteoporosis, a high risk of osteoporosis, or glucocorticoid-induced osteoporosis before coverage will be authorized.62 As of June 14, 2017, only two payers listed abaloparatide in their exchange formularies; Anthem listed the drug as non-formulary (i.e., coverage would require an exception), and Health Net listed the drug at the specialty tier. All four plans covered alendronate without prior authorization or step therapy requirements at the lowest or second-lowest formulary tier. Two insurers, Health Net and BSCA, did not include zoledronic acid in their formularies for silver-tier exchange plans; Anthem covered the drug at the fourth, or highest, tier; and Kaiser Permanente covered the drug at the lowest tier. ©Institute for Clinical and Economic Review, 2017 Page 10 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table 3. Representative Medicare Part D Plan Coverage Policies for Teriparatide, Alendronate, and Zoledronic Acid Anthem Aetna Cigna Humana UHC Health Net BSCA Teriparatide Tier 4 5 5 4 5 5 6 ST No No No Yes No No No PA Yes Yes Yes No Yes No Yes T-score ≤ -2.5 ≤ -2.5 ≤ -3.5* None ≤ -3.5* or ≤ -2.5† None ≤ -2.5 Tx Failure 1 oral BP 1 oral BP/SERM 1 oral BP and denosumab 1 oral BP 1 BP None 1 monthly BP and denosumab I/C 2 oral BP 2 oral BP or SERMs 1 oral BP and denosumab 2 oral BPs 1 BP None 1 monthly BP and denosumab Alendronate Tier 1 1 2 2 1 1 2 ST No No No No No No No PA No No No No No No No Zoledronic Acid Tier 1 4 4 4 4 4 5 ST No No No No No No No PA Yes No Yes‡ Yes Yes‡ No Yes T-score None None None None -2.5 None None Tx Failure None None None 1 oral BP None None 1 oral BP§ I/C None None None 1 oral BP None None 1 oral BP§ BP: bisphosphonate, BSCA: Blue Shield of California, I/C: intolerance/contraindication, PA: prior authorization, SERM: selective estrogen receptor modulator, ST: step therapy, Tx: treatment, UHC: United Healthcare *Individuals with a T-score of less than -3.5 do not need to meet failure/intolerance/contraindication criteria †Also requires a prior fragility fracture or tx failure/intolerance. Coverage is authorized regardless of BMD T-score for individuals with a prior fragility fracture and bisphosphonate failure/intolerance/contraindication. ‡PA only to determine whether coverage is provided under Medicare Part B or D. As an infused drug, zoledronic acid would be covered under Part D. §Coverage is also authorized for individuals with a recent hip fragility fracture ©Institute for Clinical and Economic Review, 2017 Page 11 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 3.2 Clinical Guidelines To better understand the perspective of clinical specialty societies on the appropriate treatment of postmenopausal osteoporosis, we reviewed guideline statements issued by selected US and ex-US organizations. For the purposes of this report, we have focused on recommendations that are relevant to the treatment of postmenopausal women with osteoporosis and have not summarized guiding statements related to primary prevention, secondary osteoporosis, or the treatment of osteoporosis in men. All of the guidelines used terms such as “low,” “moderate,” “high,” or “severe” risk for fracture, but did not explicitly define these levels of risk. American Association of Clinical Endocrinologists (AACE) and American College of Endocrinology (ACE), 20166 The AACE/ACE guidelines, which are based upon expert opinion, recommend that all postmenopausal women over the age of 50 be screened for osteoporosis risk. Osteoporosis may be diagnosed in patients who meet one of four criteria: 1) a T-score of ≤ -2.5 in the lumbar spine, femoral neck, total hip, and/or radius; 2) a fragility fracture at any BMD T-score; 3) osteopenia (T- score of -1.0 to -2.5) and a fragility fracture of the humerus, pelvis, or distal forearm; 4) or osteopenia and a high FRAX probability of fracture. Low-, moderate-, and high-risk categories are not conclusively defined, but risk factors include ethnicity, age, sex, body mass index (BMI), BMD, family history, long-term glucocorticoid use, rheumatoid arthritis, secondary osteoporosis, more than three units of alcohol intake per day, smoking, and several other factors. Pharmacologic therapy is strongly recommended for individuals who meet the above criteria; for individuals with osteopenia, treatment is indicated when the FRAX 10-year probability of major osteoporotic fracture or hip fracture is ≥ 20% or 3%, respectively. Alendronate, risedronate, zoledronic acid, and denosumab are recommended as first-line agents for most patients with osteoporosis, with oral agents (alendronate and risedronate) being recommended for individuals who are at low to moderate risk of fracture (e.g., younger postmenopausal women without prior fractures and a “moderately low” T-score). Teriparatide, zoledronic acid, and denosumab are recommended for individuals with the highest fracture risk (e.g., older women with multiple prior fractures or a very low T-score; individuals in whom oral therapy is contraindicated due to intolerance, likelihood of poor medication absorption, or difficulties with treatment adherence). Teriparatide should be followed by treatment with an antiresorptive agent to preserve bone density gains and reduction in fracture risk. Combination therapy is not recommended for the treatment or prevention of postmenopausal osteoporosis. The AACE/ACE recommend that the use of teriparatide is limited to 2 years. Oral bisphosphonates may be used for longer periods, but the guidelines suggest “bisphosphonate holidays” may be appropriate after five years of stable treatment for individuals at low to moderate risk, and after six ©Institute for Clinical and Economic Review, 2017 Page 12 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents to 10 years of stable treatment in patients at the highest risk. For individuals treated with zoledronic acid, treatment holidays should be considered after three annual doses for women at low to moderate risk, and after six annual doses for individuals at higher risk. Teriparatide may be used during bisphosphonate holidays for high-risk patients. The AACE/ACE guidelines note that there are no clear data on the optimal duration of a drug holiday, but suggest that the duration of treatment holidays may be longest for zoledronic acid, of moderate length with alendronate, and of shortest duration with risedronate due to each drug’s bone-binding affinity. Resumption of therapy should be considered in patients who experience a fracture or substantial decline in BMD. American College of Physicians (ACP), 20179 The ACP guidelines, which are based on a systematic review conducted for the Agency for Healthcare Research and Quality (AHRQ), recommend pharmacologic treatment for women with osteoporosis, defined as a T-score of ≤ -2.5 or a history of fragility fractures. Alendronate, risedronate, zoledronic acid, and denosumab are recommended as first-line treatment options, while the use of both raloxifene and estrogen therapy with or without a progestogen is discouraged. The ACP recommends that pharmacologic treatment continue for up to five years, but noted that there is low-quality evidence to suggest an optimal treatment duration. BMD monitoring during pharmacologic treatment is not recommended; the ACP notes that there is no current evidence demonstrating that such monitoring provides any benefit. Decisions to treat women with osteopenia who are over the age of 65 should be made based on individual patient preference, fracture risk, the balance of benefits and harms, and cost considerations. American Society for Bone and Mineral Research (ASBMR), 20165 The ASBMR guidelines, which are based upon expert opinion, pertain to the management of osteoporosis in patients who are on long-term bisphosphonate treatment. For postmenopausal women on bisphosphonate therapy, the ASBMR recommends that physicians reassess their patients’ fracture risk after five years of oral bisphosphonate treatment, or three years of intravenous bisphosphonate treatment. Physicians should consider continuing therapy or switching to an alternative agent in patients who experience a hip, spine, or multiple other osteoporotic fractures during the initial treatment period; who have an on-therapy hip BMD T-score of ≤ -2.5; or who remain at high risk for fracture based on factors including age, body mass index (BMI), or a history of major osteoporotic fracture. Postmenopausal women who meet these criteria should be considered candidates for up to 10 years of treatment with an oral bisphosphonate or six years of treatment with an IV bisphosphonate. The guidelines recommend that patients who continue treatment be re-evaluated for fracture risk every two to three years. Patients with low to moderate risk of fracture after treatment may be considered candidates for a drug holiday of two to three years in length, and patients on a drug holiday should be reassessed for fracture risk every two to three years. Earlier reassessment should be considered for patients ©Institute for Clinical and Economic Review, 2017 Page 13 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents who experience a fracture during the drug holiday, and for individuals who are likely to experience rapid bone loss due to other factors such as treatment with glucocorticoids. National Osteoporosis Foundation (NOF), 20147 The NOF guidelines, which are based upon a cost-effectiveness analyses plus expert opinion, recommend pharmacologic treatment in women with a history of clinical or radiographic hip or vertebral fractures; in patients with a T-score of ≤ -2.5 at the femoral neck, total hip, or lumbar spine; or in patients with osteopenia (T-score between -1.0 and -2.5 at the femoral neck, total hip, or lumbar spine) and FRAX 10-year probability of major osteoporotic fracture or hip fracture of ≥ 20% or 3%, respectively. Patients with severe osteoporosis should initiate treatment with an anabolic agent, and anabolic therapy should be immediately followed by a bisphosphonate. Combination therapy with teriparatide and an anti-resorptive therapy may be considered in rare cases, such as in patients those with very severe osteoporosis (e.g., a hip and spine fracture), and short-term combination therapy with a two anti-resorptive agents may be considered for women who experience bone loss while being treated with low-dose hormone therapy for menopausal symptoms or raloxifene for breast cancer prevention. In contrast to the AACE/ACE guidelines, the NOF recommendations do not specify which treatments (bisphosphonates or anabolic agents) are most appropriate for patients of various levels of risk. The NOF guidelines do not recommend indefinite treatment with any agent. The guidelines note that the benefits of anabolic therapy diminish rapidly if not followed by an anti-resorptive treatment, but that the benefits of anti-resorptive therapy persist after treatment discontinuation. As such, it is appropriate to consider treatment discontinuation for patients at “modest” risk of fracture after three to five years of treatment with bisphosphonates. Patients with a high fracture risk despite treatment should continue to take bisphosphonates or an alternative therapy. North American Menopause Society (NAMS), 20108 The NAMS guidelines, which are based upon expert opinion, recommend pharmacologic treatment in postmenopausal women who have had an osteoporotic fracture of the vertebra or hip; or who have a T-score ≤ -2.5 in the lumbar spine, femoral neck, or total hip; or who have a T-score between -1.0 and -2.5 and a 10-year risk of major osteoporotic fracture or hip fracture ≥ 20% and 3%, respectively. NAMS recommends that bisphosphonates be used as first-line treatments, and that teriparatide be reserved for individuals at high risk for osteoporotic fracture. The guidelines do not recommend an optimal duration for bisphosphonate treatment. Teriparatide may be used for a maximum of 24 months. The guidelines do not include recommendations related to treatment sequencing or combination therapy. Treatment discontinuation should be guided by individual patient characteristics, including fracture risk and response to therapy. ©Institute for Clinical and Economic Review, 2017 Page 14 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents National Institute for Health and Care Excellence (NICE), 201763 The NICE guidelines (from the United Kingdom) pertain only to the prevention of fracture in women with osteoporosis who have had a fragility fracture. Alendronate is recommended as a first-line treatment option. Risedronate and etidronate are listed as second-line treatment options for patients who cannot comply with alendronate’s administration requirements or have a contraindication or intolerance to the drug. In addition, patients must meet several criteria related to age, BMD, and the presence of independent clinical risk factors (e.g., history of hip fracture in the patient’s parent, more than 4 units of alcohol consumption per day, rheumatoid arthritis). For example, treatment would be recommended for a woman aged 55-59 years with a T-score of -4.0 and no independent risk factors, and for women in the same age range with a T-score of -3.5 and one independent risk factor. Strontium ranelate and raloxifene are considered third-line therapies for patients who cannot comply with administration instructions for first- or second-line treatments or have a contraindication or intolerance to those options. Similar additional criteria related to BMD, age, and independent risk factors are also applied. The guidelines recommend teriparatide as a third-line option for women who have been unsuccessfully treated with alendronate and either risedronate or etidronate, or who have a contraindication/intolerance to the aforementioned drugs (including strontium ranelate, but not raloxifene). In addition, candidates for teriparatide should be ages 65+ with a T-score of -4.0, or ages 65+ with a T-score of ≤ -3.5 and more than two fractures, or ages 55-64 with a T-score of ≤ -4.0 and more than two fractures. Denosumab is also listed as a third-line treatment option for patients unable to appropriately administer alendronate and either risedronate or etidronate, or who have a contraindication or intolerance to those drugs. The guidelines recommend that physicians and patients discuss cessation of bisphosphonate treatment after three years of therapy. Several factors may inform these discussions, including individual choice, fracture risk, and life expectancy. ©Institute for Clinical and Economic Review, 2017 Page 15 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 4. Comparative Clinical Effectiveness 4.1 Overview We abstracted data from the pivotal randomized trials of teriparatide, abaloparatide and romosozumab. We focused primarily on fracture outcomes (vertebral, hip, wrist, non-vertebral) and potential harms. Given the paucity of head-to-head trials, we performed a network meta- analysis (NMA) to generate indirect comparisons between teriparatide, abaloparatide, and zoledronic acid. We included the pivotal trial for zoledronic acid in the NMA because it is the bisphosphonate that the recent AACE/ACE guidelines recommend for patients at highest risk for fracture along with teriparatide.6 Similarly, the FDA indication for teriparatide and abaloparatide are identical, with both drugs being “indicated for the treatment of postmenopausal women with osteoporosis at high risk for fracture, defined as a history of osteoporotic fracture, multiple risk factors for fracture, or patients who have failed or are intolerant to other available osteoporosis therapy.”13,15 We did not include denosumab because it is not an anabolic agent (the primary focus of this report) and because multiple stakeholders recommended that we use zoledronic acid as the primary comparator. We expect that the two anabolic agents will also be primarily used in patients at highest risk for fracture. 4.2 Methods Data Sources and Searches Procedures for the systematic literature review assessing the evidence on anabolic therapies for osteoporosis followed established best methods.64,65 The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.66 The PRISMA guidelines include a list of 27 checklist items, which are described further in Appendix A. We searched MEDLINE/PubMed, EMBASE, and Cochrane Central Register of Controlled Trials for relevant studies. The search was limited to English-language studies of human subjects and focused on trials of at least one year’s duration; articles indexed as guidelines, letters, editorials, narrative reviews, or news items were excluded. The search strategies included a combination of indexing terms (MeSH terms in MEDLINE/PubMed and EMTREE terms in EMBASE), as well as free-text terms, and are presented in Appendix Tables A2-A4. In order to supplement the above searches and ensure optimal and complete literature retrieval, we performed a manual check of the references of recent relevant reviews and meta- ©Institute for Clinical and Economic Review, 2017 Page 16 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents analyses. We also contacted manufacturers, specialty societies, and patient advocacy organizations to ensure that we captured all of the relevant literature. Study Selection After the literature search and removal of duplicate citations using both online and local software tools, study selection was performed using two levels of screening, at the abstract and full-text level. Two reviewers screened the titles and abstracts of all publications identified through electronic searches per the inclusion and exclusion criteria defined by the PICOTS elements; a third reviewer worked with the initial two reviewers to resolve any issues of disagreement through consensus. No study was excluded at abstract-level screening due to insufficient information. For example, an abstract that did not report an outcome of interest in the abstract would be accepted for further review in full text. Citations accepted during abstract-level screening were retrieved in full text for review. Reasons for exclusion will be categorized according to the PICOTS elements during both title/abstract and full- text review. Key inclusion criteria included studies of 1) at least one year’s duration that 2) reported fracture outcomes for 3) postmenopausal women with osteoporosis treated with 4) at least one of the drugs of interest (teriparatide, abaloparatide, romosozumab) compared to 5) another of the drugs of interest or placebo. Data Extraction and Quality Assessment For the systematic literature review, the data abstraction was performed using the following steps: 1. Two reviewers abstracted information from the full articles. 2. Abstracted data were reviewed for logic, and a random proportion of data was validated by a third investigator for additional quality assurance. Information from the accepted studies was extracted into data extraction forms and summarized in Appendix Tables E1-E7. Assessment of Level of Certainty in Evidence We used the ICER Evidence Rating Matrix (Figure 2) to evaluate the evidence for a variety of outcomes. The evidence rating reflects a joint judgment of two critical components: a) The magnitude of the difference between a therapeutic agent and its comparator in “net health benefit” – the balance between clinical benefits and risks and/or adverse effects AND b) The level of certainty in the best point estimate of net health benefit.67 ©Institute for Clinical and Economic Review, 2017 Page 17 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Figure 2. ICER Evidence Rating Matrix Data Synthesis and Statistical Analyses Data on relevant outcomes were summarized in Appendix Tables E5-E7, and synthesized qualitatively below. In addition, we conducted NMAs using a mixed treatment comparison approach.68 Quantitative analyses were conducted using WinBUGS statistical software for Bayesian analysis (MRC Biostatistics Unit, Cambridge, UK). We fit fixed treatment effect models using non-informative ©Institute for Clinical and Economic Review, 2017 Page 18 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents normal priors. A total of 10,000 iterations each were used for both “burn-in” (for model convergence) and model (for model results) simulations. The fixed effects model was chosen as the primary analysis because it is standard practice within a Bayesian environment when the network is almost entirely made up of single study connections. We report the results from random effects models with vague priors in Appendix Tables E12-13: the credible intervals from the random effects NMA are many orders of magnitude wider than those of the original trial results, which reflects the poor accuracy of random effects models when there are primarily single study connections in the network. We performed NMAs for morphometric vertebral fractures and non-vertebral fractures. There was insufficient data to evaluate hip fractures in an NMA. As noted above, we excluded romosozumab from all comparative effectiveness analyses including the NMAs. We reviewed the deviance information criterion (DIC) statistics as well as comparison of the residual deviance (resdev) to the number of unconstrained data points to assess model fit under multiple alternative assumptions. The paucity of studies precluded meta-regression and extensive sensitivity analyses. We did sensitivity analyses excluding the data from the open label teriparatide arm of the ACTIVE trial. 4.3 Results For each of the three anabolic drugs, there is only one pivotal trial. Each pivotal trial is described in detail in the key studies section below. The pivotal study of zoledronic acid is also described because it is the comparator bisphosphonate therapy in the cost-effectiveness model and we included it in the NMA that provides estimates for the reduction in fractures used in the cost model. Multiple stakeholders recommended using zoledronic acid as the comparator because it is a parenteral therapy like the anabolic agents and is used for the same indications (women at very high risk for fracture and those unable to tolerate oral therapy). Study Selection The literature search identified 788 citations (Appendix Figure A1). After reviewing the titles and abstracts, 222 full-text articles were evaluated. Three randomized trials met all inclusion and exclusion criteria.17-19 Details of the studies are summarized in Appendix Tables E1-E7 and briefly in Table 4. ©Institute for Clinical and Economic Review, 2017 Page 19 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table 4. Summary of the Randomized Trials of Anabolic Agents and Zoledronic Acid for Osteoporosis F/U, Age, BMI, Prior Reference Study Group N months years kg/m2 Fracture Neer 200119 Fracture Teriparatide 541 21 69 26.8 100% V Prevention Placebo 544 21 69 26.7 Trial Miller 201618 ACTIVE Abaloparatide 824 18 69 25.0 24% V Teriparatide 818 18 69 25.2 63% any Placebo 821 25.1 Cosman FRAME Romosozumab 3589 12 71 24.7 18% V 2016 17 Placebo 3591 12 71 24.7 22% non-V Black 200720 HORIZON Zoledronic acid 3889 36 73 25.1 63% V Placebo 3876 36 73 25.4 BMI: body mass index, F/U: follow-up, Non-V: non-vertebral fracture, V: vertebral fracture Key Studies The Fracture Prevention Trial – Teriparatide19 The Fracture Prevention Trial randomized 1,085 patients to daily subcutaneous (SC) injections of teriparatide 20 mcg or identical placebo and followed them for 21 of the planned 24 months.19 The study was terminated early to investigate concerns raised because of the development of osteosarcomas in rats during a toxicology study. No osteosarcomas developed in the human participants in this trial. The participants were women at least five years after their menopause who had at least one moderate or two mild vertebral fractures. At baseline, the mean T-score was not reported and 100% had existing vertebral fracture. The primary outcome was not specified, but was likely new morphometric vertebral fractures assessed using the semiquantitative (SQ) approach by a single reading by a radiologist at a central location who was blinded to treatment allocation, but not the order of the radiographs. New vertebral fractures occurred in 5% of women in the teriparatide group and 14% of women in the placebo group (relative risk [RR] 0.35, 95% confidence interval [CI] 0.22-0.55). Non-vertebral fragility fractures occurred in 6% of women in the teriparatide group and 10% of women in the placebo group (RR 0.47, 95% CI 0.25-0.88). Hip fractures occurred in 0.2% of women in the teriparatide group and 0.7% of women in the placebo group (RR not reported). Discontinuation of the study drug due to adverse events was identical in the two groups (6%). Dizziness (9% vs. 3%) and leg cramps (3% vs. 1%) were more common in the teriparatide group. Hypercalcemia was also more common in the teriparatide group (11% vs. 2%). Adherence, based on returned medication, was approximately 81% for both teriparatide and placebo injections at each follow-up visit. ©Institute for Clinical and Economic Review, 2017 Page 20 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Eight years later, the investigators re-analyzed the vertebral fracture data using quantitative morphometry (QM).21 An incident vertebral fracture had to meet the following 3 criteria: 1) 20% decrease in height by quantitative morphometry, 2) a 4 mm decrease in height, and 3) an increase of at least one grade by visual SQ morphometry. Using this definition 1.8% of patients in the teriparatide group and 11.4% of the placebo group had incident vertebral fractures (RR 0.16, 95% CI 0.08-0.33). Note that this incidence of new vertebral fractures using this definition is markedly lower than that reported in the original publication (1.8% vs. 5% in teriparatide group and 11.4% vs. 14.3% in the placebo group). We have decided to use these estimates in our base-case analyses of vertebral fractures because the other pivotal trials used a two-step approach for diagnosing incident vertebral fractures. The ACTIVE Trial – Abaloparatide18 The ACTIVE trial randomized 2,463 patients to daily SC injections of abaloparatide 80 mcg, teriparatide 20 mcg or identical placebo and followed them for 18 months.18 The teriparatide was given open label. The participants were postmenopausal women ages 49 to 86 years who had at least one moderate or two mild vertebral fractures or other fragility fractures in the past five years and bone mineral density (BMD) T-scores between -2.5 and -5.0, or women at least 65 years of age without a history of a fragility fracture with BMD T-scores between -3.0 and -5.0. At baseline, the mean T-score at the total hip was -1.9 and 24% had existing vertebral fracture. The primary outcome was the cumulative incidence of new vertebral fractures defined using the SQ method with each fracture confirmed by a second radiologist also using the SQ technique. New vertebral fractures occurred in 0.6% of women in the abaloparatide group, 0.8% of women in the teriparatide group, and 4.2% of women in the placebo group (abaloparatide hazard ratio [HR] 0.14, 95% CI 0.05- 0.39; teriparatide HR 0.20, 95% CI 0.08-0.47, both vs. placebo). The HR for abaloparatide versus teriparatide was not reported for vertebral fractures. Non-vertebral fragility fractures occurred in 2.7% of women in the abaloparatide group, 3.3% of women in the teriparatide group and 4.7% of women in the placebo group (abaloparatide HR 0.57, 95% CI 0.32-1.00; teriparatide HR 0.72, 95% CI 0.42-1.22, both vs. placebo). The HR for abaloparatide versus teriparatide was 0.79 (95% CI 0.43- 1.45) for non-vertebral fractures. There were no hip fractures in either the abaloparatide or teriparatide groups and 2 (0.2%) in the placebo group (HRs not reported). Discontinuation of the study drug due to adverse events was higher in the abaloparatide group (9.9% vs. teriparatide 6.8% and placebo 6.1%). However, rates of significant adverse events were similar in the three groups (9.7%, 10.0%, and 11%). Hypercalcemia was more common in the PTH analog groups (3.4% abaloparatide, 6.4% teriparatide, 0.4% placebo). Adherence, based on weekly diary recording, was greater than 90% for each of the treatment groups. Patients in both the abaloparatide and placebo groups of the ACTIVE trial were offered an additional two years of follow-up receiving open-label oral alendronate 70 mg weekly and 92% of eligible patients agreed to participate. The six-month follow-up results reported lower rates of ©Institute for Clinical and Economic Review, 2017 Page 21 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents vertebral fractures (HR 0.13, 95% CI 0.04-0.41), non-vertebral fractures (HR 0.48, 95% CI 0.26-0.89), and major osteoporotic fractures (HR 0.42, 95% CI 0.21-0.85) for abaloparatide followed by alendronate compared to placebo followed by alendronate when analyzed from the beginning of the ACTIVE trial.69 However, the number of new fractures in the extension trial was low in both the abaloparatide/alendronate and placebo/alendronate groups (vertebral 0 vs. 7; non-vertebral 3 vs. 7; major osteoporotic 2 vs. 4). This suggests that alendronate therapy can preserve the fracture reduction benefits of abaloparatide, but the interim results should be considered preliminary until the full two-year extension study results are published. The FRAME Study – Romosozumab17 The FRAME study randomized 7,180 patients to monthly SC injections of romosozumab 210 mg or identical placebo for 12 months followed by an additional 12 months of denosumab.17 The participants were women ages 55 to 90 years of age with BMD T-scores between -3.0 and -5.0. Mean total hip T-score was -2.5 and 18% had vertebral fractures at baseline. The co-primary outcomes were the cumulative incidence of new vertebral fractures at 12 and 24 months. Incident vertebral fractures were defined as an increase of at least one severity grade using the SQ method. Confirmation by a second radiologist was not reported in the primary publication or the study protocol, so this outcome may be similar to that reported in the original Fracture Prevention Trial publication.19 At 12 months, new vertebral fractures occurred in 0.5% of women in the romosozumab group and 1.8% of women in the placebo group (RR 0.27, 95% CI 0.16-0.47). Non- vertebral fractures occurred in 1.6% of women in the romosozumab group and 2.1% of women in the placebo group (RR 0.75, 95% CI 0.53-1.05). Hip fractures occurred in 0.2% of women in the romosozumab group and 0.4% of women in the placebo group (RR 0.54, 95% CI 0.22-1.35). Effect modification was evaluated in 11 subgroups including age, history of fracture, T-score, and geographic region for new vertebral, clinical, and non-vertebral fractures. The treatment effects were consistent in all subgroups except for treatment by region interactions for clinical and non- vertebral fractures (nominal p values 0.03 and 0.04, respectively). Post-hoc analyses suggested that romosozumab may be less effective in the Latin American region, though this could be a chance finding given the multiple comparisons performed without any adjustment. During the first 12 months, discontinuation of the study drug due to adverse events was similar in the two groups (2.9% vs. 2.6%). There were seven patients with serious possible hypersensitivity reactions in the romosozumab group. In addition, injection site reactions were more common in the romosozumab group (5.2% vs. 2.9%). Of note in such a short study, one patient in the romosozumab group had an atypical femoral fracture and one had osteonecrosis of the jaw. These events may be due to chance, but could reflect the anti-resorptive properties of romosozumab. Adherence was not reported. After 12 months, all patients in the FRAME study received denosumab 60 mg SC every six months for an additional 12 months. The cumulative risk for the full 24-month period for new vertebral ©Institute for Clinical and Economic Review, 2017 Page 22 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents fractures (RR 0.25, 95% CI 0.16-0.40) and non-vertebral fractures (HR 0.75, 95% CI 0.57-0.97) was lower in the romosozumab/denosumab group than in the placebo/denosumab group. In the second year, there were 5 new vertebral fractures in the romosozumab/denosumab group and 25 in the placebo/denosumab group. During the second 12-month period there was one additional case of osteonecrosis of the jaw in the group treated with romosozumab followed by denosumab. The HORIZON Study – Zoledronic Acid20 The HORIZON study randomized 7,765 patients to annual IV infusions of zoledronic acid 5 mg or identical placebo and followed them for 36 months.20 The participants were women ages 65 to 90 years with BMD T-scores less than –2.5 or prior vertebral fracture with T-score less than -1.5. Mean total hip T-score was not reported, but 63% had vertebral fractures at baseline. The co-primary outcomes were the cumulative incidence of new vertebral fractures and hip fractures. Incident vertebral fractures were defined by a reduction in vertebral height of at least 20% and 4 mm by QM confirmed by an increase of one or more severity grades using the SQ method. New vertebral fractures occurred in 3.3% of women in the zoledronic acid group and 10.9% of women in the placebo group (RR 0.30, 95% CI 0.24-0.38). Non-vertebral fractures occurred in 8.0% of women in the zoledronic acid group and 10.7% of women in the placebo group (RR 0.75, 95% CI 0.64-0.87). Hip fractures occurred in 1.4% of women in the zoledronic acid group and 2.5% of women in the placebo group (RR 0.59, 95% CI 0.42-0.83). The treatment effects were consistent over time with similar reductions in vertebral fractures at years one, two, and three (RR 0.40, 0.29, and 0.30 respectively, p<0.001 at all three time points). There was no evidence of a delay in efficacy for vertebral fractures, non-vertebral fractures, hip fractures, or any clinical fractures. Discontinuation of the study drug due to adverse events was similar in the two groups (2.1% vs. 1.8%). During the three days following the infusion, more patients in the zoledronic acid group reported fever (16.1% vs. 2.1%), myalgias (9.5% vs. 1.7%), and flu-like symptoms (7.8% vs. 1.6%). The post-infusion symptoms decreased over time (first infusion 31.6%; second 6.6%, third 2.8%). Adherence was greater than 90% in both groups. Quality of Individual Studies Using the USPSTF criteria, we rated the three studies to be of good quality (Appendix Table E4). The trials all used appropriate randomization methods with comparable groups at baseline and good retention to retain comparability through the end of the study periods. The studies were all double-blinded with clearly defined interventions and blinded adjudication of outcomes. The key outcomes were addressed and appropriate intention-to-treat analyses were performed. The only exception is for the teriparatide group in the ACTIVE study. The abaloparatide and placebo groups were double-blind, but the teriparatide group received open label treatment. For this reason, we performed sensitivity analyses with and without data from this arm of the ACTIVE study in our NMAs. The HORIZON study was also rated as good quality. ©Institute for Clinical and Economic Review, 2017 Page 23 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Clinical Benefits The essential clinical benefit of the anabolic drugs for osteoporosis is the prevention of fragility fractures. The primary outcome in the pivotal trials was incident morphometric vertebral fractures, even though more than half of these fractures are not clinically apparent. Non-vertebral fragility fractures were also reported as they are relatively common and clinically important. Finally, hip fractures are clinically the most important in terms of impact on a patient’s quality of life, but they are uncommon. All three anabolic studies had insufficient power to demonstrate a reduction in hip fractures. However, the HORIZON study demonstrated that zoledronic acid significantly reduced the incidence of hip fractures in women with osteoporosis.20 Morphometric Vertebral Fractures The pivotal trials of teriparatide, abaloparatide, and zoledronic acid all reported a significant reduction in vertebral fractures versus placebo (Appendix Table E6), though the definition of incident vertebral fractures differed somewhat between trials. We elected to use a secondary analysis of incident vertebral fractures for the Fracture Prevention Trial21 for primary inputs to this NMA because the definition was closer to that used in the ACTIVE trial.18 The results of the NMA (Table 5) confirmed that all three drugs were significantly better than placebo at reducing morphometric vertebral fractures. Neither of the two anabolic agents was significantly different from the other, nor were they significantly different from zoledronic acid: the credible intervals for each of the comparisons between active drugs each contain 1 (Table 5). We performed multiple sensitivity analyses including 1) using the original definition of vertebral fractures in the Fracture Prevention Trial, 2) excluding the open-label teriparatide data from the ACTIVE trial, and 3) using a random-effects model (Appendix Table E12), but the primary conclusions did not change. ©Institute for Clinical and Economic Review, 2017 Page 24 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table 5. Network Meta-Analysis Results for the Relative Risk of Morphometric Vertebral Fractures* Abaloparatide (80 mcg) 0.76 Teriparatide (0.20 – 2.26) (20 mcg) 0.44 0.57 Zoledronic Acid (0.12 – 1.15) (0.30 – 1.02) (5 mg) 0.13 0.17 0.30 Placebo (0.03 – 0.33) (0.09 – 0.29) (0.24 – 0.37) *Includes data from open-label teriparatide arm of the ACTIVE trial18 and from the secondary analysis of vertebral fractures for the Fracture Prevention Trial.21 Fixed-effects model; resdev = 5.646, DIC = 47.781 Legend: Each box represents the estimated rate ratio and 95% credible interval for the combined direct and indirect comparisons between two drugs: the drug at the top of the column compared to the drug at the right of the row. Estimates in bold signify that the 95% credible interval does not contain 1. As expected, the NMA estimates for the reduction in vertebral fractures for each drug versus placebo are similar to the direct estimates versus placebo in the randomized trials (Table 6). Table 6. Comparison of the Relative Risk Versus Placebo for Morphometric Vertebral Fractures between the Network Meta-Analysis and the Randomized Controlled Trials Drug NMA Estimate RCT Estimates Abaloparatide 0.13 (0.03-0.33) 0.14 (0.05-0.39) Teriparatide 0.17 (0.09-0.29) 0.16 (0.08-0.33) 0.20 (0.08-0.47) Zoledronic acid 0.30 (0.24-0.37) 0.30 (0.24-0.38) Non-Vertebral Fragility Fractures In the key randomized trials, both teriparatide and abaloparatide significantly reduced non- vertebral fractures (Appendix Table E6). The results of the NMA confirmed this finding (Table 7). Again, neither of the anabolic agents were significantly different from one another, nor were they significantly different from zoledronic acid. Zoledronic acid significantly reduced non-vertebral fractures in the HORIZON trial and in the NMA. Again, we performed sensitivity analyses including 1) excluding the open-label teriparatide data from the ACTIVE trial, and 2) using a random-effects model (Appendix Table E13). ©Institute for Clinical and Economic Review, 2017 Page 25 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table 7. Network Meta-Analysis Results for the Relative Risk of Non-Vertebral Fragility Fractures* Abaloparatide (80 mcg) 0.83 Teriparatide (0.46 – 1.46) (20 mcg) 0.69 0.82 Zoledronic Acid (0.38 – 1.16) (0.54 – 1.22) (5 mg) 0.51 0.61 0.75 Placebo (0.28 – 0.85) (0.41 – 0.88) (0.64 – 0.87) *Includes data from open-label teriparatide arm of the ACTIVE trial Fixed-effects model; resdev = 6.518, DIC = 52.897 Legend: Each box represents the estimated rate ratio and 95% credible interval for the combined direct and indirect comparisons between two drugs: the drug at the top of the column compared to the drug at the right of the row. Estimates in bold signify that the 95% credible interval does not contain 1. As with vertebral fractures, the NMA estimates for the reduction in non-vertebral fractures for each drug versus placebo are similar to the direct estimates versus placebo in the randomized trials (Table 8). Table 8. Comparison of the Relative Risk versus Placebo for Non-Vertebral Fractures between the Network Meta-Analysis and the Randomized Controlled Trials Drug NMA estimate RCT estimates Abaloparatide 0.51 (0.28-0.85) 0.57 (0.32-1.00) Teriparatide 0.61 (0.41-0.88) 0.47 (0.25-0.88) 0.72 (0.42-1.22) Zoledronic acid 0.75 (0.64-0.87) 0.75 (0.64-0.87) Hip Fractures Among the anabolic studies the incidence of hip fractures was low (7 hip fractures total in the abaloparatide and teriparatide studies). Only the FRAME study reported relative risks (Appendix Table E6). Hip fractures were significantly reduced for zoledronic acid compared to placebo in the HORIZON trial (RR=0.59, CI 0.42-0.83). Relative risk estimates for abaloparatide and teriparatide were not reported in the clinical trials. We examined a network meta-analysis, but the results were unstable and unrealistically low (much lower than the estimates for vertebral fractures, which has not been observed for any other drug used to prevent fractures). Given the lack of face validity of these estimates, we did not use them in assessing the comparative effectiveness of these drugs, nor were the estimates used in our cost modeling. ©Institute for Clinical and Economic Review, 2017 Page 26 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Bone Mineral Density Change in BMD is often used as a surrogate marker in preliminary studies of drugs to prevent osteoporotic fractures. The change in BMD for the anabolic agents and zoledronic acid in the pivotal trials are summarized in Appendix Table E7. The anabolic agents had large increases in BMD of the lumbar spine (approximately 10% to 13% over 12 to 21 months), while zoledronic acid had smaller gains (6.7% over 36 months). At the total hip, the increases compared to placebo were greatest for romosozumab (6.9%) and zoledronic acid (6.0%), with somewhat smaller gains for abaloparatide (4.3%) and teriparatide (3.6%). The changes in BMD at the femoral neck were similar to those observed at the total hip. Because change in BMD is an imperfect predictor of fracture prevention, it is difficult to draw firm conclusions from these results. Harms Table 9 summarizes the harms of the anabolic therapies observed in the clinical trials. There were no important differences in serious adverse events between the anabolic therapy groups and placebo groups. In the ACTIVE trial, the abaloparatide group had a greater percentage of patients discontinue therapy due to adverse events than the teriparatide or placebo groups, but the difference was small (10% vs. 7% and 6%, respectively). There was one case of osteonecrosis of the jaw and one atypical femoral fracture observed during the one year of treatment with romosozumab during the FRAME trial, but these may be chance findings. Similarly, there was one case of osteonecrosis of the jaw observed in the placebo group of the HORIZON trial. No other cases of osteosarcoma were observed in any of the trials. As described in the Key Trials section above, there were more cases of hypercalcemia with teriparatide and abaloparatide and more injection site reactions with romosozumab, but most were mild and self-limited, though some required dose reduction or a decrease in calcium supplementation. Half of patients treated with abaloparatide developed anti-abaloparatide antibodies, but these did not significantly impact fracture efficacy or adverse events. ©Institute for Clinical and Economic Review, 2017 Page 27 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table 9. Key Harms in Randomized Trials of Anabolic Agents for Osteoporosis Discontinuation Kidney Hyper- Reference Group SAEs AFF ONJ due to AE stones Ca Teriparatide Neer 200119 Teriparatide NR 6% NR NR No 11% Placebo NR 6% NR NR sign. 2% diff. Abaloparatide Miller 201618 Abaloparatide 9.7% 9.9% NR NR NR 3.4% Teriparatide 10.0% 6.8% NR NR NR 6.4% Placebo 11.0% 6.1% 0.4% Romosozumab Cosman 201617 Romosozumab 9.6% 2.9% 1 1 NR NR Placebo 8.7% 2.6% 0 0 NR NR Key comparator: Zoledronic acid Black 200720 Zoledronic acid 29.2% 2.1% NR 1 NR NR Placebo 30.1% 1.8% NR 1 NR NR AE: adverse event, AFF: atypical femoral fracture, Hyper-ca: hypercalcemia, NR: not reported ONJ: osteonecrosis of the jaw, SAE: serious adverse event Additional considerations include the risk for atypical femoral fractures and osteonecrosis of the jaw with bisphosphonates. However, the risk is low. It is estimated that treatment of 10,000 women with zoledronic acid for three years would prevent approximately 710 vertebral fractures, 110 hip fractures, while causing one atypical femoral fracture and less than one case of osteonecrosis of the jaw.70 In addition, a substantial proportion of patients report systemic symptoms following zoledronic acid infusion, which may contribute to decreased long-term adherence. This may be an issue for the injectable anabolic agents as well, as one study found the discontinuation rates of teriparatide after one year were higher than those of zoledronic acid.12 Sensitivity Analyses There were insufficient studies to perform meta-regression, and we did not have individual level data that would have allowed for subgroup analyses. We did repeat the NMA eliminating the teriparatide data derived from the open-label arm of the ACTIVE trial. Similarly, we used the original Fracture Prevention Trial definition of vertebral fracture in a sensitivity analysis. Finally, we performed random effects models in addition to the primary fixed effects models. There were no changes in the conclusions from the NMAs and the changes in the estimates for teriparatide were modest (a slight reduction in efficacy for vertebral, and a slight increase in the reduction of non- vertebral fractures). Results from NMA sensitivity analyses are reported in Appendix Tables E8-E13 ©Institute for Clinical and Economic Review, 2017 Page 28 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Observational Study Results for the Anabolic Agents There are no observational data available for abaloparatide or romosozumab that met our search criteria. Teriparatide, on the other hand, has been in clinical use for more than a decade. No studies reported on morphometric fractures, because these require baseline and follow-up lateral spine x-rays without symptoms. The observational data report a 40% to 73% reduction in clinical vertebral fractures.22,71,72 The observed reduction in non-vertebral fractures ranged from 38% to 45%.22,71,73 Importantly, there was also a significant 45% reduction in hip fractures in one claims database (odds ratio [OR] 0.55, 95% CI 0.42-0.74).22 This is particularly important as the randomized studies of teriparatide had insufficient power to demonstrate a reduction in hip fractures. Unpublished Trials The VERO Study (Teriparatide vs. Risedronate) The Vertebral Fracture Treatment Comparisons in Osteoporotic Women (VERO) trial was presented at the World Congress on Osteoporosis, Osteoarthritis and Muscular Disease in March 2017.74 The investigators randomized 1,360 women with at least two moderate or one severe vertebral fracture and a T-score ≤ -1.5 to two years of teriparatide 20 mcg SC daily or the oral bisphosphonate risedronate 35 mg once a week. Compared with risedronate, patients treated with teriparatide had significantly fewer new vertebral fractures (5.4% vs. 12.0%, HR 0.44, p<0.001) and a non-significant reduction in non-vertebral fragility fractures (4.0% vs. 6.1%, HR 0.66, p=0.099). Adverse events that were more common in the teriparatide group included extremity pain (5.4% vs. 2.6%, p = 0.013), dizziness (4.4% vs. 1.8%, p=0.007, hypercalcemia (2.2% vs. 0.1%, p<0.001), and decreased vitamin D (1.3% vs. 0.1%, p=0.021). These results support the hypothesis that teriparatide prevents more vertebral fractures than risedronate over two years of therapy. The study did not have sufficient power to demonstrate a reduction in non-vertebral fractures or hip fractures. We performed a scenario analysis including the reported VERO study results in our NMA, with bisphosphonates considered as a class. This analysis had little impact on the estimates for the relative risks of the individual drugs compared to placebo (see Appendix Tables E10 and E11). The relative risk for morphometric vertebral fractures remained the same for abaloparatide compared to placebo (0.13), decreased slightly for teriparatide (0.17 to 0.14), and increased slightly for bisphosphates (0.30 to 0.31). Similarly, the relative risk for non-vertebral fractures decreased slightly for abaloparatide compared to placebo (0.51 to 0.49), decreased slightly for teriparatide (0.61 to 0.56), and increased slightly for bisphosphates (0.75 to 0.76). Thus, this scenario analysis confirms the validity of our primary NMA results. ©Institute for Clinical and Economic Review, 2017 Page 29 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents The ARCH Study (Romosozumab Followed by Alendronate vs. Alendronate Alone) The double-blind, placebo-controlled ARCH study results were announced in a press release by Amgen on May 21, 2017.4 The investigators randomized 4,093 postmenopausal women to either romosozumab 210 mg SC every month for 12 months followed by the oral bisphosphonate alendronate for 12 months or alendronate 70 mg orally once weekly for 24 months. The sequence of romosozumab followed by alendronate significantly reduced the incidence of new vertebral fractures (HR 0.50) and clinical fractures (HR 0.73) compared to alendronate alone. The sequenced therapy also significantly reduced the incidence of non-vertebral fractures (HR 0.81). However, there was a concerning safety signal. The incidence of new serious cardiovascular events was higher in the romosozumab group at one year (2.5% vs. 1.9%). There were also more injection site reactions in the romosozumab group (4.4% vs. 2.6%) These results support the hypothesis that starting therapy with an anabolic agent and sustaining the gains with an anti-resorptive agent may be more effective at preventing vertebral and non- vertebral fractures than treating with an anti-resorptive agent for two years. The increase in serious cardiovascular events was not observed in the larger FRAME study (n=7,180), so the observed increase in the ARCH study may be a chance finding. The ARCH study has not yet been presented at a conference or published in a peer reviewed journal, which limits our ability to fully evaluate these results. Controversies and Uncertainties The primary controversy is whether it was appropriate to combine the data from the different study populations of the three trials in a NMA. There were differences in the inclusion criteria of the studies (Appendix Table E2) and in some of the characteristics of patients included in the study, though all were women and the average age and BMI of the participants was very similar across the trials (Table 4). In order for the NMA results to be invalid, there must be effect modification in the relative rate of fractures for one or more of the drugs by patient characteristics that differ significantly between trials. Specific analyses looking for effect modification by patient characteristics such as age, BMD, prior fracture history, and baseline risk for fracture have been published for teriparatide75, abaloparatide76, and romosozumab.17 In all three analyses, risk factors for fracture did not modify the relative efficacy of the drugs. In the FRAME trial, romosozumab appeared to be less effective in participants recruited in Latin America, but this observation was of borderline statistical significance. Given the number of subgroups examined, this may be a chance finding. This finding deserves additional attention, but is not strong enough to invalidate the NMA. It is also worth examining the incidence of fractures in the placebo groups in each of the four pivotal trials as an indicator of the underlying risk for fractures in patients enrolled in the trials (Table 10). ©Institute for Clinical and Economic Review, 2017 Page 30 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table 10. Annual Incidence of Vertebral, Non-Vertebral, and Hip Fractures in Placebo Groups Reference Trial Vertebral Fx Non-Vertebral Fx* Hip Fx Teriparatide Fracture Prevention Trial 6.5 3.1 0.4 Abaloparatide ACTIVE 2.4 2.7 0.2 Romosozumab FRAME 1.8 2.1 0.4 Zoledronic acid HORIZON 3.6 3.6 0.8 Fx: fracture *Includes hip fractures The annual risk for vertebral fractures was particularly high in the Fracture Prevention Trial because all participants had prior vertebral fractures. The annual vertebral fracture rates in the other trials were higher in the trials with greater prevalence of vertebral fractures at baseline (Appendix Table E1). For non-vertebral fragility fractures and hip fractures, the annual risks were reasonably similar across the trials. It is worth noting that patients in the HORIZON trial were at as high or higher risk for fracture as patients included in the pivotal trials of the anabolic agents. None of the published NMAs of drug therapy for osteoporosis included abaloparatide (see Appendix C).23-27 Similar to our findings, the NMAs concluded that both teriparatide and zoledronic acid reduce the risk of vertebral and non-vertebral fractures compared to placebo. They found no significant differences between the drugs, though teriparatide ranked higher than zoledronic acid. They also concluded that zoledronic acid reduced hip fractures, but there was insufficient evidence for teriparatide. A major area of uncertainty is due to the relative paucity of evidence for each of the anabolic agents, particularly for the hip fracture outcome. The trials were relatively small given the large number of women with osteoporosis. In addition, active treatment continued for only one to two years. We could not model stable estimates for hip fracture reduction because of the low number of events. Indeed, the recent ACP clinical guideline did not recommend any of the anabolic agents as first-line therapy for osteoporosis because of the lack of randomized trial evidence on hip fracture prevention.9 Some have suggested that anabolic therapy may have more rapid onset of fracture prevention than antiresorptive therapy. Given the paucity of head-to-head trials, it is difficult to evaluate this hypothesis. However, in the HORIZON trial, the reductions seen with zoledronic acid in hip fractures, non-vertebral fractures, and any clinical fractures, as assessed by the Kaplan-Meier curves, appeared to begin at randomization. The reduction in clinical vertebral fractures may have been delayed, but is unlikely to be clinically or statistically significant. Indeed, the disconnect between change in bone mineral density and reduction in fractures has been widely recognized for bisphosphonates.77-79 In the ACTIVE trial, abaloparatide appeared to have a more rapid reduction in non-vertebral fractures, clinical fractures, and major osteoporotic fractures than teriparatide, but the differences were not statistically significant except for major osteoporotic fractures (p=0.03). ©Institute for Clinical and Economic Review, 2017 Page 31 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents The Kaplan-Meier curves for clinical vertebral fractures were not shown. There are insufficient data to assess the relative efficacy of the anabolic agents compared to zoledronic acid in the first three to six months of therapy. There are no significant differences in fracture reduction between anabolic therapy and zoledronic acid over longer time periods. Another important area of uncertainty is sequencing of therapies. Studies suggest that the bone density gains from anabolic agents are quickly lost if no follow-up therapy is used.28 Since anabolic agents are only used for one to two years, they will need to be followed by some form of anti- resorptive therapy to maintain the reduction in fracture risk. Other studies have found that the beneficial effects of PTH-related therapies on bone mass are blunted among individuals previously treated with anti-resorptive drugs.29 This suggests that anabolic agents may be most effective if used prior to anti-resorptive therapy. The best agent to use and the optimal length of follow-up treatment is uncertain and awaits additional fracture endpoint studies. The outcomes of greatest interest to patients are maintenance of independence and prevention of disability. These and other patient-centered outcomes were not reported in the pivotal trials. Summary The evidence to date demonstrates with high certainty that the two anabolic agents reduce vertebral fractures compared to no therapy. However, there is insufficient evidence to distinguish the anabolic agents from each other and from zoledronic acid for vertebral fractures. The differences in fracture reduction are small and the credible intervals all contain 1, so the therapies may be comparable. The evidence is even less certain for non-vertebral fragility fractures and, in particular, hip fractures. The harms of therapy are relatively small and have little influence on the net benefit for each therapy compared to the others. Adherence to both initial anabolic therapy and subsequent anti-resorptive therapy is essential to preserve the fracture reduction benefit. However, there are minimal real-world data available to compare adherence to therapy between the two anabolic agents. For the two anabolic agents, we judged the evidence to be promising, but inconclusive (P/I) for the net health benefit when compared to zoledronic acid in postmenopausal women with osteoporosis at high risk for fracture. When compared to no treatment, we judged with moderate certainty that the anabolic agents provided a small or substantial net health benefit compared to no therapy, with high certainty of at least a small net health benefit when compared to no therapy (B+). There is a substantial reduction in vertebral fractures, a small to moderate reduction in non-vertebral fractures, and uncertain benefits for hip fractures, though observational data do support a benefit for teriparatide. ©Institute for Clinical and Economic Review, 2017 Page 32 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents When abaloparatide is compared to teriparatide, we judged that the evidence is insufficient to assess the comparative clinical effectiveness of the two drugs. The extensive real world clinical experience with teriparatide without identification of new adverse events and observational evidence confirming benefits is reassuring. However, in the ACTIVE trial, there was a non-significant trend towards greater reduction in both vertebral and non-vertebral fractures with abaloparatide compared with teriparatide. ©Institute for Clinical and Economic Review, 2017 Page 33 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 5. Other Benefits or Disadvantages Our reviews seek to provide information on other benefits or disadvantages offered by the intervention to the individual patient, caregivers, the delivery system, other patients, or the public that would not have been considered as part of the evidence on comparative clinical effectiveness. 1. Unmeasured patient health benefits: There are no clear differences among the drugs. 2. Relative complexity of the treatment regimen that is likely or demonstrated to significantly affect adherence and outcomes: There are important differences in the treatments that may be important for some patients and preferences will differ among patients. Abaloparatide and teriparatide require daily injections, which is a barrier to adherence for some patients. The comparator, zoledronic acid requires an annual visit for a 15-minute infusion that can be associated with systemic symptoms, particularly following the first dose. The once-a-year dosing may be an advantage, but the requirement for an intravenous infusion may decrease acceptability. In addition, some patients may have concerns about a drug that remains in the body for a long time. 3. Impact on productivity and ability of the patient to contribute to personal and national economic activity: No clear differences among the different drugs. 4. Impact on caregiver burden: No clear differences among the drugs, although daily injections may be burdensome if a caregiver is required to perform the injection. 5. Impact on spread of infectious disease: Not applicable. 6. New mechanism of action that is likely to help patients who have not responded to other treatments: Abaloparatide acts through a similar mechanism as teriparatide. However, both anabolic drugs work through a fundamentally different mechanism from the other available agents, including zoledronic acid. There is evidence that starting with an anabolic agent followed by an antiresorptive agent may result in greater long-term fracture prevention than treating with an antiresorptive agent for the same length of time. However, to date, there are no published randomized trials demonstrating that this is the optimal approach. 7. Severity of the untreated condition: Based upon fracture outcomes in controlled trials, no clear differences among the different drugs 8. Lifetime burden of illness: No clear differences among the different drugs 9. Lack of availability of any previous treatment for the condition: There are existing anabolic and anti-resorptive treatments for osteoporosis. ©Institute for Clinical and Economic Review, 2017 Page 34 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 10. Other ethical, legal, or social considerations that might strongly influence the overall value of an intervention to patients, families, and caregivers, the health system, or society: There are no clear differences among the drugs. ©Institute for Clinical and Economic Review, 2017 Page 35 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 6. Economic Analyses 6.1 Long-Term Cost-Effectiveness Overview We conducted a cost-effectiveness analysis using a simulation model comparing the two FDA- approved anabolic therapies, abaloparatide and teriparatide, to treatment with a bisphosphonate (zoledronic acid) in a representative cohort of postmenopausal women who are at high risk for osteoporotic fractures. Zoledronic acid was chosen as a comparator because 1) it is commonly used, 2) adherence with treatment is significantly higher than with oral bisphosphonates, 3) patients at higher risk of fracture are recommended to receive this drug, and 4) clinical experts indicated it was the most appropriate comparator. We estimated the costs, numbers of fractures, quality-adjusted life years (QALY) gained, life-years gained and the incremental cost-effectiveness of the anabolic agents relative to zoledronic acid, using estimates of relevant clinical parameters from trial data and estimates of drug and other related health care costs. Model outcomes of interest include: • Incidence of clinical vertebral, hip, and all other non-vertebral fractures • Life expectancy • Quality-adjusted life-years (QALYs) • Osteoporosis drug treatment costs • Fracture costs • Total costs • Costs per QALY gained Cost-Effectiveness Model: Methods Model Structure The primary aim of this analysis was to estimate the cost-effectiveness of treatments indicated for the prevention of osteoporotic fractures in postmenopausal women who have not been recently treated for osteoporosis, but have an indication for treatment to prevent osteoporotic fractures. The model adopted a health care system perspective. This de novo model was built in Microsoft Excel and the model structure is depicted in Figure 3, and is based in part on a literature review of prior published models of osteoporosis.80 A representative cohort of patients at high fracture risk who are untreated or have not recently received treatment were modeled from treatment initiation until death. Patients transitioned between health states during one-year cycles over a lifetime time ©Institute for Clinical and Economic Review, 2017 Page 36 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents horizon. The model used a 3% discount rate for costs and health outcomes, and costs are presented in 2016 US dollars. The model consists of several health states, including osteoporosis without a new fracture (the origination state for patients entering the model), clinical vertebral fracture, hip fracture, other (i.e., non-hip) non-vertebral fracture, and death; fracture health states reflect those examined in the network meta-analysis (NMA). Patients may also have a morphometric vertebral fracture but we assumed they do not change health states, due to the negligible cost and QALY impacts of morphological vertebral fractures; we explored a potential QALY loss for these patients in a scenario analysis. Patients enter an acute fracture health state for one year upon experiencing a new fracture; after one year, patients transition to a post-fracture health state, where they remain until they transition to a subsequent fracture or death. Once they enter a post-fracture health state, patients may only transition to a worse subsequent fracture or die, so that patients who experience a serious fracture do not forfeit the long-term costs and utilities associated with it by transitioning to a less severe fracture in the “memory-less” Markov model framework. Given this constraint, we calibrated the model so that the cumulative fracture probabilities for an untreated (baseline) population were representative of expected rates for each fracture type, as described below. The assumed hierarchy of fracture severity is hip > vertebral > other. Figure 3. Markov Model Structure for Osteoporosis Patients Target Population The population of focus is postmenopausal women who are untreated or have not been recently treated for osteoporosis, but who have an indication for treatment to prevent osteoporotic fractures. In our base-case analysis, we assumed the fracture risk was similar to that observed in ©Institute for Clinical and Economic Review, 2017 Page 37 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents the clinical trials of the anabolic agents; this estimate was varied in a scenario analysis. Patients were assumed to enter the cohort at age 70 years, based on the demographic data from the pivotal trials of the anabolic agents, in which average age was 68.8 – 70.9 years.17-19 Key Model Characteristics The model utilizes results from the NMAs in the evidence review as well as imputed values for hip fracture (see Table 11 for a detailed explanation) as the effectiveness estimates for fracture prevention for each drug regimen (Table 12). We applied the relative risk estimates derived from the NMA and the imputed hip fracture relative risks to the baseline fracture probabilities, which were derived from a combination of clinical trials, the published literature, and the FRAX Fracture Risk Assessment Tool.17-19,31,81,82 Survival time in each health state was weighted by published health state-specific utilities to model health-related quality of life. The model includes separate utilities for the different types of fractures.83 Patient mortality was based on US background age- related mortality estimates for females; hip fracture was assumed to increase the risk of mortality. The model includes treatment costs associated with each individual regimen, including drug acquisition costs, administration costs, and acute care costs for fractures. The base-case analysis uses a health care system perspective (i.e., focuses on direct medical care costs only). All costs and health outcomes were discounted by 3% per year.84 Key Model Assumptions Table 11 contains a list of key model assumptions along with the rationale for each assumption. ©Institute for Clinical and Economic Review, 2017 Page 38 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table 11. Key Model Assumptions Assumption Rationale From a post-fracture state, patients can transition to a worse Prevents patients who experience more serious fractures fracture state only (or death). The hierarchy for fracture severity from forfeiting serious fracture states’ associated long-term is hip > vertebral > other. costs and utilities by transitioning to a less severe fracture in the “memory-less” Markov model framework. Subject to the fracture hierarchy, patients may have an unlimited Real-world patients may experience any number of number of fractures over the modeled time horizon. fractures. Hip fracture relative risk estimates for anabolic drugs were based Trial-observed hip fractures were rare for teriparatide and on the ratio of hip fracture relative risk versus non-vertebral abaloparatide; thus, the NMA results for hip-only fractures fracture relative risk reported in the HORIZON trial (zoledronic were unstable and lacked face validity. Studies of acid vs. placebo). Briefly, the HORIZON-derived ratio was 0.59 osteoporosis drugs with adequate power to assess hip (hip) / 0.75 (non-vertebral) = 0.79, which was multiplied by the fractures consistently find that the reduction in hip NMA-derived relative risks for non-vertebral fractures fractures lies between that for vertebral fractures and non- (abaloparatide = 0.51, teriparatide = 0.61) to obtain base case vertebral fractures (see Appendix Table E6 for estimates. We then derived confidence intervals for sensitivity romosozumab and zoledronic acid). analyses based on the NMA-derived non-vertebral fracture ranges for each drug. We used the HORIZON trial’s zoledronic acid relative risk (and confidence interval) for hip fracture directly. We did not model serious adverse events in the base case Anabolic regimens as well as zoledronic acid exhibited analysis. We explored the impact of infusion reactions stemming similar serious adverse event rates compared to placebo from IV infusion of zoledronic acid in a scenario analysis based on and each other in their respective trials. These small event the approach of a previous cost-effectiveness analysis of rate differences are unlikely to impact cost-effectiveness bisphosphonates.23 results. Anabolic therapies are administered for a duration of two years Anabolic treatment duration: FDA label according to their labeled indication, and are followed by six Zoledronic acid treatment duration: AACE guidelines state years of zoledronic acid. We assume that time to benefit for that patients at high risk should be treated for six years. anabolic agents and zoledronic acid is immediate and that 100% The HORIZON Extension Trials demonstrated added efficacy anabolic efficacy is maintained throughout the anabolic and post- for six but not nine years of therapy and maintenance of anabolic zoledronic acid periods, plus an additional 3 years, then efficacy for three years following treatment cessation.85,86 efficacy declines linearly to a relative risk of 1 over a period of 10 Time to benefit: Data show that the benefit of treatment is years. immediate for hip fracture and for any clinical fracture. Bisphosphonate therapy with zoledronic acid is administered for Efficacy maintenance: Expert opinion. six years, 100% efficacy is maintained throughout the six-year Efficacy decline: Parity with previous cost-effectiveness administration period plus an additional three years, then models that model a decline over time. efficacy declines linearly to a relative risk of 1 over a period of 10 years. All comparators’ adherence rates were 100% in base case Lack of real-world adherence data for abaloparatide, and on analysis. the impact of lower adherence on efficacy for abaloparatide and teriparatide. We applied a 27% discount to the WAC price of abaloparatide. Net price information for abaloparatide is not yet available. The average industry-wide reduction (including discounts, rebates and other price concessions) for brand drugs is 27.1%.33 ©Institute for Clinical and Economic Review, 2017 Page 39 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Treatment Strategies The interventions assessed in the model were the same as those in the clinical evidence review (abaloparatide, teriparatide, and zoledronic acid). We compared the anabolic agents to zoledronic acid in the base-case analysis, which allowed us to evaluate the relative incremental benefits and harms of these agents when used first-line in patients with risk for fragility fractures. All patients received treatment upon entering the model. Patients received anabolic therapy for two years, immediately followed by six years of therapy with zoledronic acid. We assumed all therapies had 100% efficacy throughout the treatment regimen (i.e., no efficacy ramp-up time), and that the anabolic therapies’ efficacy was maintained throughout the zoledronic acid administration period plus three years (i.e., 11 years total of 100% anabolic efficacy), before declining to a relative risk of 1.0 over a 10-year period. For zoledronic acid, we modeled a three-year efficacy maintenance period after the administration period ended (i.e., nine years total of 100% zoledronic acid efficacy), followed by an efficacy decline over 10 years. We also assumed 100% treatment adherence for all agents. Figure 4 represents an example of treatment sequencing and effect over time for hip fractures; the same approach was applied to clinical vertebral and to other non- vertebral fractures. We explored the impacts of our assumptions regarding efficacy onset, maintenance, and decline in scenario analyses. Figure 4. Treatment Sequencing and Effect Over Time for Hip Fractures Note: Each treatment line is color-coded to match the X-axis labels at the top of the chart; vertical black lines indicate transitions to the next stage in sequence/efficacy. Line placement is not exact. Fx: fracture, RR: relative risk, Tx: treatment ©Institute for Clinical and Economic Review, 2017 Page 40 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Clinical Inputs Annual relative risks of fracture for each drug (Table 12) were derived from 1) the NMA (vertebral and non-vertebral fractures) and 2) the ratio of hip to non-vertebral fracture relative risks reported in the HORIZON trial (see Table 11 for additional hip fracture relative risk explanation); each relative risk estimate represents the differential risk of fracture versus placebo per year. Note that the NMA relative risk estimates for vertebral fracture include both clinical and morphometric vertebral fractures; because our modeled health states include only clinical vertebral fracture, while morphometric vertebral fracture patients are assumed to remain in their current health states, we assumed a 35% proportion of overall vertebral fractures were clinical vertebral fractures, reflecting the results of a retrospective cohort analysis.31 In probabilistic sensitivity analyses (PSA), relative risk estimates were varied using a log-normal distribution; uncertainty in the proportion of vertebral fractures that were clinical was modeled as ±20% with a beta distribution. Table 12. Fracture Relative Risk Parameters Model Input Default Lower Upper Source Zoledronic acid (baseline) Hip Fracture 0.59 0.42 0.83 Black et al., HORIZON trial20 Vertebral Fracture (all)* 0.30 0.24 0.37 NMA Other Non-Vertebral Fractures 0.75 0.64 0.87 NMA Teriparatide Hip Fracture 0.48 0.28 0.75 Derived from NMA and HORIZON Vertebral Fracture (all)* 0.17 0.09 0.29 NMA Other Non-Vertebral Fractures 0.61 0.41 0.88 NMA Abaloparatide Hip Fracture 0.40 0.17 0.74 Derived from NMA and HORIZON Vertebral Fracture (all)* 0.13 0.03 0.33 NMA Other Non-Vertebral Fractures 0.51 0.28 0.85 NMA *Relative risks for vertebral fractures were estimated from studies including morphometric vertebral fractures; 35% of estimated vertebral fractures were modeled as clinical vertebral fractures. Baseline Fracture Inputs The relative risk estimates from the NMA were applied to age-stratified baseline (placebo) estimates of annual probability of fracture to derive each comparator’s annual fracture probabilities (Table 13). We derived the age-stratified baseline annual fracture probabilities by calculating the fracture risks of an average 70-year old patient from the pooled placebo arms of the Fracture Prevention, ACTIVE, FRAME, and HORIZON trials. Briefly, we summed 1) the number of each fracture type (hip, vertebral, and non-vertebral) from the trials, as well as 2) the fracture types’ associated follow-up time in person-years, then calculated annualized rates of each fracture type. These annualized rates were then converted to annual probabilities that we used as the baseline fracture probabilities for patients age 70-74 years. ©Institute for Clinical and Economic Review, 2017 Page 41 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents To model increasing fracture risk as patients age, we extrapolated pooled estimates based on 1) previously published age-stratified fracture estimates31 and 2) the 10-year probability of fracture based on FRAX Fracture Risk Assessment Tool output for a 70-year-old US Caucasian woman with a T-score of -3; note that FRAX was only utilized for the derivation of hip fracture estimates.83 We used data reported by Melton et al. for age-weighted estimates of the increasing risk of fracture over time. These estimates are for a mixed population of people with and without a prior fracture, so are higher than for someone who has never had a fracture but somewhat lower than for someone who has. We then calibrated hip fracture estimates so the modeled 10-year cumulative incidence of hip fracture matched the FRAX 10-year probability of hip fracture (9.5%). Each resultant estimate was varied by ±20% in sensitivity analyses. Annual probabilities were linearly interpolated from the five-year estimates. All baseline fracture parameters were varied using a beta distribution in the PSA. Table 13. Baseline (Placebo) Annual Fracture Probabilities by Age Strata Fracture and Age (in years) Default Lower Upper Source Groups Hip Fracture Age 70-74 0.006 0.005 0.007 Pooled trials Age 75-79 0.011 0.009 0.013 Pooled trials & Melton31/FRAX extrapolation Age 80-84 0.023 0.019 0.028 Pooled trials & Melton/FRAX extrapolation Age 85+ 0.031 0.025 0.038 Pooled trials & Melton/FRAX extrapolation Vertebral Fracture (Clinical and Morphometric) Age 70-74 0.034 0.027 0.041 Pooled trials Age 75-79 0.046 0.037 0.055 Pooled trials & Melton extrapolation Age 80-84 0.076 0.061 0.091 Pooled trials & Melton extrapolation Age 85+ 0.091 0.074 0.111 Pooled trials & Melton extrapolation Other Non-Vertebral Fracture Age 70-74 0.024 0.019 0.029 Pooled trials Age 75-79 0.037 0.030 0.044 Pooled trials & Melton extrapolation Age 80-84 0.053 0.042 0.063 Pooled trials & Melton extrapolation Age 85+ 0.079 0.063 0.095 Pooled trials & Melton extrapolation Fracture-Related Excess Mortality Inputs A review of studies reporting excess mortality following fractures showed that all but one study did not control for comorbidities. The study that did control for underlying health status found that excess mortality occurred after hip fractures (vertebral and non-vertebral fractures were not considered) at a rate roughly 50% lower than studies that adjusted for age and gender only.87 We therefore applied fracture-related excess mortality to hip fractures only, by applying the Tosteson formula (=[baseline probability *{hazard ratio-1}]/[baseline probability*{hazard ratio-1}+1]) to baseline hip fracture probabilities (Table 14). The excess mortality estimates were then added to ©Institute for Clinical and Economic Review, 2017 Page 42 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents the background mortality estimates of the US population at each model cycle for hip fracture patients.83 All excess mortality parameters were varied using a log-normal distribution in a PSA. Table 14. Absolute Mortality Increase for Hip Fracture Age Range Default Lower Upper Source Age 70-74 0.0025 0.0020 0.0029 Tosteson87 Age 75-79 0.0075 0.0060 0.0090 Tosteson Age 80-84 0.0336 0.0269 0.0403 Tosteson Age 85+ 0.0727 0.0581 0.0872 Tosteson Quality-of-Life Inputs Health state utilities were derived from publicly-available literature and/or manufacturer-submitted data, and applied to the fracture and post-fracture health states (Table 15).30,88-91 The baseline utility estimates for patients with no new fracture were from a study of the non-institutionalized US adult population for 7 health-related quality-of-life scores; we used the EuroQoL-5D (EQ-5D) age- stratified estimates for US women.30 We applied utility multipliers to baseline estimates for each fracture health state; utility multipliers were also derived using the EQ-5D index. The utility multipliers for vertebral fracture were applied to only 35% of patients with vertebral fracture, reflecting the proportion of these fractures that were clinical fractures in a retrospective cohort analysis;31 non-clinical vertebral fractures had no utility multiplier applied in the base case analysis, however we explored this assumption in a scenario analysis (see Appendix Figure F5). Health state utility values did not vary across treatments evaluated in the model. All utility parameters were varied using a beta distribution in the PSA. Table 15. Utility Values by Age Strata and Utility Multipliers Model Input Default Lower Upper Source General Population Utilities Age 70-79 0.770 0.616 0.924 Hanmer et al.30 Age 80+ 0.720 0.576 0.864 Hanmer et al. Utility Multipliers Hip Fracture Year 1 0.700 0.560 0.840 Peasgood et al.90 Hip Fracture Year 2+ 0.800 0.640 0.960 Peasgood et al. Clinical Vertebral Fracture 0.590 0.472 0.708 Peasgood et al. Year 1 Clinical Vertebral Fracture 0.931 0.745 1.000 Kanis/Oleksik et al.88,89 Year 2+ Other Non-Vertebral 0.902 0.722 1.000 Burstrom et al.91 Fracture Year 1 Other Non-Vertebral 1.000 0.800 1.000 Assumption Fracture Year 2+ ©Institute for Clinical and Economic Review, 2017 Page 43 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Drug Cost Inputs We used the average wholesale acquisition cost (WAC) for generic zoledronic acid and assumed that treatment was administered for six years.92 For the price of a teriparatide pen, we obtained data from SSR Health that combined information on net US dollar sales through the first quarter of 2017 with information on unit sales to derive net pricing at the unit level across all payer types.32 We estimated net prices by comparing the four-quarter rolling averages (i.e., second quarter of 2016 through first quarter of 2017) of both net prices and WAC per unit to arrive at a mean discount from WAC for the drug. Finally, we applied this average discount to the WAC as of June 201792 to arrive at an estimated net price per pen. The derived discount for teriparatide was 38%, which was then applied to the WAC for a 2.4 ml (250 mcg/ml) package that resulted in a net price of $1,866.34 per pen. This discount may not reflect the negotiated price for any one payer, but rather the average discount across all payers. Each teriparatide pen contains 28 doses, so patients use approximately 13 pens per year. For abaloparatide, we used the announced list price of $1,625 per pen (as of June 1, 2017) and applied a 27% discount, representing the average industry-wide discount on brand drugs.33,93 Each abaloparatide pen contains 30 doses, so patients use approximately 12 pens per year. In addition, threshold analyses on these costs are provided in the results section of this chapter. All drug costs were varied by ±20% using a normal distribution in the PSA. Table 16. Drug Cost Inputs Base- Acquisition Drug Name, Labeled Dose, Strength WAC/Pen Net Price* Case Tx Cost Per Tx Administration Route (Pen Size) Duration Course† Teriparatide 20 mcg SC QD 250 mcg/ml $2,997.90 $1,866.34‡ 2 years $48,691 (2.4 ml) Abaloparatide 80 mcg SC QD 3,120 $1,625 $1,186.25§ 2 years $29,312 mcg/1.56 ml Zoledronic Acid 5 mg IV Q year 5 mg/100 ml $306 # $306# 6 years $1,837 IV: intravenous, SC: subcutaneous, QD: once daily, Q mo: once monthly, Q year: once yearly, Tx: treatment, WAC: wholesale acquisition cost *Net price is the estimated price after discounts and rebates from WAC. No discounts have been applied to generic zoledronic acid. †Acquisition cost of initial drug using net price (or average generic WAC for zoledronic acid) and assuming full course of treatment; costs would be lower if a modeled patient died before completing a course of therapy. Costs do not include the additional costs of post-anabolic zoledronic acid therapy. ‡Price per pen including 38% discount §Price per pen based on announced list price and assumed 27% discount #Annual dose cost based on average generic WAC ©Institute for Clinical and Economic Review, 2017 Page 44 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Healthcare Cost Inputs Fracture-related healthcare costs were derived from publicly available literature, and applied to the fracture and post-fracture health states (Table 17).94,95 All cost estimates were from US cohort studies in representative populations, and inflated to 2016 US dollars. Costs for vertebral fracture were applied to only 35% of patients, reflecting the proportion of clinical vertebral fractures in a retrospective cohort;31 non-clinical vertebral fractures had no fracture-related costs applied. We modeled administration cost for zoledronic acid intravenous administration ($168)96 but not for anabolic drugs as they are self-administered. We assumed supportive care costs were similar among comparators and thus would not contribute to cost differences. All healthcare costs were varied by ±20% using a log-normal distribution in the PSA. Table 17. Acute and Long-Term Annual Fracture Costs Model Input Default Lower Upper Source Hip Fracture Cost $44,395 $35,516 $53,274 Bonafede94 Post-Hip Fracture Annual Cost $10,835 $8,668 $13,002 Parthan95 Clinical Vertebral Fracture Cost $27,906 $22,325 $33,487 Bonafede Post-Clinical Vertebral Fracture Annual Cost $309 $247 $371 Parthan Other Non-Vertebral Fracture Cost $12,764 $10,211 $15,317 Bonafede Post-Other Non-Vertebral Fracture Annual Cost $0 $0 $0 Assumption Sensitivity Analyses We ran one-way sensitivity analyses to identify the key drivers of model outcomes. One-way sensitivity analyses used 95% confidence intervals from clinical evidence where available. When 95% confidence intervals were not available, uncertainty ranges were varied by ±20%. We also conducted a PSA by jointly varying all model parameters over 5,000 simulations, then calculating 95% credible range estimates for each model outcome (Appendix Tables F1-F2). In addition, we also conducted scenario analyses to explore the impacts of our assumptions on model results, by varying: 1. Baseline fracture risk probabilities, by increasing the baseline fracture rates by up to 100%. 2. Years of maintenance of full treatment effect after stopping zoledronic acid treatment. 3. The duration of the efficacy decline for anabolic agents, including no decline over lifetime horizon. 4. Zoledronic acid ramp-up time to full efficacy (base case was full efficacy throughout). 5. Comparison to no treatment, rather than to zoledronic acid. 6. NMA relative risk estimates by excluding open-label teriparatide data from the ACTIVE trial. 7. Increased relative risk of subsequent fracture. 8. Excess mortality risk after vertebral and other non-hip fractures ©Institute for Clinical and Economic Review, 2017 Page 45 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 9. Duration of teriparatide and abaloparatide therapy as studied in the trials (i.e., 21 and 18 months, respectively, vs. the labeled indication of two years in the base case). 10. The inclusion of a disutility for zoledronic acid infusion-related reaction (adverse event). 11. The inclusion of a disutility for morphometric vertebral fractures 12. Baseline fracture risk in higher risk patient groups who cannot tolerate zoledronic acid. 13. Inclusion of excess mortality risk after vertebral and other (i.e., non-hip) non-vertebral fractures 14. Efficacy estimates from NMA including VERO-trial data i 15. Baseline fracture risks from a retrospective observational cohort study using US administrative claims data from fee-for-service Medicare beneficiaries i Cost-Effectiveness Model: Results Base-Case Results The anabolic therapies resulted in increased costs, QALYs, and life years compared to zoledronic acid (Table 18). The QALYs gained versus zoledronic acid were 0.066 for abaloparatide and 0.046 for teriparatide over the lifetime horizon (Table 19). Incremental costs were $22,061 for abaloparatide and $43,440 for teriparatide. The base case incremental cost-effectiveness ratios (ICERs) for each anabolic drug compared to zoledronic acid far exceeded the commonly-cited cost- effectiveness threshold of $150,000 per QALY (Table 19). Table 18. Base-Case Results Regimen Cost QALYs Life Years Zoledronic acid $25,465 8.933 12.188 Teriparatide $68,905 8.979 12.193 Abaloparatide $47,525 8.999 12.195 QALY: quality-adjusted life year Table 19. Pairwise Results for Anabolic Therapies Compared to Zoledronic Acid Regimen Incr. Cost Incr. QALYs Incr. LYs ICER vs. Zoledronic Acid Teriparatide $43,440 0.046 0.005 $941,537 Abaloparatide $22,061 0.066 0.007 $333,892 ICER: incremental cost-effectiveness ratio, Incr.: incremental, LY: life year, QALY: quality-adjusted life year Appendix Tables F1-F2 provide additional detail regarding the model findings. First, there were moderate cost offsets compared to zoledronic acid due to fracture prevention, ranging from i These scenario analyses were added after the June 30 public meeting to review the findings of the Evidence Report that was released on June 16, and were not considered by the CTAF Panel during their votes (see Section 7 for details of the public meeting). ©Institute for Clinical and Economic Review, 2017 Page 46 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents approximately -$5,300 for abaloparatide versus zoledronic acid to approximately -$3,500 for teriparatide versus zoledronic acid. These cost savings from prevention of fractures offset approximately 7-19% of the anabolic drug costs. The benefits to patients (measured in QALYs) resulted from small contributions across hip, clinical vertebral, and other non-vertebral fractures; however, because fracture events are relatively rare, most QALYs for each regimen are accrued by patients who remain in the “no new fracture” health state. In general, the modest clinical differences of the anabolic agents compared to zoledronic acid were not large enough to offset the cost increases. Probabilistic sensitivity analysis indicated that our ICER results are highly uncertain, but the probability that the ICERs for the anabolic therapies were below $150,000 per QALY gained were either low (abaloparatide: 7.1%) or zero (teriparatide) (Figure 5). This was primarily due to the small QALY gains and higher prices of anabolics versus zoledronic acid. Figure 5. Cost-effectiveness Acceptability Curve for Anabolic Agents Compared to Zoledronic Acid One-way Sensitivity Analyses Detailed findings from the one-way sensitivity analyses of model inputs for anabolic agents versus zoledronic acid can be found in Figures 6-7. Parameters associated with hip fractures were by far the largest contributors to uncertainty for abaloparatide and teriparatide versus zoledronic acid, particularly the anabolics’ relative risks for hip fracture (the most expensive and severe of the fracture types) as they approached 1.0 (i.e., no efficacy vs. untreated patients). Results were also sensitive to uncertainty in the long-term utility multipliers and drug costs. None of the modeled parameters’ range values resulted in an ICER less than $150,000 per QALY gained. (Negative ICERs shown below result from negative incremental QALYs vs. zoledronic acid.) ©Institute for Clinical and Economic Review, 2017 Page 47 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Figure 6. One-way Sensitivity Analysis: Incremental Cost-Effectiveness Ratios for Teriparatide Versus Zoledronic Acid -$18M -$15M -$12M -$9M -$6M -$3M $0 $3M Parameter Low Value High Value Low Result High Result Spread Teriparatide Relative Risk: Hip Fracture 0.280 0.750 $479,590 -$16,532,705 $17,012,296 Zoledronic Acid Relative Risk: Hip Fracture 0.420 0.830 $2,384,332 $482,093 $1,902,239 Utility Multiplier: Hip Fracture Year 2+ 0.640 0.960 $700,502 $1,435,465 $734,963 Cost: Teriparatide 600 mcg/2.4mL pen $1,493 $2,240 $737,109 $1,145,966 $408,857 Teriparatide Relative Risk: Vertebral Fracture 0.090 0.290 $833,277 $1,165,710 $332,433 Utility Multiplier: Other Fracture Year 2+ 0.800 1.000 $637,548 $941,537 $303,990 Teriparatide Relative Risk: Other Fracture 0.410 0.880 $840,319 $1,110,995 $270,675 General Population Utility: Age 80+ 0.576 0.864 $1,081,405 $833,707 $247,698 Utility Multiplier: Vertebral Fracture Year 2+ 0.745 1.000 $781,722 $1,018,714 $236,992 Utility Multiplier: Other Fracture Year 1 0.722 1.000 $838,241 $1,009,089 $170,848 Figure 7. One-way Sensitivity Analysis: Incremental Cost-Effectiveness Ratios for Abaloparatide Versus Zoledronic Acid $0 $2M $4M $6M $8M Parameter Low Value High Value Low Result High Result Spread Abaloparatide Relative Risk: Hip Fracture 0.171 0.741 $165,144 $7,807,367 $7,642,223 Zoledronic Acid Relative Risk: Hip Fracture 0.420 0.830 $624,585 $181,708 $442,877 Utility Multiplier: Hip Fracture Year 2+ 0.640 0.960 $243,159 $532,644 $289,485 Cost/pen: Abaloparatide $949 $1,424 $250,425 $417,358 $166,933 Abaloparatide Relative Risk: Vertebral Fracture 0.030 0.330 $298,026 $435,752 $137,726 Utility Multiplier: Other Fracture Year 2+ 0.800 1.000 $197,820 $333,892 $136,072 Abaloparatide Relative Risk: Other Fracture 0.280 0.850 $293,843 $405,122 $111,279 General Population Utility: Age 80+ 0.576 0.864 $381,325 $296,954 $84,371 Utility Multiplier: Other Fracture Year 1 0.722 1.000 $289,681 $364,076 $74,395 Utility Multiplier: Vertebral Fracture Year 2+ 0.745 1.000 $284,412 $356,900 $72,488 ©Institute for Clinical and Economic Review, 2017 Page 48 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Scenario Analyses Below, we report the results of the most relevant or influential scenario analyses. Results from the scenario analyses pertaining to treatment efficacy ramp-up, maintenance, and rates of decline as well as reduced quality of life after morphometric fractures, did not yield major differences in conclusions from the base case and can be found in Appendix Figures F1-F5 and Tables F3-F10. In general, because efficacy maintenance was tied to the use of post-anabolic zoledronic acid, and because assumptions about zoledronic acid in both the baseline comparator arm and the anabolic arms were similar, changes in these parameters tended to impact all three arms similarly; thus, the small incremental QALY differences between anabolic drugs and zoledronic acid were relatively consistent with each scenario iteration. None of these scenarios produced a large enough QALY difference to lower the ICER below $150,000 per QALY. Higher-Risk Group To model patient populations with a higher baseline risk of fracture than in the base-case, which was derived from the key clinical trials, a historical cohort,31 and the online FRAX tool97 we increased the age-dependent baseline fracture risks up to 100% of their base-case value. The corresponding incremental cost-effectiveness ratios for each anabolic treatment in such higher-risk patient populations is shown in Figure 8 below. Annual fracture risks must be approximately 118% higher for abaloparatide to approach the $150,000 per QALY threshold (i.e. 0.013 for hip, 0.075 for vertebral, and 0.052 for other non-vertebral fractures annually). This corresponds with a cohort in which the average 70-year-old has the approximate risk of an 85-year-old woman with a T-score of - 4.0. Teriparatide did not approach commonly-cited cost-effectiveness thresholds until a greater than 1,000% increased risk of fracture was applied. Figure 8. Results of Higher Baseline Fracture Risk Scenario Analysis $1,000,000 $900,000 $800,000 $700,000 $600,000 ICER $500,000 Teriparatide $400,000 Abaloparatide $300,000 $200,000 $100,000 $0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Baseline Fracture Risk Increase ©Institute for Clinical and Economic Review, 2017 Page 49 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Comparisons to No Treatment We also considered a scenario in which patients may not be able to take zoledronic acid, and thus the comparator is no treatment. To do this, we compared the anabolics to a baseline fracture risk population (i.e. no relative risks were applied to baseline fracture estimates), and assumed the anabolic-treated patients did not receive zoledronic acid or its efficacy maintenance benefits following initial anabolic therapy. We assumed anabolic efficacy linearly declined to 1.0 (i.e. no efficacy vs. placebo) over three years in the absence of zoledronic acid. The results for this scenario are shown below in Tables 20 and 21. In this scenario, incremental QALYs decreased due to the shortened efficacy time window for the anabolics, and none of the treatments reached the $150,000 per QALY threshold. None of the comparators approached commonly-cited cost- effectiveness thresholds when we varied our assumption of the number of years of efficacy decline from three years up to 10 years (as in the base case analysis). Table 20. Results of Scenario Analysis Comparing Anabolic Drugs to No Treatment Regimen Cost QALYs Life Years No Treatment $30,038 8.825 12.181 Teriparatide $73,162 8.886 12.182 Abaloparatide $52,919 8.893 12.183 QALY: quality-adjusted life year Table 21. Pairwise Results of Anabolic Drugs Compared to No Treatment Incr. Regimen Incr. QALYs Incr. LYs ICER vs. No Treatment Cost Teriparatide $43,124 0.060 0.002 $715,878 Abaloparatide $22,881 0.067 0.002 $339,027 Incr.: incremental, LY: life year, QALY: quality-adjusted life year Increased Refracture Risk In this scenario analysis, we further increased the refracture risk from baseline, using published estimates (Table 22)98 to explore the impact on model results. ©Institute for Clinical and Economic Review, 2017 Page 50 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table 22. Relative Risk of Subsequent Fracture for Scenario Analysis of Increased Refracture Risk Model Input Default Source Post-Hip Fracture RR: Hip Fracture 2.30 Klotzbuecher98 Post-Vertebral Fracture RR: Hip Fracture 2.30 Klotzbuecher Post-Vertebral Fracture RR: Vertebral Fracture 4.40 Klotzbuecher Post-Other Non-Vertebral Fracture RR: Hip Fracture 1.90 Klotzbuecher Post-Other Non-Vertebral Fracture RR: Vertebral Fracture 1.70 Klotzbuecher Post-Other Non-Vertebral Fracture RR: Other Fracture 3.30 Klotzbuecher Increasing refracture risks impacted the results by slightly amplifying the differences in relative risk parameters between the anabolic agents and zoledronic acid, resulting in modest improvements in incremental QALYs and cost (Table 24); however, none of these improvements were sufficient to make the incremental cost-effectiveness ratios for anabolic agents fall below $150,000 per QALY. Table 23. Results When Including Increased Refracture Risk Regimen Cost QALYs Life Years Zoledronic Acid $32,129 8.864 12.170 Teriparatide $74,109 8.927 12.179 Abaloparatide $52,078 8.953 12.183 QALY: quality-adjusted life year Table 24. Pairwise Results of Anabolic Drugs Compared to Zoledronic Acid When Including Increased Refracture Risk Regimen Incr. Cost Incr. QALYs Incr. LYs ICER vs. Zoledronic Acid Teriparatide $41,980 0.063 0.010 $662,149 Abaloparatide $19,949 0.090 0.013 $222,548 Incr.: incremental, LY: life year, QALY: quality-adjusted life year Duration of Teriparatide and Abaloparatide Therapy as Studied in the Trials In the Fracture Prevention Trial (teriparatide) and the ACTIVE trial (abaloparatide), patients were treated for 21 months and 18 months, respectively; however, their FDA labels both recommend treatment for up to two years, and we used the labeled indication in the base-case analysis. In this scenario analysis, we modeled the respective trial treatment durations, but assumed that efficacy was maintained for the entire two years. Zoledronic acid was administered for six years in both arms, as in the base case. This scenario showed lowered anabolic cost of both regimens, but did not impact QALYs because our assumptions regarding efficacy maintenance were unchanged from the base case. ©Institute for Clinical and Economic Review, 2017 Page 51 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table 25. Results When Using Trial-Reported Time on Treatment for Anabolics Regimen Cost QALYs Life Years Zoledronic Acid $25,465 8.933 12.188 Teriparatide $63,486 8.979 12.193 Abaloparatide $41,020 8.999 12.195 QALY: quality-adjusted life year Table 26. Pairwise Results When Using Trial-Reported Time on Treatment for Anabolics Regimen Incr. Cost Incr. QALYs Incr. LYs ICER vs. Zoledronic Acid Teriparatide $38,021 0.046 0.005 $824,094 Abaloparatide $15,555 0.066 0.007 $235,430 Incr.: incremental, LY: life year, QALY: quality-adjusted life year Excess Mortality after Vertebral and Other Non-vertebral Fractures This scenario considers the impact of excess mortality after vertebral and other non-vertebral fractures on the incremental cost-effectiveness by multiplying the age-stratified background mortality of the US population by age-stratified relative risks for excess mortality. We obtained age-stratified estimates of relative risk for excess mortality from Johnell et al. These estimates are adjusted for age and gender, but not underlying health status. Tosteson et al. showed that adjusting for underlying health status in addition to age and gender reduced the relative risk estimates for excess mortality after hip fractures by 50% (no data were reported for vertebral and other non-hip fractures). Consistent with that finding we adjusted the published relative risk estimates of Johnell et al., by a factor 0.5 to account for underlying health status and applied these adjusted relative risk estimates to the age-stratified background mortality of the US population.99 Table 27. Results When Considering Excess Mortality after Vertebral and Other Non-vertebral Fractures Regimen Cost QALYs Life Years Zoledronic Acid $25,051 8.819 12.027 Teriparatide $68,588 8.886 12.062 Abaloparatide $47,260 8.917 12.080 QALY: quality-adjusted life year Table 28. Pairwise Results When Considering Excess Mortality after Vertebral and Other Non- vertebral Fractures Regimen Incr. Cost Incr. QALYs Incr. LYs ICER vs. No Treatment Teriparatide $43,537 0.067 0.035 $649,845 Abaloparatide $22,209 0.098 0.053 $226,259 Incr.: incremental, LY: life year, QALY: quality-adjusted life year ©Institute for Clinical and Economic Review, 2017 Page 52 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Drug Price Threshold Analysis Prices for each drug that would achieve commonly-cited cost-effectiveness thresholds ranging from $50,000 to $150,000 per QALY gained are presented in Table 29, along with net price per pen. Table 29. Resulting Package Prices for Each Anabolic Therapy to Reach Cost per QALY Thresholds Drug Base-Case Cost $50,000/QALY $100,000/QALY $150,000/QALY Teriparatide $1,866.34 $238.47 $329.77 $421.07 (cost per pen) Abaloparatide $1,186.25 $379.30 $521.42 $663.55 (cost per pen) QALY: quality-adjusted life year Model Validation and Prior Published Evidence on Costs and Cost-Effectiveness Model validation followed standard practices in the field. We tested all mathematical functions in the model to ensure they were consistent with the report (and supplemental Appendix materials). We also conducted sensitivity analyses with null input values to ensure the model produced findings consistent with expectations. Three independent modelers tested the mathematical functions in the model as well as the therapy-specific inputs and corresponding outputs. We also compared the ICER model to previously published models. We searched the literature to identify models that were similar to our own, with comparable populations, settings, perspective, and treatments. One manufacturer-funded study by Tosteson et al., compared teriparatide with bisphosphonates and no therapy in postmenopausal women eligible for osteoporosis treatment.34 Both the ICER and Tosteson model were structurally similar Markov models, with differences in the included therapies, modeled time-horizon, and certain model-specific inputs. Teriparatide was the common intervention in both studies. The bisphosphonates in the Tosteson model did not include zoledronic acid, which was included in the ICER model. Costs and QALYs for teriparatide were higher in the ICER model compared to the Tosteson model ($68,905 and 8.979 vs. $20,800 and 6.608, respectively). Several key differences between the two models contributed to the differences in results. 1) The ICER model adopted a lifetime time horizon while the Tosteson model time horizon was 10 years. The additional time in the ICER model contributed to the greater number of QALYs accrued and additional therapy costs. When treatment was modeled over a 10- year time span (results not shown), the ICER model showed QALY results that were similar to those in the Tosteson model. 2) The costs of therapy have increased substantially over time, with an annual teriparatide cost of approximately $6,300 in the Tosteson model versus approximately $24,350 in the ICER model. Additionally, the ICER model included bisphosphonate therapy post- anabolic therapy, and assumed that the full anabolic treatment effect was maintained by zoledronic ©Institute for Clinical and Economic Review, 2017 Page 53 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents acid for up to nine years after cessation of anabolic therapy. The Tosteson model assumed no residual treatment efficacy after treatment was completed. 3) The fracture probabilities in the Tosteson model were higher compared to the ICER model. 4) The base-case utilities in the ICER model were lower than in the Tosteson model. Additionally, utility multipliers and costs associated with vertebral fractures were applied to only 35% of the patients in the ICER model cohort to mirror the proportion of these fractures that were clinically apparent in a retrospective cohort analysis. 5) When comparing health state costs, the first-year costs for hip fracture were higher and first-year costs for vertebral fractures were lower in the ICER model compared to the Tosteson model. Subsequent-year fracture costs in the ICER model were higher. Furthermore, we calculated the cumulative lifetime risk of fracture, whereas Tosteson et al. calculated fracture risk over only a 10- year time horizon. The excess mortality inputs for hip fracture were similar in both studies, as the ICER inputs were derived from the Tosteson model. Other US-based models that we reviewed compared treatments that were not included in our analysis, so we did not conduct in-depth comparisons between these models and our own.95,100-104 We found one model by Murphy et al. that compared teriparatide to no treatment in Swedish osteoporosis patients who had a T-score of -3.0 or less.105 This model, which was run over a lifetime horizon with six-month cycles, resulted in incremental cost-effectiveness ratios of €5,897 per QALY ($7,990 per QALY) in patients with historical as well as incident vertebral fracture, and an incremental cost-effectiveness ratio of €18,701 per QALY gained ($25,340 per QALY) in those with only incident vertebral fractures. Compared to our model, these incremental cost-effectiveness results were significantly lower. One of the key drivers of the differences in the results between the two models is drug costs. When converted to US dollars, the annual cost of teriparatide in the Murphy et al., model was $7,290 (using 2011 currency exchange rates), while in the ICER model it was $21,243. Another key difference between the two models is in the assumed relative risk reduction of fractures compared to no treatment (0.17 for vertebral fractures and 0.47 for non- vertebral in Murphy et al. vs. 0.28 and 0.65, respectively, in the ICER model). This and other differences in the models resulted in a greater incremental QALY gain in the Murphy et al. analysis compared to the ICER model (0.189 vs. 0.019 QALYs). 6.2 Value-Based Benchmark Prices Our value-based benchmark prices for abaloparatide and teriparatide are presented in Table 30. As noted in the initial ICER methods document (http://icer- review.org/wpcontent/uploads/2016/02/Value-Assessment-Framework-slides-for-July-29-webinar- FINALcorrected-8-22-1.pdf), the value-based benchmark price for a drug is defined as the price range that would achieve incremental cost-effectiveness ratios between $100,000 and $150,000 per QALY gained. ©Institute for Clinical and Economic Review, 2017 Page 54 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents For both abaloparatide and teriparatide, the discounts required to meet both threshold prices are greater than the current discounts from WAC (assumed 27% for abaloparatide, 38% for teriparatide). Table 30. Value-based Benchmark Prices for Abaloparatide and Teriparatide for Osteoporosis Treatment in Postmenopausal Women Discount from Average Net Pen Price to Pen Price to WAC to reach WAC per Net Price* Price Within Drug Name Achieve Achieve $100,000 and Pen per Pen Benchmark $100,000/QALY $150,000/QALY $150,000/QALY Range? Threshold Teriparatide $2,997.90 $1,866.34‡ $329.77 $421.07 86% to 89% No Abaloparatide $1,625.00 $1,186.25† $521.42 $663.55 59% to 68% No QALY: quality-adjusted life year, WAC: wholesale acquisition cost, WTP: willingness to pay *Net price is the estimated price after discounts and rebates from WAC. † Price per pen based on announced list price and assumed 27% discount ‡ Price per pen including 38% discount 6.3 Potential Budget Impact We used the cost-effectiveness model to estimate the potential total budgetary impact of abaloparatide for postmenopausal women with osteoporosis and high risk of fracture. We used the WAC, an estimate of discounted WAC, and the three threshold prices for abaloparatide in our estimates of budget impact. Teriparatide was not included in this analysis because of its established presence in the market. Potential Budget Impact Model: Methods We used results from the same model employed for the cost-effectiveness analyses to estimate total potential budget impact. Potential budget impact was defined as the total differential cost of using the new therapy rather than relevant existing therapy for the treated population, calculated as differential health care costs (including drug costs) minus any offsets in these costs from averted health care events. All costs were undiscounted and estimated over one- and five-year time horizons. The five-year timeframe was of primary interest, given the potential for cost offsets to accrue over time and to allow a more realistic impact on the number of patients treated with the new therapy. The potential budget impact analysis included the entire candidate population for treatment, which consisted of postmenopausal women (assumed to be women over 50 years of age) diagnosed with osteoporosis and with a high risk of fractures. To estimate the size of the potential candidate population for treatment with abaloparatide, we first determined the number of women over 50 ©Institute for Clinical and Economic Review, 2017 Page 55 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents years of age in the US, approximately 62.6 million. Of those women, we assumed that 13% currently receive treatments for osteoporosis, based on a claims database analysis by Parthan et al., conducted to identify this percentage for a published budgetary impact analysis of denosumab in a hypothetical health plan.35 Of those receiving treatment, 66% were diagnosed with osteoporosis while the remainder were treated for osteopenia.35 We assumed that 46% of those women diagnosed and treated for osteoporosis had a high risk of osteoporotic fractures, based on occurrence of previous fractures and/or intolerance to previous osteoporosis treatment.35 This high-risk population was assumed to be eligible to receive treatment with abaloparatide. Applying these estimates to the projected 2017 US population resulted in an estimate of approximately 2.47 million eligible patients in the US. ICER’s methods for estimating potential budget impact are described in detail elsewhere and have recently been updated. The intent of our revised approach to budgetary impact is to document the percentage of patients that could be treated at selected prices without crossing a budget impact threshold that is aligned with overall growth in the US economy. Briefly, when we evaluate a new drug or device that would take market share from one or more drugs, we calculate the blended budget impact associated with displacing use of existing therapies with the new intervention. We assumed that abaloparatide would take market shares from teriparatide and zoledronic acid in a ratio of 80:20 (i.e., abaloparatide would take 80% from teriparatide and 20% from zoledronic acid). We tested the potential budget impact of abaloparatide by assuming different unit price points (WAC, discounted WAC, and the three cost- effectiveness threshold prices for abaloparatide) against the combination of the discounted WAC for teriparatide and the average generic price for zoledronic acid. Using this approach to estimate potential budget impact, we then compared our estimates to an updated budget impact threshold that represents a potential trigger for policy mechanisms to improve affordability, such as changes to pricing, payment, or patient eligibility. As described in ICER’s methods presentation, this threshold is based on an underlying assumption that health care costs should not grow much faster than growth in the overall national economy. From this foundational assumption, our potential budget impact threshold is derived using an estimate of growth in US gross domestic product (GDP) +1%, the average number of new drug approvals by the FDA over the most recent two-year period, and the contribution of spending on retail and facility- based drugs to total health care spending. Calculations are performed as shown in Table 31. For 2017-18, therefore, the five-year annualized potential budget impact threshold that should trigger policy actions to manage access and affordability is calculated to total approximately $915 million per year for new drugs. ©Institute for Clinical and Economic Review, 2017 Page 56 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table 31. Calculation of Potential Budget Impact Threshold Item Parameter Estimate Source 1 Growth in US GDP, 2017 (est.) +1% 3.20% World Bank, 2016 2 Total health care spending, 2016 ($) $2.71 trillion CMS NHE, 2014 3 Contribution of drug spending to total health 17.7% CMS National Health care spending (%) Expenditures (NHE), 2016; Altarum Institute, 2014 4 Contribution of drug spending to total health $479 billion Calculation care spending ($) (Row 2 x Row 3) 5 Annual threshold for net health care cost $15.3 billion Calculation growth for ALL new drugs (Row 1 x Row 4) 6 Average annual number of new molecular 33.5 FDA, 2016 entity approvals, 2013-2014 7 Annual threshold for average cost growth per $457.5 million Calculation individual new molecular entity (Row 5 ÷ Row 6) 8 Annual threshold for estimated potential $915 million Calculation budget impact for each individual new molecular entity (doubling of Row 7) Potential Budget Impact Model: Results Table 32 illustrates the per-patient budget impact results in more detail. Costs for abaloparatide were calculated using the WAC, discounted WAC, and threshold prices. The discounted WAC price of teriparatide and average WAC price for generic zoledronic acid were used to calculate costs for those treatments. When treating the eligible cohort with abaloparatide, the average potential budgetary impact (adjusted for differing periods of drug utilization and associated cost-offsets over the five-year period) resulted in cost-savings using the WAC, discounted WAC and across all three cost- effectiveness thresholds, ranging from approximately -$120 per patient using the WAC price ($1,625), to approximately -$10,500 per patient using the price to achieve $50,000 per QALY ($379). ©Institute for Clinical and Economic Review, 2017 Page 57 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table 32. Per-Patient Budget Impact Calculation Over a Five-year Time Horizon Average Annual Per-Patient Budget Impact Discounted WAC $150,000/QALY $100,000/QALY $50,000/QALY WAC Abaloparatide $13,952 $10,290 $5,928 $4,742 $3,556 Teriparatide + $14,072 Zoledronic acid* (Discounted WAC Only) Difference -$120† -$3,782† -$8,144† -$9,330† -$10,516† *Weighted in the ratio 80:20 for teriparatide:zoledronic acid †Indicates cost-saving N/A: not available, QALY: quality-adjusted life year, WAC: wholesale acquisition cost 6.4 Summary and Comment: Long-Term Cost Effectiveness and Potential Budget Impact We estimated the cost-effectiveness of anabolic treatments compared to zoledronic acid in patients with osteoporosis at high risk for fragility fractures. The cost per additional QALY was estimated to be above $150,000 per QALY for each anabolic agent, assuming a 38% and 27% discount on list prices of teriparatide and abaloparatide, respectively. This finding remained over a wide range of sensitivity and scenario analyses. These included analyses of patients at even higher risk for fracture, varying the ramp-up time to full zoledronic acid efficacy, and varying the rate of decline in benefit after treatment is stopped. The results were most sensitive to uncertainty in relative risk estimates for hip fracture, long-term fracture utility multipliers, and drug costs. When the anabolic agents are compared to no treatment, the results suggest that anabolic treatments would not produce incremental cost-effectiveness ratios of less than $150,000 per QALY. Our study has some limitations that are worth noting. First, our model assumes a fracture hierarchy that prevents patients from having a fracture classified as less severe than their last fracture. This likely underestimates the number of less severe fractures, and potentially overestimates impacts of hip fractures, which was the most severe fracture in the hierarchy. We attempted to mitigate the influence of hip fracture by calibrating our base-case hip fracture estimates to reflect those predicted by the FRAX Fracture Assessment Tool. Second, we did not consider adverse events, given that anabolic regimens and zoledronic acid exhibited similar serious adverse event rates compared to placebo and to each other in their respective trials. These small event rate differences would have minimal impact on the results. Third, we assumed 100% adherence to all treatments, which would not occur in actual practice. Finally, our base-case cost and cost-effectiveness results for anabolics reflect our current assumptions about drug prices. Despite this, one-way sensitivity analysis showed that drug prices were much less influential on results than differences in fracture ©Institute for Clinical and Economic Review, 2017 Page 58 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents prevention efficacy, and we provided threshold analysis results to offer insight into the drug prices that would make each agent cost-effective under traditional thresholds. Finally, our budget impact analysis for abaloparatide indicates that its use in place of teriparatide and zoledronic acid is not likely to generate access or affordability alerts when using WAC, discounted WAC, or the prices to achieve cost-effectiveness thresholds of $150,000 per QALY or lower. ©Institute for Clinical and Economic Review, 2017 Page 59 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 7. Summary of the Votes and Considerations for Policy 7.1 About the CTAF Process During CTAF public meetings, the CTAF Panel deliberates and votes on key questions related to the systematic review of the clinical evidence, an economic analysis of the applications of treatments under examination, and the supplementary information presented. Panel members are not pre- selected based on the topic being addressed and are intentionally selected to represent a range of expertise and diverse perspectives. Acknowledging that any judgment of evidence is strengthened by real-life clinical and patient perspectives, subject matter experts are recruited for each meeting topic and provide input to CTAF Panel members before the meeting to help clarify their understanding of the different interventions being analyzed in the evidence review. The same clinical experts serve as a resource to the CTAF Panel during their deliberation, and help to shape recommendations on ways the evidence can apply to policy and practice. After the CTAF Panel votes, a policy roundtable discussion is held with the CTAF Panel, clinical experts, patient advocates, payers, and when feasible, manufacturers. The goal of this discussion is to bring stakeholders together to apply the evidence to guide patient education, clinical practice, and coverage and public policies. Participants on policy roundtables are selected for their expertise on the specific meeting topic, are different for each meeting, and do not vote on any questions. At the June 30, 2017 meeting, the CTAF Panel discussed issues regarding the application of the available evidence to help patients, clinicians, and payers address important questions related to the use of anabolic therapies for osteoporosis in postmenopausal women. Following the evidence presentation and public comments (public comments from the meeting can be accessed here, starting at minute 1:18:00), the CTAF Panel voted on key questions concerning the comparative clinical effectiveness and comparative value of the anabolic therapies. These questions are developed by the ICER research team for each assessment to ensure that the questions are framed to address the issues that are most important in applying the evidence to support clinical practice, medical policy decisions, and patient decision-making. The voting results are presented below, along with comments reflecting considerations mentioned by CTAF Panel members during the voting process. In its deliberations and votes related to value, the CTAF Panel made use of a value assessment framework with four different components of “long-term value for money,” a concept that represents the long-term perspective, at the individual patient level, on patient benefits with a ©Institute for Clinical and Economic Review, 2017 Page 60 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents given intervention and the incremental costs to achieve those benefits. The four components of long term value for money are comparative clinical effectiveness, estimated incremental cost- effectiveness, other benefits or disadvantages, and contextual considerations regarding the illness or therapy. There are four elements to consider when deliberating on long-term value for money: 1. Comparative clinical effectiveness is a judgment of the overall difference in clinical outcomes between two interventions (or between an intervention and placebo), tempered by the level of certainty possible given the strengths and weaknesses of the body of evidence. CTAF uses the ICER Evidence Rating Matrix as its conceptual framework for considering comparative clinical effectiveness. 2. Estimated incremental cost-effectiveness is the average per-patient incremental cost of one intervention compared to another to achieve a desired “health gain,” such as an additional stroke prevented, case of cancer diagnosed, or gain of a year of life. Alternative interventions are compared in terms of cost per unit of effectiveness, and the resulting comparison is presented as a cost-effectiveness ratio. Relative certainty in the cost and outcome estimates continues to be a consideration. As a measure of cost-effectiveness, ICER follows common academic and World Health Organization (WHO) standards by using cost per quality-adjusted life years (QALYs) and adopting thresholds at $100,000 per QALY and $150,000 per QALY as guides to reasonable ratios for cost-effectiveness. 3. Other benefits or disadvantages refers to any significant benefits or disadvantages offered by the intervention to the individual patient, caregivers, the delivery system, other patients, or the public that would not have been considered as part of the evidence on comparative clinical effectiveness. Examples of other benefits include better access to treatment centers, mechanisms of treatment delivery that require fewer visits to the clinician’s office, treatments that reduce disparities across various patient groups, and new potential mechanisms of action for treating clinical conditions that have demonstrated low rates of response to currently available therapies. Other disadvantages could include increased burden of treatment on patients or their caregivers. For each intervention evaluated, it will be open to discussion whether other benefits or disadvantages such as these are important enough to factor into the overall judgment of long-term value for money. There is no quantitative measure for other benefits or disadvantages. 4. Contextual considerations include ethical, legal, or other issues (but not cost) that influence the relative priority of illnesses and interventions. Examples of contextual considerations include whether there are currently any existing treatments for the condition, whether the ©Institute for Clinical and Economic Review, 2017 Page 61 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents condition severely affects quality of life or not, and whether the condition affects priority populations. There is no quantitative measure for contextual considerations. 7.2 Voting Results 1) For postmenopausal women with osteoporosis and a high risk of fracture, is the evidence adequate to demonstrate that the net health benefit of treatment with teriparatide (Forteo®, Eli Lilly and Co.), is greater than that of treatment with zoledronic acid? Yes: 2 votes No: 13 votes Comments: The majority of the voting Panel judged there to be inadequate evidence to demonstrate an improved net health benefit for teriparatide over zoledronic acid. Panel members who voted “no” cited several considerations, including the lack of statistically significant differences between the drugs in the network meta-analysis and the absence of data on patient-reported outcomes. Several members noted that the use of surrogate outcomes (i.e., bone mineral density) in osteoporosis trials was disappointing because the fracture prevention effect of osteoporosis therapy appears rapidly as opposed to a condition such as hepatitis C, in which many of the benefits of treatment do not appear for decades. The two panelists who dissented were persuaded by the trend toward superior clinical effectiveness in the network meta-analysis. One Panel member who voted “yes” noted that the greater bone mineral density gains with teriparatide added further evidence in favor of an affirmative vote. 2) For postmenopausal women with osteoporosis and a high risk of fracture, is the evidence adequate to demonstrate that the net health benefit of treatment with abaloparatide (Tymlos™, Radius Health Inc.), is greater than that of treatment with zoledronic acid? Yes: 2 votes No: 13 votes Comments: Panel members voiced nearly identical considerations in their responses to this question. 3) For postmenopausal women with osteoporosis and a high risk of fracture, is the evidence adequate to distinguish between the net health benefit of teriparatide and abaloparatide? Yes: 2 votes No: 13 votes Comments: Panel members who voted “no” noted the wide credible intervals for the comparison between teriparatide and abaloparatide in the ICER network meta-analysis. One of the CTAF members who voted “yes” was persuaded by the QALY gain shown by the economic model and the consistent, although not statistically-significant, trend toward improved health outcomes with abaloparatide. ©Institute for Clinical and Economic Review, 2017 Page 62 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 4) Given the available evidence on comparative effectiveness and incremental cost-effectiveness, and considering other benefits, disadvantages, and contextual considerations, what is the long- term value for money of treatment with teriparatide followed by zoledronic acid versus treatment with zoledronic acid alone for postmenopausal women with osteoporosis at high risk for fracture? Low: 13 votes Intermediate: 2 votes High: 0 votes Comment: “No” votes on this question were primarily driven by the Panel’s earlier judgment that there was inadequate evidence to demonstrate an improved net health benefit for teriparatide versus zoledronic acid. In addition, Panel members who voted “no” cited the results the economic modeling, which demonstrated an incremental cost- effectiveness ratio in excess of $900,000 per quality-adjusted life year (QALY) gained for teriparatide followed by zoledronic acid versus zoledronic acid alone. One panelist noted that, in contrast to the votes on clinical effectiveness that involved a large amount of uncertainty due to wide credible intervals in the network meta-analysis, the cost- effectiveness analysis began with the assumption that each anabolic drug was superior to zoledronic acid, and yet still produced cost per QALY ratios that greatly exceeded commonly-cited ranges for cost-effective therapies. Two panelists noted that daily injections pose a burden to patients, an important other benefit or disadvantage, drawing parallels to the difficulty that patients with diabetes have in adhering to daily injections of insulin. 5) Given the available evidence on comparative effectiveness and incremental cost-effectiveness, and considering other benefits, disadvantages, and contextual considerations, what is the long- term value for money of treatment with abaloparatide followed by zoledronic acid versus treatment with zoledronic acid alone for postmenopausal women with osteoporosis at high risk for fracture? Low: 13 votes Intermediate: 2 votes High: 0 votes Comments: Panel members noted similar considerations for this vote as in to the previous question. Two panelists who voted “low” remarked that abaloparatide’s lower price versus teriparatide was a favorable development, but that further discounts would still be needed to sway them toward a vote of intermediate value. One panelist also highlighted the fact that abaloparatide requires refrigeration up until the first dose as an important consideration for patients; by contrast, teriparatide requires refrigeration throughout the use of each pen injector. ©Institute for Clinical and Economic Review, 2017 Page 63 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 7.3 Roundtable Discussion and Key Policy Implications Following its deliberation on the evidence, the CTAF Panel engaged in a moderated discussion with a policy roundtable about how best to apply the evidence on anabolic therapies for osteoporosis in postmenopausal women to policy and practice. The policy roundtable members included a patient, three clinical experts, two payers, and two representatives from pharmaceutical manufacturers. The discussion reflected multiple perspectives and opinions, and therefore, none of the statements below should be taken as a consensus view held by all participants. The names of the Policy Roundtable participants are shown below, and conflict of information disclosures for all meeting participants can be found in Appendix H. Table 33. Policy Roundtable Members Victoria Dang, PharmD Director, CDAG Program Performance, UnitedHealthcare Medicare and Retirement Matthew Drake, MD, PhD Consultant, Division of Endocrinology, Department of Medicine; Associate Professor of Medicine, Mayo Clinic Deborah Kado, MD, MS Professor, Department of Family Medicine and Public Health; Osteoporosis Clinic Director, Department of Medicine; Deputy Director of Clinical Research and Education, Sam and Rose Stein Institute for Research on Aging, University of California, San Diego John Krege, MD, FAHA Medical Fellow, Eli Lilly and Co. Shireen Fatemi, MD Healthy Bones Regional Co-Lead, Kaiser Permanente Southern California; National Clinical Lead for Osteoporosis, Kaiser Permanente, Assistant Area Medical Director, Kaiser Permanente Panorama City Stuart L. Silverman, MD, FACP, Clinical Professor of Medicine, Cedars-Sinai Medical Center and UCLA FACR School of Medicine; Medical Director, Osteoporosis Medical Center Clinical Research Center; Member, National Bone Health Alliance Osteoporosis Messaging Group Roselyne Smith Patient Martin Zagari, MD Vice President, Global Health Economics, Amgen, Inc. The roundtable discussion was facilitated by Dr. Steven Pearson, MD, MSc, President of ICER. The main themes and recommendations from the discussion are organized by audience and summarized below. Manufacturers Reduce the prices of anabolic agents to align with the clinical benefits they bring to patients. The high cost of anabolic therapy was a recurrent theme during the policy roundtable discussion, and the prevailing sentiment among the CTAF panel was that the current prices were not commensurate with the added value of anabolic therapy. One patient advocate stated “The prices ©Institute for Clinical and Economic Review, 2017 Page 64 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents are too damn high. No subgroup justifies that price.” The lower annual price of abaloparatide compared with teriparatide was viewed as a step in the right direction, but both drugs were still considered to be too expensive. Abstain from direct to consumer advertising and detailing to primary care providers. The policy roundtable appeared to have consensus that only providers with expertise in caring for patients with osteoporosis should prescribe anabolic drugs. Thus, it would be inappropriate to promote their use to PCPs or to patients, who may not be able to appropriately identify the individuals at highest risk of fracture who would merit treatment with an anabolic drug instead of an anti-resorptive agent. Include broader patient groups in randomized trials Experts said that patients they commonly treat for osteoporosis, such as those who are frail and cognitively impaired, were excluded from the randomized trials. In addition, the patients at highest risk for fractures were excluded from the pivotal trials of the anabolic drugs, likely because it would be unethical to randomize such patients in placebo controlled trials. The highest risk patients should be included in randomized trials with active controls in order to demonstrate greater efficacy with anabolic drugs than with standard therapy in a high-risk population. The recent VERO trial is an example of such a study. Payers Given the lack of clinical expert consensus on how to identify patients who would benefit most from consideration of anabolic therapy, design coverage policies with a broad set of criteria by which to determine whether the risk of fracture and the underlying bone pathology would make anabolic therapy a more appropriate first choice than intensive anti-resorptive therapy. Current evidence and expert opinion support the use of anti-resorptive therapy before consideration of an anabolic therapy for most patients with moderate to severe osteoporosis. Expert clinicians on the policy roundtable indicated that they would choose initial anabolic therapy only for the small group of patients at the highest risk of fracture. However, current guidelines and academic models are not precise in their ability to identify those patients at highest future fracture risk, and many clinical experts also believe there are multiple lifestyle, age, former treatment history, and other parameters that often come into play in deciding whether anabolic therapy may be preferable. Payers should therefore design prior authorization and step therapy policies with criteria broad and flexible enough to allow expert clinicians to identify patients at extremely high risk for whom initial treatment with an anabolic agent will be covered. ©Institute for Clinical and Economic Review, 2017 Page 65 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents As noted in the report, specialty society clinical guidelines differ in how they combine clinical criteria and BMD scores to define risk thresholds for different treatments. This has led to significant variation in coverage terms across payers. Payers should seek to combine different elements of criteria in a way that provides clarity for coverage review purposes while being comprehensive enough to capture multiple ways that the nature of the risk for future fracture should qualify a patient for coverage of anabolic therapy. Some of the coverage criteria used by payers for this purpose are shown below. • T-scores of -3.5 or lower without any other requirements of treatment failure. • T-scores of -2.5 or lower with prior fragility fractures and/or prior treatment failure, contraindication, or intolerance to another osteoporosis therapy. Treatment failure is often left undefined, but Blue Shield of California defines it as a T-score that remains ≤ -2.5 with or without a low-impact fracture while on bisphosphonate treatment, while Cigna defines failure as a “significant” decrease in BMD after one year of treatment or a new fracture while on bisphosphonate treatment. • T-scores of -2.5 and otherwise at “high risk” for fracture through prior history of fragility fractures or a combination of considerations including more than three months of systemic corticosteroid use, advanced age, family history of osteoporosis at a young age, cigarette smoking, and three or more alcoholic drinks per day. Create a prior authorization process for anabolic therapies that is clear and efficient for providers Patients and providers spoke of delays of several months in obtaining decisions about authorization for anabolic therapy. Specialists in osteoporosis spoke of having one employee working more than half time solely on authorization for medications for osteoporosis and of primary care physicians referring patients solely for assistance with such authorization. Given these concerns, payers should review their prior authorization policies for anabolic therapies to ensure that coverage decisions are being made in a timely manner. If the prices of anabolic agents are reduced, ease access restrictions A consistent message from patients and clinicians was the desire to have access to the full range of drugs in order to tailor treatment to meet individual clinical characteristics and personal preferences. However, the high cost of anabolic agents makes it difficult for payers to accommodate wider access. If manufacturers bring prices of these agents in line with the value the treatments provide, insurers should ease access restrictions. The cost of these agents may still create substantial burdens because most patients with severe osteoporosis are older and on fixed incomes. To address this concern, payers should also implement policies that protect patients from excessive costs, such as placing anabolic agents on formulary tiers that require lower out-of-pocket spending. ©Institute for Clinical and Economic Review, 2017 Page 66 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Patient Advocacy Organizations Demand the inclusion of patient-centered outcomes in clinical trials Patients highlighted the lack of patient centered outcomes such as independence, mobility status, ability to perform activities of daily living, and overall quality of life in clinical trials of therapies for osteoporosis. Patient groups should advocate for appropriate patient-centered outcomes to be integrated into future clinical trials. Such groups can also promote the development of high-quality registries of patients treated with drugs for osteoporosis, with a goal of answering questions about long-term outcomes such as loss of independence that are not addressed by most trials. Continue to promote lifestyle changes that protect against osteoporosis Lifestyle factors such as adequate vitamin D and calcium intake, weight bearing exercise, and smoking cessation play important roles in bone health and form the foundation of both the prevention and treatment of osteoporosis. Osteoporosis advocacy organizations should continue to reach out to the broader patient community to educate them on the importance of these lifestyle factors. Specialty Societies Develop clear guidelines for use of anabolic agents Clinicians, patients, and payers need guidelines that clearly define which patients are sufficiently high risk to warrant treatment with anabolic agents. Additionally, while awaiting better evidence, recommendations should also discuss appropriate sequencing of agents for osteoporosis for patients at various levels of risk. The current literature finds that BMD increases are greater when anabolic therapy is followed by anti-resorptive therapy rather than anti-resorptive therapy followed by anabolic therapy. Studies of drug sequencing must use fracture outcomes to confirm hypotheses supported by BMD and bone turnover markers. Experts advocated for the use of anabolic drugs when treatment with anti-resorptive agents fail. However, there is no clear definition of treatment failure and no trials comparing continued therapy with switching to a new therapy based on a proposed definition of treatment failure. Clinical guidelines should include a conclusive definition of treatment failure to help direct clinical care. Regulators Promote hip fracture as the most important outcome in pivotal clinical trials Manufacturers and others pointed out that FDA guidance has led to morphometric vertebral fractures being the primary outcome in pivotal trials for treatments of osteoporosis. Although ©Institute for Clinical and Economic Review, 2017 Page 67 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents currently available drugs have been approved using less patient-centric outcome measures, regulators should work with patient groups and manufacturers to seek every possible route to power pivotal trials to measure improvements in hip fracture rates. Require that pivotal trials include an active comparator Given the large body of evidence that treatment of osteoporosis with anti-resorptive agents prevents vertebral, non-vertebral, and hip fractures, it is unethical to continue to perform placebo controlled trials in high-risk patients. Requiring an active comparator in clinical trials would improve patient and clinician understanding of the relative benefits, risks, and value of available treatment options. Researchers Develop better risk assessment tools to identify patients at extreme risk for fracture There was frequent reference to the existence of a group of patients with osteoporosis who were at extremely high risk for fracture and warranted treatment with therapies other than standard bisphosphonate therapy. Risk assessment tools should be developed that can accurately predict fracture incidence in patients felt to be at particularly high risk. However, the existence of a more appropriate model would not by itself increase certainty about the relative benefits of anabolic therapy compared to a potent bisphosphonate; head-to-head trials are still needed to conclusively determine whether there is a net benefit **** This is the first CTAF review of anabolic therapies for the treatment of osteoporosis in postmenopausal women. ©Institute for Clinical and Economic Review, 2017 Page 68 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents References 1. Wright NC, Looker AC, Saag KG, et al. The recent prevalence of osteoporosis and low bone mass in the United States based on bone mineral density at the femoral neck or lumbar spine. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2014;29(11):2520-2526. 2. United States. Public Health Service. Office of the Surgeon General. Bone health and osteoporosis : a report of the Surgeon General. Rockville, Md.: U.S. Dept. of Health and Human Services, Public Health Service, Office of the Surgeon General; 2004. 3. Burge R, Dawson-Hughes B, Solomon DH, Wong JB, King A, Tosteson A. Incidence and economic burden of osteoporosis-related fractures in the United States, 2005-2025. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2007;22(3):465-475. 4. Amgen And UCB Announce Top-Line Phase 3 Data From Active-Comparator Study Of EVENITY™ (Romosozumab) In Postmenopausal Women With Osteoporosis [press release]. May 21 2017. 5. Adler RA, El-Hajj Fuleihan G, Bauer DC, et al. Managing Osteoporosis in Patients on Long-Term Bisphosphonate Treatment: Report of a Task Force of the American Society for Bone and Mineral Research. Journal of Bone and Mineral Research. 2016;31(1):16-35. 6. Camacho PM, Petak SM, Binkley N, et al. AMERICAN ASSOCIATION OF CLINICAL ENDOCRINOLOGISTS AND AMERICAN COLLEGE OF ENDOCRINOLOGY CLINICAL PRACTICE GUIDELINES FOR THE DIAGNOSIS AND TREATMENT OF POSTMENOPAUSAL OSTEOPOROSIS - 2016. Endocrine practice : official journal of the American College of Endocrinology and the American Association of Clinical Endocrinologists. 2016;22(Suppl 4):1-42. 7. Cosman F, de Beur SJ, LeBoff MS, et al. Clinician's Guide to Prevention and Treatment of Osteoporosis. Osteoporos Int. 2014;25(10):2359-2381. 8. North American Menopause Society. Management of osteoporosis in postmenopausal women: 2010 position statement of The North American Menopause Society. Menopause (New York, NY). 2010;17(1):25-54; quiz 55-26. 9. Qaseem A, Forciea M, McLean RM, Denberg TD, for the Clinical Guidelines Committee of the American College of P. Treatment of low bone density or osteoporosis to prevent fractures in men and women: A clinical practice guideline update from the american college of physicians. Annals of internal medicine. 2017. 10. Jha S, Wang Z, Laucis N, Bhattacharyya T. Trends in Media Reports, Oral Bisphosphonate Prescriptions, and Hip Fractures 1996-2012: An Ecological Analysis. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2015;30(12):2179-2187. 11. Karlsson L, Lundkvist J, Psachoulia E, Intorcia M, Strom O. Persistence with denosumab and persistence with oral bisphosphonates for the treatment of postmenopausal osteoporosis: a retrospective, observational study, and a meta-analysis. Osteoporos Int. 2015;26(10):2401-2411. 12. Modi A, Sajjan S, Insinga R, Weaver J, Lewiecki EM, Harris ST. Frequency of discontinuation of injectable osteoporosis therapies in US patients over 2 years. Osteoporos Int. 2017;28(4):1355- 1363. 13. US Food and Drug Administration (FDA). Forteo (teriparatide [rDNA origin] injection) for subcutaneous use) label. 2012; https://www.accessdata.fda.gov/drugsatfda_docs/label/2013/021318s036lbl.pdf. ©Institute for Clinical and Economic Review, 2017 Page 69 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 14. FDA Approves Radius Health's TYMLOS™ (abaloparatide), a Bone Building Agent for the Treatment of Postmenopausal Women with Osteoporosis at High Risk for Fracture [press release]. April 28, 2017. 15. US Food and Drug Administration (FDA). Tymlos (abaloparatide) injection, for subcutaneous use: label. 2017; https://www.accessdata.fda.gov/drugsatfda_docs/label/2017/208743lbl.pdf. 16. National Osteoporosis Foundation. Bone Health Index Survey: Final Report - September 2016. 2016. 17. Cosman F, Crittenden DB, Adachi JD, et al. Romosozumab Treatment in Postmenopausal Women with Osteoporosis. N Engl J Med. 2016;375(16):1532-1543. 18. Miller PD, Hattersley G, Riis BJ, et al. Effect of Abaloparatide vs Placebo on New Vertebral Fractures in Postmenopausal Women With Osteoporosis: A Randomized Clinical Trial. Jama. 2016;316(7):722-733. 19. Neer RM, Arnaud CD, Zanchetta JR, et al. Effect of parathyroid hormone (1-34) on fractures and bone mineral density in postmenopausal women with osteoporosis. N Engl J Med. 2001;344(19):1434-1441. 20. Black DM, Delmas PD, Eastell R, et al. Once-yearly zoledronic acid for treatment of postmenopausal osteoporosis. N Engl J Med. 2007;356(18):1809-1822. 21. Prevrhal S, Krege JH, Chen P, Genant H, Black DM. Teriparatide vertebral fracture risk reduction determined by quantitative and qualitative radiographic assessment. Current medical research and opinion. 2009;25(4):921-928. 22. Burge RT, Disch DP, Gelwicks S, Zhang X, Krege JH. Hip and other fragility fracture incidence in real-world teriparatide-treated patients in the United States. Osteoporos Int. 2017;28(3):799- 809. 23. Davis S, Martyn-St. James M, Sanderson J, et al. Bisphosphonates for preventing osteoporotic fragility fractures (including a partial update of NICE technology appraisal guidance 160 and 161). National Institute for Health and Care Excellence (NICE); 2015. 24. Freemantle N, Cooper C, Diez-Perez A, et al. Results of indirect and mixed treatment comparison of fracture efficacy for osteoporosis treatments: a meta-analysis. Osteoporos Int. 2013;24(1):209-217. 25. Murad MH, Drake MT, Mullan RJ, et al. Clinical review. Comparative effectiveness of drug treatments to prevent fragility fractures: a systematic review and network meta-analysis. The Journal of clinical endocrinology and metabolism. 2012;97(6):1871-1880. 26. Yang XC, Deng ZH, Wen T, et al. Network Meta-Analysis of Pharmacological Agents for Osteoporosis Treatment and Fracture Prevention. Cellular physiology and biochemistry : international journal of experimental cellular physiology, biochemistry, and pharmacology. 2016;40(3-4):781-795. 27. Zhang L, Pang Y, Shi Y, et al. Indirect comparison of teriparatide, denosumab, and oral bisphosphonates for the prevention of vertebral and nonvertebral fractures in postmenopausal women with osteoporosis. Menopause (New York, NY). 2015;22(9):1021-1025. 28. Black DM, Bilezikian JP, Ensrud KE, et al. One year of alendronate after one year of parathyroid hormone (1-84) for osteoporosis. N Engl J Med. 2005;353(6):555-565. 29. Cosman F, Nieves JW, Dempster DW. Treatment Sequence Matters: Anabolic and Antiresorptive Therapy for Osteoporosis. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2017;32(2):198-202. 30. Hanmer J, Lawrence WF, Anderson JP, Kaplan RM, Fryback DG. Report of nationally representative values for the noninstitutionalized US adult population for 7 health-related ©Institute for Clinical and Economic Review, 2017 Page 70 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents quality-of-life scores. Medical decision making : an international journal of the Society for Medical Decision Making. 2006;26(4):391-400. 31. Melton LJ, 3rd, Crowson CS, O'Fallon WM. Fracture incidence in Olmsted County, Minnesota: comparison of urban with rural rates and changes in urban rates over time. Osteoporos Int. 1999;9(1):29-37. 32. SSR Health. US Brand Rx Net Price. 2016. 33. Aitken M, Kleinrock M, Pennente K, Lyle J, Nass D, Caskey L. Medicines Use and Spending in the U.S.: A review of 2015 and Outlook to 2020. Parsippany, NJ: QuintilesIMS;2016. 34. Tosteson AN, Burge RT, Marshall DA, Lindsay R. Therapies for treatment of osteoporosis in US women: cost-effectiveness and budget impact considerations. The American journal of managed care. 2008;14(9):605-615. 35. Parthan A, Emptage N, Taylor D, et al. Budgetary impact analysis of denosumab in a US health plan. American Journal of Pharmacy Benefits. 2013;5(5):e129-e138. 36. Padhi D, Jang G, Stouch B, Fang L, Posvar E. Single-dose, placebo-controlled, randomized study of AMG 785, a sclerostin monoclonal antibody. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2011;26(1):19-26. 37. Wright NC, Saag KG, Curtis JR, et al. Recent trends in hip fracture rates by race/ethnicity among older US adults. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2012;27(11):2325-2332. 38. US Preventive Services Task Force. Screening for osteoporosis: U.s. preventive services task force recommendation statement. Annals of internal medicine. 2011;154(5):356-364. 39. Gillespie CW, Morin PE. Trends and Disparities in Osteoporosis Screening Among Women in the United States, 2008-2014. Am J Med. 2017;130(3):306-316. 40. National Quality Measures Clearinghouse. Osteoporosis: percentage of women age 50 to 85 who suffered a fracture and who either had a bone mineral density test or received a prescription for a drug to treat osteoporosis. 2015. 41. California Department of Health Care Services. Contract Drugs List. 2017; http://files.medi- cal.ca.gov/pubsdoco/manual/man_query.asp?wSearch=%28%23filename+drugscdl%2a.doc+OR +%23filename+drugscdl%2a.zip%29&wFLogo=Contract+Drugs+List&wFLogoH=52&wFLogoW=51 6&wAlt=Contract+Drugs+List&wPath=N. Accessed March 28, 2017. 42. Aetna. 2017 Prescription Drug Search. 2017; https://rxtools.aetnamedicare.com/PlanCompare/indeptools2017/Tools/FormularySearch/Form ularySearch.aspx. Accessed March 29, 2017. 43. Anthem Blue Cross. Your Blue Cross MedicareRx (PDP) with Senior Rx Plus Plan: 2017 Part D Formulary (List of Covered Drugs). 2017; https://www11.anthem.com/ca/provider/f0/s0/t0/pw_e234971.pdf. Accessed March 29, 2017. 44. Blue Shield of California. Blue Shield Medicare Basic Plan (PDP): 2017 Formulary. 2017; https://fm.formularynavigator.com/MemberPages/pdf/2017Basic_9107_BS%20CA%20Two%20 Column_2614.pdf. Accessed March 29, 2017. 45. Cigna HealthSpring. 2017 Cigna-Healthspring Rx Comprehensive Drug List (Formulary). 2017; https://www.cigna.com/iwov-resources/medicare-2017/docs/formulary-ea-pdp- a.pdf?WT.z_nav=medicare%2Fpart-d%2Fdrug-list-formulary%3BBody%3BEnglish. Accessed March 29, 2017. 46. Health Net. Pharmacy Information Overview. 2017; https://www.healthnet.com/portal/provider/content/iwc/provider/unprotected/pharmacy_inf o/book/pharmacy_information.action#drug_info_medicare_plansContent. Accessed June 14, 2017. ©Institute for Clinical and Economic Review, 2017 Page 71 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 47. Humana. Provider Drug List Search. 2017; http://apps.humana.com/UnsecuredDrugListSearch/Search.aspx. Accessed June 14, 2017. 48. United Healthcare. 2017 Comprehensive Formulary (complete list of covered drugs): AARP MedicareRX Walgreens (PDP). 2017; https://www.uhcmedicaresolutions.com/alphadms/ovdms10g/groups/ov/@ov/@highrespdf/do cuments/highrespdf/4044271.pdf. Accessed March 29, 2017. 49. Aetna. Aetna Medicare Rx Saver (PDP). 2017; https://www.aetnamedicare.com/documents/individual/2017/formularies/PA_2017_17020AET _a1_EN.pdf. Accessed March 29, 2017. 50. Anthem Blue Cross. Prior Authorization Criteria. 2017; https://www11.anthem.com/ca/provider/f0/s0/t0/pw_e234981.pdf. Accessed March 29, 2017. 51. Blue Shield of California. Medicare Part D Coverage Criteria: Forteo (teriparatide). 2017; https://www.blueshieldca.com/sites/medicare/documents/PA_CY2017_FORTEO%20teriparatid e_MCweb.pdf. Accessed March 29, 2017. 52. United Healthcare. Prior Authorization Criteria. 2017; https://www.uhcmedicaresolutions.com/online_documents/ovation/pdf/pdp/en/2017/Prior_A uth_PWAG_2017.pdf. Accessed March 29, 2017. 53. Cigna HealthSpring. 2017 Cigna-HealthSpring Prior Authorization Criteria. 2017; https://www.cigna.com/iwov-resources/medicare-2017/docs/prior-authorization- chs.pdf?WT.z_nav=medicare%2Fpart-d%2Fdrug-list-formulary%3BBody%3BCigna- HealthSpring%20Medicare%20Plans. Accessed March 29, 2017. 54. Humana. Forteo (teriparatide) Pharmacy Coverage Policy (Medicare and Puerto Rico). Vol 20172016. 55. Blue Shield of California. Medicare Part D Coverage Criteria: zoledronic acid (generic Reclast). 2017; https://www.blueshieldca.com/sites/medicare/documents/PA_CY2017_zoledronic%20acid%20( generic%20RECLAST)_MCweb.pdf. Accessed March 29, 2017. 56. Humana. Reclast (zoledronic acid) Pharmacy Coverage Policy (Medicare, Exchanges, Puerto Rico, Commercial). 2016. 57. Humana. Zometa (zoledronic acid) Pharmacy Coverage Policy (Medicare, Exchanges, Puerto Rico, Commercial). 2017. 58. Anthem Blue Cross. Individual Select Drug List (Searchable). 2017; https://www11.anthem.com/ca/pharmacyinformation/. Accessed June 14, 2017. 59. Blue Shield of California. Blue Shield Standard Drug Formulary. 2017; https://fm.formularynavigator.com/MemberPages/pdf/CommercialStandard2017_Closed_1014 3_BS%20CA%20Standard_2810.pdf. Accessed March 28, 2017. 60. Blue Shield of California. Specialty Drug List for Standard Drug Formulary. 2017; https://www.blueshieldca.com/bsca/documents/pharmacy/Specialty_Drugs_List_Standard_For mulary.pdf. Accessed March 28, 2017. 61. Kaiser Permanente. 2017 California Marketplace Formulary. 2017; https://healthy.kaiserpermanente.org/static/health/pdfs/formulary/cal/2017_ca_marketplace_f ormulary.pdf. Accessed March 28, 2017. 62. Health Net. Prior Authorization Protocol: Forteo (teriparatide). 2016; https://www.healthnet.com/portal/common/content/iwc/common/unprotected/pharmacy_inf o/prior_auth_criteria.action#f. Accessed March 29, 2017. 63. National Institute for Health and Care Excellence. Secondary prevention of fragility fractures in postmenopausal women. 2017; ©Institute for Clinical and Economic Review, 2017 Page 72 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents https://pathways.nice.org.uk/pathways/osteoporosis#content=view-node%3Anodes-secondary- prevention-of-fragility-fractures-in-postmenopausal-women. Accessed April 5, 2017. 64. Cook DJ, Mulrow CD, Haynes RB. Systematic reviews: synthesis of best evidence for clinical decisions. Annals of internal medicine. 1997;126(5):376-380. 65. Higgins J, Green S. Cochrane Handbook for Systematic Reviews of Interventions Version 5.0.0 [updated February 2008). The Cochrane Collaboration; 2008. 66. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Annals of internal medicine. 2009;151(4):264-269, w264. 67. Ollendorf DA, Pearson SD. An integrated evidence rating to frame comparative effectiveness assessments for decision makers. Medical care. 2010;48(6 Suppl):S145-152. 68. Caldwell DM, Ades AE, Higgins JPT. Simultaneous comparison of multiple treatments: combining direct and indirect evidence. BMJ. 2005;331(7521):897-900. 69. Cosman F, Miller PD, Williams GC, et al. Eighteen Months of Treatment With Subcutaneous Abaloparatide Followed by 6 Months of Treatment With Alendronate in Postmenopausal Women With Osteoporosis: Results of the ACTIVExtend Trial. Mayo Clinic proceedings. 2017;92(2):200-210. 70. Black DM, Rosen CJ. Postmenopausal Osteoporosis. New England Journal of Medicine. 2016;374(3):254-262. 71. Langdahl BL, Rajzbaum G, Jakob F, et al. Reduction in fracture rate and back pain and increased quality of life in postmenopausal women treated with teriparatide: 18-month data from the European Forsteo Observational Study (EFOS). Calcified tissue international. 2009;85(6):484- 493. 72. Yu S, Burge RT, Foster SA, Gelwicks S, Meadows ES. The impact of teriparatide adherence and persistence on fracture outcomes. Osteoporos Int. 2012;23(3):1103-1113. 73. Silverman S, Miller P, Sebba A, et al. The Direct Assessment of Nonvertebral Fractures in Community Experience (DANCE) study: 2-year nonvertebral fragility fracture results. Osteoporosis International. 2013;24(8):2309-2317. 74. Kendler D, Zerbini C, Russo L, et al. Effects of 24 Months Treatment of Teriparatide Compared with Risedronate on New Fractures in Postmenopausal Women with Severe Osteoporosis: a Randomised, Double-Dummy Clinical Trial (VERO). Paper presented at: World Congress on Osteoporosis, Osteoarthritis, and Musculoskeletal Diseases2017; Florence, Italy. 75. Harvey NC, Kanis JA, Oden A, et al. FRAX and the effect of teriparatide on vertebral and non- vertebral fracture. Osteoporos Int. 2015;26(11):2677-2684. 76. Cosman F, Hattersley G, Hu MY, Williams GC, Fitzpatrick LA, Black DM. Effects of Abaloparatide- SC on Fractures and Bone Mineral Density in Subgroups of Postmenopausal Women With Osteoporosis and Varying Baseline Risk Factors. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2017;32(1):17-23. 77. Cefalu CA. Is bone mineral density predictive of fracture risk reduction? Current medical research and opinion. 2004;20(3):341-349. 78. Miller PD. Bone density and markers of bone turnover in predicting fracture risk and how changes in these measures predict fracture risk reduction. Current Osteoporosis Reports. 2005;3(3):103-110. 79. Seeman E. Is a change in bone mineral density a sensitive and specific surrogate of anti-fracture efficacy? Bone. 2007;41(3):308-317. 80. Si L, Winzenberg TM, Palmer AJ. A systematic review of models used in cost-effectiveness analyses of preventing osteoporotic fractures. Osteoporos Int. 2014;25(1):51-60. ©Institute for Clinical and Economic Review, 2017 Page 73 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 81. Center for Metabolic Bone Diseases. FRAX Fracture Risk Assessment Tool - Calculation Tool (US). https://www.shef.ac.uk/FRAX/tool.aspx?country=9. Accessed May 1, 2017. 82. Kanis JA, Johansson H, Oden A, Dawson-Hughes B, Melton LJ, 3rd, McCloskey EV. The effects of a FRAX revision for the USA. Osteoporos Int. 2010;21(1):35-40. 83. Centers for Disease Control and Prevention. United States Life Tables, 2008. https://www.cdc.gov/nchs/data/nvsr/nvsr61/nvsr61_03.pdf. 84. Neumann PJ, Sanders GD, Russell LB, Siegel JE, Ganiats TG. Cost-effectiveness in health and medicine. Oxford University Press; 2016. 85. Black DM, Reid IR, Boonen S, et al. The effect of 3 versus 6 years of zoledronic acid treatment of osteoporosis: a randomized extension to the HORIZON-Pivotal Fracture Trial (PFT). Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2012;27(2):243-254. 86. Black DM, Reid IR, Cauley JA, et al. The effect of 6 versus 9 years of zoledronic acid treatment in osteoporosis: a randomized second extension to the HORIZON-Pivotal Fracture Trial (PFT). Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2015;30(5):934-944. 87. Tosteson AN, Gottlieb DJ, Radley DC, Fisher ES, Melton LJ, 3rd. Excess mortality following hip fracture: the role of underlying health status. Osteoporos Int. 2007;18(11):1463-1472. 88. Kanis JA, Johnell O, Oden A, et al. The risk and burden of vertebral fractures in Sweden. Osteoporos Int. 2004;15(1):20-26. 89. Oleksik A, Lips P, Dawson A, et al. Health-related quality of life in postmenopausal women with low BMD with or without prevalent vertebral fractures. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2000;15(7):1384- 1392. 90. Peasgood T, Herrmann K, Kanis JA, Brazier JE. An updated systematic review of Health State Utility Values for osteoporosis related conditions. Osteoporos Int. 2009;20(6):853-868. 91. Burstrom K, Johannesson M, Diderichsen F. A comparison of individual and social time trade-off values for health states in the general population. Health policy (Amsterdam, Netherlands). 2006;76(3):359-370. 92. Redbook Online. 2017. 93. Radius Health. TYMLOS™ and 1q 2017 Financial Results Update. [Investor presentation]. 2017; http://edge.media-server.com/m/p/z5woypf3. Accessed 38. 94. Bonafede M, Shi N, Viswanathan HN, Yurgin N. PMS13 OSTEOPOROSIS-RELATED FRACTURE COSTS AMONG FEMALE COMMERCIALLY INSURED AND MEDICARE PATIENTS. Value in Health. 2011;14(3):A125. 95. Parthan A, Kruse M, Yurgin N, Huang J, Viswanathan HN, Taylor D. Cost effectiveness of denosumab versus oral bisphosphonates for postmenopausal osteoporosis in the US. Applied health economics and health policy. 2013;11(5):485-497. 96. Insinga R. Administration Costs of Denosumab and Zoledronic Acid for Postmenopausal Osteoporosis. Am J Pharm Benefits. 2016;8(3):e42-e47. 97. Amgen and UCB Announce U.S. FDA Accepatance of Biologics License Application for Romosozumab [press release]. 2016. 98. Klotzbuecher CM, Ross PD, Landsman PB, Abbott TA, 3rd, Berger M. Patients with prior fractures have an increased risk of future fractures: a summary of the literature and statistical synthesis. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research. 2000;15(4):721-739. ©Institute for Clinical and Economic Review, 2017 Page 74 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 99. Johnell O, Kanis JA, Oden A, et al. Mortality after osteoporotic fractures. Osteoporos Int. 2004;15(1):38-42. 100. Ivergard M, Strom O, Borgstrom F, Burge RT, Tosteson AN, Kanis J. Identifying cost-effective treatment with raloxifene in postmenopausal women using risk algorithms for fractures and invasive breast cancer. Bone. 2010;47(5):966-974. 101. Lekander I, Borgstrom F, Strom O, Zethraeus N, Kanis JA. Cost effectiveness of hormone therapy in women at high risks of fracture in Sweden, the US and the UK--results based on the Women's Health Initiative randomised controlled trial. Bone. 2008;42(2):294-306. 102. Lekander I, Borgström F, Ström O, Zethraeus N, Kanis JA. Cost-Effectiveness of Hormone Therapy in the United States. Journal of Women's Health. 2009;18(10):1669-1677. 103. Pham AN, Datta SK, Weber TJ, Walter LC, Colon-Emeric CS. Cost-effectiveness of oral bisphosphonates for osteoporosis at different ages and levels of life expectancy. Journal of the American Geriatrics Society. 2011;59(9):1642-1649. 104. Salpeter SR, Buckley NS, Liu H, Salpeter EE. The cost-effectiveness of hormone therapy in younger and older postmenopausal women. Am J Med. 2009;122(1):42-52.e42. 105. Murphy DR, Smolen LJ, Klein TM, Klein RW. The cost effectiveness of teriparatide as a first-line treatment for glucocorticoid-induced and postmenopausal osteoporosis patients in Sweden. BMC musculoskeletal disorders. 2012;13:213. 106. Agency for Healthcare Research and Quality. U.S. Preventive Services Task Force Procedure Manual. 2008. 107. Keshishian A, Boytsov N, Burge R, et al. Examining the treatment gap and risk of subsequent fractures among females with a fragility fracture in the US Medicare population. Osteoporos Int. 2017. ©Institute for Clinical and Economic Review, 2017 Page 75 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents APPENDICES ©Institute for Clinical and Economic Review, 2017 Page 76 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Appendix A. Search Strategies and Results Table A1. PRISMA 2009 Checklist # Checklist item TITLE Title 1 Identify the report as a systematic review, meta-analysis, or both. ABSTRACT Structured summary 2 Provide a structured summary including, as applicable: background; objectives; data sources; study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number. INTRODUCTION Rationale 3 Describe the rationale for the review in the context of what is already known. Objectives 4 Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, outcomes, and study design (PICOS). METHODS Protocol and registration 5 Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide registration information including registration number. Eligibility criteria 6 Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale. Information sources 7 Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched. Search 8 Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated. Study selection 9 State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis). Data collection process 10 Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators. Data items 11 List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made. Risk of bias in individual 12 Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done studies at the study or outcome level), and how this information is to be used in any data synthesis. Summary measures 13 State the principal summary measures (e.g., risk ratio, difference in means). Synthesis of results 14 Describe the methods of handling data and combining results of studies, if done, including measures of consistency (e.g., I2) for each meta-analysis. ©Institute for Clinical and Economic Review, 2017 Page 77 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Risk of bias across studies 15 Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies). Additional analyses 16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which were pre-specified. RESULTS Study selection 17 Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram. Study characteristics 18 For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations. Risk of bias within studies 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12). Results of individual studies 20 For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot. Synthesis of results 21 Present results of each meta-analysis done, including confidence intervals and measures of consistency. Risk of bias across studies 22 Present results of any assessment of risk of bias across studies (see Item 15). Additional analysis 23 Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, meta-regression [see Item 16]). DISCUSSION Summary of evidence 24 Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers). Limitations 25 Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias). Conclusions 26 Provide a general interpretation of the results in the context of other evidence, and implications for future research. FUNDING Funding 27 Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review. From: Moher D, Liberati A, Tetzlaff J, Altman DG. The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(6): e1000097. doi:10.1371/journal.pmed1000097 ©Institute for Clinical and Economic Review, 2017 Page 78 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table A2. PubMed search, January 24, 2017 #1 ((("teriparatide"[MeSH Terms] OR teriparatide) OR ("abaloparatide"[Supplementary Concept] OR abaloparatide OR "AMG 785"[Supplementary Concept] OR "AMG 785" OR "romosozumab"[All Fields]) #2 #1 AND ("osteoporosis"[All Fields] OR "osteoporosis"[MeSH Terms] OR "osteopenia"[All Fields])))) #3 ((("addresses"[Publication Type] OR "autobiography"[Publication Type] OR "bibliography"[Publication Type] OR "biography"[Publication Type] OR "book illustrations"[Publication Type] OR "case reports"[Publication Type] OR "classical article"[Publication Type] OR "clinical conference"[Publication Type] OR "clinical trial, phase i"[Publication Type] OR "collected works"[Publication Type] OR "comment"[Publication Type] OR "congresses"[Publication Type] OR "consensus development conference"[Publication Type] OR "consensus development conference, nih"[Publication Type] OR "dataset"[Publication Type] OR "dictionary"[Publication Type] OR "directory"[Publication Type] OR "duplicate publication"[Publication Type] OR "editorial"[Publication Type] OR "electronic supplementary materials"[Publication Type] OR "ephemera"[Publication Type] OR "evaluation studies"[Publication Type] OR "festschrift"[Publication Type] OR "government publications"[Publication Type] OR "guideline"[Publication Type] OR "historical article"[Publication Type] OR "interactive tutorial"[Publication Type] OR "interview"[Publication Type] OR "introductory journal article"[Publication Type] OR "lectures"[Publication Type] OR "legal cases"[Publication Type] OR "legislation"[Publication Type] OR "letter"[Publication Type] OR "news"[Publication Type] OR "newspaper article"[Publication Type] OR "patient education handout"[Publication Type] OR "periodical index"[Publication Type] OR "personal narratives"[Publication Type] OR "pictorial works"[Publication Type] OR "portraits"[Publication Type] OR "practice guideline"[Publication Type] OR "retracted publication"[Publication Type] OR "retraction of publication"[Publication Type] OR "review"[Publication Type] OR "video audio media"[Publication Type] OR "webcasts"[Publication Type]))) #4 (((("clinical study"[Publication Type] OR "clinical trial"[Publication Type] OR "comparative study"[Publication Type] OR "meta analysis"[Publication Type] OR "observational study"[Publication Type]))) #5 #2 AND #4 #6 #5 NOT #3 Table A3. Cochrane Central Register of Controlled Trials search, January 24, 2017 (via Ovid) 1 Exp teriparatide/ 2 Teriparatide 3 Abaloparatide 4 Romosozumab 5 Osteopenia 6 Exp osteporosis 7 1 or 2 or 3 or 4 8 5 or 6 9 7 and 8 ©Institute for Clinical and Economic Review, 2017 Page 79 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table A4. Embase search, January 24, 2017 #1 'parathyroid hormone[1-34]'/exp OR 'parathyroid hormone[1-34]' OR 'teriparatide'/exp OR teriparatide OR 'forteo'/exp OR forteo OR 'abaloparatide'/exp OR 'abaloparatide' OR 'amg 785'/exp OR 'amg 785' OR 'romosozumab'/exp OR 'romosozumab' AND ('osteoporosis'/exp OR 'osteoporosis' OR 'osteopenia'/exp OR 'osteopenia') #2 'chapter'/it OR 'conference review'/it OR 'editorial'/it OR 'letter'/it OR 'note'/it OR 'review'/it OR 'short survey'/it #3 #1 NOT #2 #4 'animal'/exp OR 'nonhuman'/exp OR 'animal experiment'/exp #5 'human'/exp #6 #4 AND #5 #7 #4 NOT #6 #8 #3 NOT #7 #9 #8 AND [english]/lim #10 #9 AND [medline]/lim #11 #9 NOT #10 #12 #11 AND ('conference abstract'/it OR 'conference paper'/it) #13 #11 NOT #12 ©Institute for Clinical and Economic Review, 2017 Page 80 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Figure A1. PRISMA flow Chart Showing Results of Literature Search for Anabolic Therapies for Osteoporosis 788 potentially relevant references screened 566 citations excluded Population: 165 Intervention: 89 Comparator: 14 Outcomes: 162 222 references for full text Study Type: 136 review 219 citations excluded (different intervention, non-labeled dosing, mixed population without stratification of results) 3 TOTAL 3 RCTs ©Institute for Clinical and Economic Review, 2017 Page 81 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Appendix B. California Health Exchange and Medicaid Coverage Policies Table B1. Representative Medi-Cal and Silver-Tier Covered California Coverage Policies for Abaloparatide, Teriparatide, Alendronate, and Zoledronic Acid Medi-Cal Anthem Health Net Kaiser Permanente BSCA Teriparatide Tier Not listed Non-formulary, Specialty 4 (Specialty) Specialty Specialty ST - No No No No PA - No Yes No No Abaloparatide Tier - Non-formulary Specialty - - ST - - - - - PA - - - - - Alendronate Tier Covered 1 1 1, 2 1 ST - No No No No PA - No No No No Zoledronic Acid Tier Covered 4 N/C 1 N/C ST - No - No - PA - Yes - No - N/C: not covered, PA: prior authorization, ST: step therapy ©Institute for Clinical and Economic Review, 2017 Page 82 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Appendix C. Previous Systematic Reviews and Technology Assessments Systematic reviews and meta-analyses that compared fracture outcomes for two or more drugs in postmenopausal women with osteoporosis are summarized below. Murad et al., 201225 Murad and colleagues performed a systematic review and network meta-analysis of drugs for women with postmenopausal osteoporosis using data from 116 randomized studies. Teriparatide, alendronate, zoledronic acid, risedronate, denosumab, and the combination of calcium and vitamin D all significantly reduced hip fractures. There was a significant reduction in vertebral fractures compared to placebo for teriparatide, alendronate, zoledronic acid, risedronate, denosumab, ibandronate, and raloxifene. Similarly, there was a significant reduction in non-vertebral fractures compared to placebo for teriparatide, alendronate, zoledronic acid, risedronate, and denosumab. Teriparatide consistently had the highest probability of being ranked as the most effective, but was not significantly more effective than the other agents. Fremantle et al., 201324 Fremantle and colleagues performed a systematic review and network meta-analysis of therapies for osteoporosis using data from 34 randomized studies. They found that all agents significantly reduced the risk of vertebral fractures compared to placebo, alendronate and teriparatide significantly reduced non-vertebral fractures, and zoledronic acid, denosumab, and risedronate significantly reduced the risk for non-vertebral and hip fractures. NICE, 201523 David and colleagues performed a systematic review and network meta-analysis of bisphosphonate therapies for osteoporosis using data from the 27 of 46 randomized studies with fracture data. They found that all agents significantly reduced the risk of vertebral fractures compared to placebo and that there were no significant pairwise differences between active therapies. Zoledronic acid had the greatest effect on vertebral fracture rate reduction and increase in bone mineral density. Zhang et al., 201527 Zhang and colleagues performed a systematic review and network meta-analysis of teriparatide, denosumab, and oral bisphosphonates for women with postmenopausal osteoporosis using data from 15 randomized studies. Zoledronic acid was not considered. The concluded that teriparatide, denosumab, alendronate and risedronate were effective at reducing vertebral and non-vertebral ©Institute for Clinical and Economic Review, 2017 Page 83 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents fractures compared to placebo and that denosumab, alendronate and risedronate reduce the risk of hip fractures. There were no significant differences in head to head comparisons of the drugs. Yang, 201626 Yang and colleagues performed a systematic review and network meta-analysis of drugs for women with postmenopausal osteoporosis using data from 36 randomized studies. Patients treated with alendronate, denosumab, and teriparatide had significantly lower rates of non-vertebral fractures than placebo. Alendronate, zoledronic acid, and denosumab were associated with a significantly lower risk of hip fractures compared to placebo. They did not consider vertebral fractures in their analysis. ©Institute for Clinical and Economic Review, 2017 Page 84 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Appendix D. Ongoing Studies Estimated Title/ Trial Sponsor Study Design Comparators Patient Population Primary Outcomes Completion Date Abaloparatide Twenty-Four Month Open-label Alendronate N = 1,200 Incidence and October 2016 Extension Study of extension trial (following 24 Women only severity of adverse BA058-05-003 months of Patients enrolled and randomized to abaloparatide or events, fractures, A 6-month pre- (ACTIVExtend) abaloparatide placebo arm of ACTIVE trial and changes in planned interim treatment in ACTIVE No participants who withdrew from ACTIVE trial laboratory values analysis has been NCT01657162 trial) No participants with serious adverse events during published69 ACTIVE trial ©Institute for Clinical and Economic Review, 2017 Page 85 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Estimated Title/ Trial Sponsor Study Design Comparators Patient Population Primary Outcomes Completion Date Teriparatide VERtebral Fracture RCT Teriparatide 20 mcg N = 1,327 Proportion of July 2016 (study Treatment Comparisons once daily, weekly Ages 45 and older patients with new completed, but in Osteoporotic Women oral placebo, daily Postmenopausal women only vertebral fractures not yet published) (VERO) calcium and vitamin BMD ≤ -1.5 at 24 months D At least 2 moderate or 1 severe vertebral fragility NCT01709110 fractures Risedronate 35 mg No increased risk of osteosarcoma once weekly, daily No history of unresolved skeletal disease that affect placebo injection, bone metabolism daily calcium and No history of atypical femoral fractures vitamin D No abnormally high/low calcium levels No abnormally high parathyroid hormone levels No severe vitamin D deficiency No abnormal, uncorrected thyroid function No malignant neoplasms in previous 5 years No active liver disease, jaundice No significant impairment of hepatic/renal function No history of nephro/urolithiasis No previous/planned kypho/vertebroplasty No current or risk of osteonecrosis of the jaw No active or recent upper gastrointestinal disorders No inability to stand/sit upright for at least 30 minutes ©Institute for Clinical and Economic Review, 2017 Page 86 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Estimated Title/ Trial Sponsor Study Design Comparators Patient Population Primary Outcomes Completion Date Romosozumab Study to Determine the RCT Romosozumab and N = 4,093 Incidence of clinical November 2017 Efficacy and Safety of placebo Ages 55-90 fracture at 24 Romosozumab in the alendronate for 12 Postmenopausal women only months Treatment of months, then open- Hip BMD T-score of ≤ -2.5 and a vertebral fracture or hip Postmenopausal Women label alendronate BMD T-score of ≤ -2.0 and a recent hip fracture or two Incidence of new With Osteoporosis for 12+ months vertebral fractures vertebral fracture at No history of metabolic/bone disease other than 24 months NCT01631214 Alendronate and osteoporosis placebo No use of agents that affect bone metabolism romosozumab for No vitamin D insufficiency 12 months, then No prior solid organ or bone marrow transplant open-label No hypo/hypercalcemia alendronate for 12+ No hypo/hyperthyroidism months No hypo/hyperparathyroidism No intolerance to alendronate Source: www.ClinicalTrials.gov (NOTE: studies listed on site include both clinical trials and observational studies) ©Institute for Clinical and Economic Review, 2017 Page 87 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Appendix E. Comparative Clinical Effectiveness Supplemental Information We performed screening at both the abstract and full-text level. A single investigator screened all abstracts identified through electronic searches according to the inclusion and exclusion criteria described earlier. We did not exclude any study at abstract-level screening due to insufficient information. For example, an abstract that did not report an outcome of interest would be accepted for further review in full text. We retrieved the citations that were accepted during abstract-level screening for full text appraisal. One investigator reviewed full papers and provided justification for exclusion of each excluded study. We used criteria published by the US Preventive Services Task Force (USPSTF) to assess the quality of RCTs and comparative cohort studies, using the categories “good,” “fair,” or “poor” (see Appendix Table F2)106 Guidance for quality ratings using these criteria is presented below, as is a description of any modifications we made to these ratings specific to the purposes of this review. Good: Meets all criteria: Comparable groups are assembled initially and maintained throughout the study; reliable and valid measurement instruments are used and applied equally to the groups; interventions are spelled out clearly; all important outcomes are considered; and appropriate attention is paid to confounders in analysis. In addition, intention to treat analysis is used for RCTs. Fair: Studies were graded "fair" if any or all of the following problems occur, without the fatal flaws noted in the "poor" category below: Generally comparable groups are assembled initially but some question remains whether some (although not major) differences occurred with follow-up; measurement instruments are acceptable (although not the best) and generally applied equally; some but not all important outcomes are considered; and some but not all potential confounders are addressed. Intention to treat analysis is done for RCTs. Poor: Studies were graded "poor" if any of the following fatal flaws exists: Groups assembled initially are not close to being comparable or maintained throughout the study; unreliable or invalid measurement instruments are used or not applied equally among groups (including not masking outcome assessment); and key confounders are given little or no attention. For RCTs, intention to treat analysis is lacking. ©Institute for Clinical and Economic Review, 2017 Page 88 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table E1. Summary of the Randomized Trials of Anabolic Agents for Osteoporosis F/U Prior Reference Study Group N T-score (months) Fracture Teriparatide Neer 200119 Fracture Teriparatide 20 mcg SC QD 541 21 - 100% vertebral Prevention Placebo SC QD 544 Trial Abaloparatide Miller 201618 ACTIVE Abaloparatide 80 mcg SC QD 824 18 -2.5 to 24% vertebral Teriparatide 20 mcg SC QD 818 -5.0 63% any Placebo SC QD 821 Romosozumab Cosman FRAME Romosozumab 210 mg SC 3589 12 -2.5 to 18% vertebral 201617 Qmo 3591 -3.5 22% non- Placebo SC Qmo vertebral Key comparator: Zoledronic acid Black 200720 HORIZON Zoledronic acid 5 mg IV Q 3889 36 -2.5 or 63% vertebral year 3876 lower Placebo IV Q year F/U: follow-up, QD: once daily; Qmo: once monthly, Q year: once yearly ©Institute for Clinical and Economic Review, 2017 Page 89 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table E2. Inclusion/Exclusion Criteria for the Randomized Trials of Anabolic Agents for Osteoporosis Reference Study Inclusion Exclusion Co-intervention Teriparatide Neer 200119 Fracture Female Illnesses that affect Vitamin D 400- Prevention 5+ years postmenopausal bone 1200 IU daily Trial ≥ 1 moderate or 2 mild V Fx Kidney stone in past Calcium 1000 If ≤1 moderate V Fx, then additionally 5 years mg daily T-score < -1.0 Cr > 2.0 mg/dL Liver disease Substance abuse Recent use of drugs for osteoporosis Abaloparatide Miller 201618 ACTIVE Female More than 4 V Fx None Postmenopausal Illnesses that affect Ages 49-86 years bone T-score -2.5 to -5.0 Recent use of drugs ≥ 1 moderate or 2 mild V Fx or other for osteoporosis fragility fracture in past 5 years Women ≥ 65 years with fracture eligible if T-score ≤ - 2.0 and > -5.0 Women ≥ 65 years without fracture if T-score ≤ - 3.0 and > -5.0 Normal serum calcium, PTH, phosphorus, alkaline phosphatase, and vitamin D levels Romosozumab Cosman FRAME Female Hip fracture None 201617 Postmenopausal Severe or >2 Ages 55-90 years moderate V Fx T-score -2.5 to -3.5 Illnesses that affect Normal serum calcium, PTH, bone phosphorus, alkaline phosphatase, ONJ and 25(OH) vitamin D levels Low vitamin D Recent use of drugs for osteoporosis Key comparator: Zoledronic acid Black 200720 HORIZON Female Use of PTH or Vitamin D 400- Postmenopausal sodium fluoride 1200 IU daily Ages 65-89 years Recent use of Calcium 1000 – T-score -2.5 to -3.5 corticosteroids 1500 mg daily CrCl < 30 ml/min Cr: creatinine, ONJ: osteonecrosis of the jaw, PTH: parathyroid hormone, V Fx: vertebral fracture ©Institute for Clinical and Economic Review, 2017 Page 90 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table E3. Baseline Characteristics of Patients in Randomized Trials of Anabolic Agents for Osteoporosis Current Prior BMD, Prior % BMI, Reference Group Age %F Smoker, treatment, mg/cm2 V Fx, W kg/m2 % % L-Spine n Teriparatide Neer 200119 Teriparatide 69 100 99 26.8 15.8 16 820 2.3 Placebo 69 100 99 26.7 18.5 15 820 2.3 Abaloparatide Miller 201618 Abaloparatide 69 100 80 25.0 NR NR 829 NR Teriparatide 69 100 79 25.2 NR NR 831 NR Placebo 69 100 80 25.1 NR NR 823 NR Romosozumab Cosman 201617 Romosozumab 71 100 NR 24.7 NR NR NR NR Placebo 71 100 NR 24.7 NR NR NR NR Key comparator: Zoledronic acid Black 200720 Zoledronic 73 100 NR 25.1 NR 59% 790 NR acid 73 100 NR 25.4 NR 59% 790 NR Placebo BMD: bone mineral density, BMI: body mass index, F: female, W: white, V Fx: vertebral fractures ©Institute for Clinical and Economic Review, 2017 Page 91 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table E4. Quality Assessment of the Included Randomized Trials of Anabolic Agents for Osteoporosis Comparable Maintain Double Measurements Clear Definition of Key Outcomes Analysis Reference Quality Groups Comparability Blind Equal and Valid Intervention Assessed Appropriate Teriparatide Neer Yes Yes Yes Yes Yes Yes Yes Good 2001 19 Abaloparatide Miller Yes Yes Yes* Yes Yes Yes Yes Good vs. 2016 18 placebo Romosozumab Cosman Yes Yes Yes Yes Yes Yes Yes Good 201617 Key comparator: Zoledronic acid Black Yes Yes Yes Yes Yes Yes Yes Good 200720 *Open-label teriparatide, double-blind abaloparatide and placebo ©Institute for Clinical and Economic Review, 2017 Page 92 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table E5. Risk for Fracture in the Randomized Trials of Anabolic Agents for Osteoporosis Reference Group V Fx Non-V Fx Hip Fx Wrist Fx Major osteoporotic Fx Clinical fracture Teriparatide Neer 200119 Teriparatide 22 (5.0%) 14 (2.6%) 1 (0.2%) 7 (1.3%) NR NR 21 months Placebo 64 (14.3%) 30 (5.5%) 4 (0.7%) 13 (2.4%) Prevrhal 200921 Teriparatide 8 (1.8%)* Placebo 51 (11.4%) Abaloparatide Miller 201618 Abaloparatide 4 (0.6%) 18 (2.7%) 0 (0%) 7 (0.8%) 10 (1.5%) 27 (4.0%) 18 months Teriparatide 6 (0.8%) 24 (3.3%) 0 (0%) 17 (2.1% 23 (3.1%) 35 (4.8%) Placebo 30 (4.2%) 33 (4.7%) 2 (0.2%) 15 (1.8%) 34 (6.2%) 49 (8.3%) Romosozumab Cosman 201617 Romosozumab 16 (0.5%) 56 (1.6%) 7 (0.2%) NR 38 (1.1%) 58 (1.6%) 12 months Placebo 59 (1.8%) 75 (2.1%) 13 (0.4%) 63 (1.8%) 90 (2.5%) Key comparator: Zoledronic acid Black 200720 Zoledronic acid 92 (3.3%) 292 (8.0%) 52 (1.4%) NR NR 308 (8.4%) 36 months Placebo 310 (10.9%) 388 (10.7%) 88 (2.5%) 456 (12.8%) NR: not reported, V Fx: vertebral fracture, Non-V Fx: non-vertebral, non-hip fractures * Using alternative definition for incident vertebral fractures: decrease in height of at least 20% and 4 mm using quantitative morphometry plus an increase in grade by the semiquantitative assessment. The primary analysis (Neer 2001) used a single reader increase in grade using the semiquantitative assessment of vertebral fracture. ©Institute for Clinical and Economic Review, 2017 Page 93 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table E6. Relative Risk for Fractures in the Randomized Trials of Anabolic Agents for Osteoporosis Reference Group V Fx Non-V Fx Hip Fx Major osteoporotic Fx Clinical fracture Teriparatide Neer 200119 Teriparatide 0.35 (0.22-0.55) 0.47 (0.25-0.88) NR NR NR Placebo 1 (ref) 1 (ref) 1 (ref) 1 (ref) 1 (ref) Prevrhal 200921 Teriparatide 0.16 (0.08-0.33)* Placebo 1 (ref) Abaloparatide Miller 201618 Abaloparatide 0.14 (0.05-0.39) 0.57 (0.32-1.00) NR 0.30 (0.15-0.61) 0.57 (0.35-0.91) Teriparatide 0.20 (0.08-0.47) 0.72 (0.42-1.22) NR 0.67 (0.39-1.14) 0.71 (0.46-1.09) Placebo 1 (ref) 1 (ref) 1 (ref) 1 (ref) 1 (ref) Romosozumab Cosman 201617 Romosozumab 0.27 (0.16-0.47) 0.75 (0.53-1.05) 0.54 (0.22-1.35) 0.60 (0.40-0.90) 0.64 (0.46-0.89) Placebo 1 (ref) 1 (ref) 1 (ref) 1 (ref) 1 (ref) Key comparator: Zoledronic acid Black 200720 Zoledronic acid 0.30 (0.24-0.38) 0.75 (0.64-0.87) 0.59 (0.42-0.83) 0.67 (0.58-0.77) Placebo 1 (ref) 1 (ref) 1 (ref) NR NR: not reported, ref: referent group, V Fx: vertebral fracture, Non-V Fx: non-vertebral, non-hip fractures * Using alternative definition for incident vertebral fractures: decrease in height of at least 20% and 4 mm using quantitative morphometry plus an increase in grade by the semiquantitative assessment. The primary analysis (Neer 2001) used a single reader increase in grade using the semiquantitative assessment of vertebral fracture. ©Institute for Clinical and Economic Review, 2017 Page 94 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table E7. Bone Mineral Density Outcomes in Randomized Trials of Anabolic Agents for Osteoporosis Reference Group BMD L spine BMD femoral neck BMD total hip Teriparatide Neer 200119 Teriparatide +9.7% +2.8% +2.6% 21 months Placebo +1.1% -0.7% -1.0% Abaloparatide Miller 201618 Abaloparatide +11.2% +3.6% +4.2% 18 months Teriparatide +10.5% +2.7% +3.3% Placebo +0.6% -0.4% -0.1% Romosozumab Cosman 201617 Romosozumab 13.3 % 5.9% difference 6.9% 12 months Placebo difference difference Key comparator: Zoledronic acid Black 200720 Zoledronic acid 6.7% 5.1% difference 6.0% 36 months Placebo difference difference BMD: bone mineral density, L spine: lumbar spine Table E8. Network Meta-Analysis Results for the Relative Risk of Morphometric Vertebral Fractures, Excluding Open-Label Teriparatide Arm from ACTIVE Trial Abaloparatide (80 mcg) 0.91 Teriparatide (0.22 – 3.20) (20 mcg) 0.45 0.51 Zoledronic Acid (0.13 – 1.21) (0.22 – 1.00) (5 mg) 0.13 0.15 0.30 Placebo (0.04 – 0.34) (0.07 – 0.28) (0.24 – 0.37) Fixed-effects model; resdev = 5.352, DIC = 43.663 Legend: Each box represents the estimated rate ratio and 95% credible interval for the combined direct and indirect comparisons between two drugs: the drug at the top of the column compared to the drug at the right of the row. Estimates in bold signify that the 95% credible interval does not contain 1. ©Institute for Clinical and Economic Review, 2017 Page 95 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table E9. Network Meta-Analysis Results for the Relative Risk of Non-Vertebral Fractures, Excluding Open-Label Teriparatide Arm from ACTIVE Trial Teriparatide (20 mcg) 0.81 Abaloparatide (0.34 – 1.87) (80 mcg) 0.62 0.75 Zoledronic Acid (0.31 – 1.09) (0.41 – 1.30) (5 mg) 0.45 0.55 0.75 Placebo (0.23 – 0.81) (0.31 – 0.95) (0.64 – 0.86) Fixed-effects model; resdev = 5.387, DIC = 46.775 Legend: Each box represents the estimated rate ratio and 95% credible interval for the combined direct and indirect comparisons between two drugs: the drug at the top of the column compared to the drug at the right of the row. Estimates in bold signify that the 95% credible interval does not contain 1. Table E10. Network Meta-Analysis Results for the Relative Risk of Morphometric Vertebral Fractures, Including Data from the VERO Trial Abaloparatide (80 mcg) 0.88 Teriparatide (0.26 – 2.38) (20 mcg) 0.41 0.47 Bisphosphonate* (0.13 – 1.09) (0.33 – 0.65) 0.13 0.14 0.31 Placebo (0.03 – 0.32) (0.10 – 0.21) (0.25 – 0.38) *Key assumption: the fracture prevention effects of zoledronic acid and risedronate are similar and represent a class effect for bisphosphonates. Fixed-effects model; resdev = 7.382, DIC = 62.014 Legend: Each box represents the estimated rate ratio and 95% credible interval for the combined direct and indirect comparisons between two drugs: the drug at the top of the column compared to the drug at the right of the row. Estimates in bold signify that the 95% credible interval does not contain 1. ©Institute for Clinical and Economic Review, 2017 Page 96 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table E11. Network Meta-Analysis Results for the Relative Risk of Non-Vertebral Fractures, Including Data from the VERO Trial Abaloparatide (80 mcg) 0.87 Teriparatide (0.49 – 1.51) (20 mcg) 0.66 0.75 Bisphosphonate* (0.37 – 1.09) (0.55 – 1.01) 0.49 0.56 0.76 Placebo (0.29 – 0.81) (0.41 – 0.76) (0.65 – 0.87) *Key assumption: the fracture prevention effects of zoledronic acid and risedronate are similar and represent a class effect for bisphosphonates. Fixed-effects model; resdev = 8.135, DIC = 66.184 Legend: Each box represents the estimated rate ratio and 95% credible interval for the combined direct and indirect comparisons between two drugs: the drug at the top of the column compared to the drug at the right of the row. Estimates in bold signify that the 95% credible interval does not contain 1. ©Institute for Clinical and Economic Review, 2017 Page 97 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table E12. NMA Sensitivity Analyses: Morphometric Vertebral Fracture Comparisons to Placebo Sensitivity Analysis Sensitivity Analysis Sensitivity Analysis Excluding Study Random Effects, Random Effects, Using Neer 2001 Using Unpublished Fixed Effects Teriparatide Arm of Drug Publication RR Vague Priors Informative Priors Teriparatide Data, VERO Data, RR (95% CrI) ACTIVE Trial, Fixed (95% CrI) RR (95% CrI) RR (95% CrI) Fixed Effects, Fixed Effects, Effects RR (95% CrI) RR (95% CrI) RR (95% CrI) Abaloparatide 0.14 0.13 0.13 0.13 0.13 0.14 0.13 (80 mcg) (0.05 – 0.39) (0.03 – 0.33) (0.01 – 0.95) (0.03 – 0.38) (0.04 – 0.34) (0.04 – 0.35) (0.03 – 0.32) Teriparatide* 0.16 0.17 0.17 0.17 0.15 0.30 0.14 (20 mcg) (0.08 – 0.33) (0.09 – 0.29) (0.03 – 0.75) (0.09 – 0.34) (0.07 – 0.28) (0.19 – 0.45) (0.10 – 0.21) Zoledronic Acid 0.30 0.30 0.30 0.30 0.30 0.30 0.31† (5 mg) (0.24 – 0.38) (0.24 – 0.37) (0.03 – 1.94) (0.15 – 0.55) (0.24 – 0.37) (0.24 – 0.38) (0.25 – 0.38) CrI: credible interval, RR: relative risk *Teriparatide results were calculated using Prevrhal, 200921 with the exception of the final column, which used data from Neer, 2001.19 Estimates in bold signify that the 95% credible interval does not contain 1. †Key assumption: the fracture prevention effects of zoledronic acid and risedronate are similar and represent a class effect for bisphosphonates. ©Institute for Clinical and Economic Review, 2017 Page 98 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table E13. NMA Sensitivity Analyses: Non-Vertebral Fracture Comparisons to Placebo Sensitivity Analysis Sensitivity Analysis Using Study Random Effects, Random Effects, Excluding Teriparatide Fixed Effects Unpublished VERO Data, Drug Publication RR Vague Priors Informative Priors Arm of ACTIVE Trial, Fixed RR (95% CrI) Fixed Effects, (95% CrI) RR (95% CrI) RR (95% CrI) Effects RR (95% CrI) RR (95% CrI) Abaloparatide 0.57* 0.51 0.50 0.50 0.55 0.49 (80 mcg) (0.32 – 1.00) (0.28 – 0.85) (0.07 – 2.80) (0.23 – 1.04) (0.31 – 0.95) (0.29 – 0.81) Teriparatide 0.47 0.61 0.60 0.60 0.45 0.56 (20 mcg) (0.25 – 0.88) (0.41 – 0.88) (0.13 – 2.32) (0.34 – 1.04) (0.23 – 0.81) (0.41 – 0.76) Zoledronic Acid 0.75* 0.75 0.75 0.75 0.75 0.76† (5 mg) (0.64 – 0.87) (0.64 – 0.87) (0.10 – 4.08) (0.40 – 1.36) (0.64 – 0.86) (0.65 – 0.87) CrI: credible interval, NR: not reported, RR: relative risk *Denotes use of hazard ratios instead of relative risks; RRs were not reported in the trial publication. †Key assumption: the fracture prevention effects of zoledronic acid and risedronate are similar and represent a class effect for bisphosphonates. Estimates in bold signify that the 95% credible interval does not contain 1. ©Institute for Clinical and Economic Review, 2017 Page 99 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Appendix F. Comparative Value Supplemental Information Table F1. Detailed Results Per Regimen Zoledronic Acid Abaloparatide Teriparatide Deterministic Credible Range Deterministic Credible Range Deterministic Credible Range (from PSA) (from PSA) (from PSA) Total Cost $25,465 ($20,844 - $31,176) $47,525 ($40,081 - $56,361) $68,905 ($58,313 - $80,110) Anabolic Cost -- $27,574 ($22,185 - $33,303) $47,159 ($37,791 - $56,803) Zoledronic $2,498 ($2,150 - $2,895) $2,243 ($1,931 - $2,600) $2,243 ($1,931 - $2,600) Acid Cost Hip Fracture $7,276 ($5,058 - $9,910) $5,708 ($3,594 - $8,877) $6,211 ($4,029 - $9,120) Cost Clinical Vert $1,202 ($863 - $1,617) $806 ($510 - $1,303) $864 ($591 - $1,248) Fracture Cost Other Non- $5,084 ($3,896 - $6,538) $4,147 ($2,739 - $6,201) $4,556 ($3,164 - $6,348) Vertebral Fracture Cost Post-Fracture $9,404 ($6,636 - $13,116) $7,048 ($4,209 - $11,550) $7,871 ($5,064 - $11,849) Cost Total QALYs 8.93 (7.54 - 10.13) 9.00 (7.60 - 10.21) 8.98 (7.58 - 10.18) Pre-Fracture 6.29 (5.16 - 7.28) 6.96 (5.62 - 8.14) 6.72 (5.50 - 7.81) QALYs Hip Fracture 0.08 (0.06 - 0.12) 0.07 (0.04 - 0.11) 0.07 (0.04 - 0.11) QALYs Vert Fracture 0.05 (0.04 - 0.07) 0.04 (0.02 - 0.06) 0.04 (0.03 - 0.05) QALYs Other Fracture 0.27 (0.20 - 0.34) 0.22 (0.14 - 0.33) 0.24 (0.16 - 0.33) QALYs Post-Fracture 2.24 (1.81 - 2.67) 1.72 (1.23 - 2.29) 1.91 (1.46 - 2.41) QALYs Lifetime Cumulative Fracture Probabilities Hip Fractures 0.24 (0.18 - 0.31) 0.19 (0.13 - 0.28) 0.21 (0.15 - 0.29) Clinical Vert 0.18 (0.15 - 0.22) 0.13 (0.09 - 0.19) 0.14 (0.11 - 0.17) Fractures Other Non- 0.54 (0.46 - 0.63) 0.46 (0.33 - 0.64) 0.50 (0.38 - 0.64) Vertebral Fractures ©Institute for Clinical and Economic Review, 2017 Page 100 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table F2. Detailed Incremental Results versus Zoledronic Acid Abaloparatide Teriparatide Deterministic Credible Range (from PSA) Deterministic Credible Range (from PSA) ICER $333,892 (-$1,183,554 - $2,480,241) $941,537 (-$4,498,030 - $7,758,067) Incremental Cost $22,061 ($14,728 - $30,542) $43,440 ($33,081 - $54,238) Anabolic Cost $27,574 ($22,185 - $33,303) $47,159 ($37,791 - $56,803) Zoledronic Acid Cost -$255 (-$295 - -$219) -$255 (-$295 - -$219) Hip Fracture Cost -$1,568 (-$3,775 - $1,023) -$1,065 (-$3,047 - $1,046) Clinical Vert Fracture -$396 (-$698 - $37) -$339 (-$572 - -$94) Cost Other Non-Vertebral -$937 (-$2,340 - $836) -$528 (-$1,747 - $824) Fracture Cost Post-Fracture Cost -$2,356 (-$5,941 - $1,771) -$1,533 (-$4,733 - $1,846) Incremental QALYs 0.07 (-0.01 - 0.15) 0.05 (-0.01 - 0.11) Pre-Fracture QALYs 0.68 (0.05 - 1.22) 0.43 (-0.04 - 0.87) Hip Fracture QALYs -0.02 (-0.05 - 0.01) -0.01 (-0.04 - 0.01) Vert Fracture QALYs -0.02 (-0.03 - 0.00) -0.02 (-0.03 - 0.00) Other Fracture QALYs -0.05 (-0.12 - 0.04) -0.03 (-0.09 - 0.04) Post-Fracture QALYs -0.52 (-0.97 - -0.01) -0.33 (-0.68 - 0.05) Lifetime Cumulative Fracture Probabilities Hip Fractures -0.05 (-0.11 - 0.03) -0.03 (-0.09 - 0.03) Clinical Vert Fractures -0.05 (-0.08 - 0.00) -0.05 (-0.07 - -0.01) Other Non-vertebral -0.08 (-0.21 - 0.09) -0.05 (-0.16 - 0.09) Fractures ©Institute for Clinical and Economic Review, 2017 Page 101 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Supplemental Scenario Analyses Years Maintaining Full Treatment Effect After Stopping Bisphosphonate Treatment The base-case analyses assumed that the treatment effect of anabolic agents is maintained by follow-up treatment with zoledronic acid. Once zoledronic acid is stopped, we assumed the anabolic treatment effect is maintained for another three years before declining. Given the uncertainty in this assumption, we varied the duration of full treatment effect post-zoledronic acid from 0-10 years. Figure F1 shows how the incremental cost-effectiveness ratio of each anabolic treatment declines with longer duration of full treatment effect post-zoledronic acid treatment. The incremental cost-effectiveness ratios corresponding with the three-year post-bisphosphonate treatment reflect the base case scenario. Regardless of the assumed duration of effect, the incremental cost-effectiveness ratios did not approach $150K per QALY. Figure F1. Results of Anabolic Treatment Efficacy Maintenance Scenario Analysis $2,500,000 Base Case $2,000,000 $1,500,000 ICER Teriparatide $1,000,000 Abaloparatide $500,000 $0 0 years 1 years 2 years 3 years 4 years 5 years 6 years 7 years 8 years 9 years10 years Years Maintaining Full Efficacy Post-Bisphosphonate Rate of Treatment Effect Decline Another key assumption was the rate of treatment effect decline over time once zoledronic acid therapy is stopped. Figure F2 below shows how the ICER for each anabolic treatment varied with the number of years it takes for the treatment effect to decline from full treatment effect to the baseline fracture rates, assuming the decline starts 3-years post-bisphosphonate treatment and declines linearly. The incremental cost-effectiveness ratios corresponding with the 10-year decline time reflect the base case scenario. Similar to the scenario analysis above, the ICERs stay well above the upper cost-effectiveness threshold of $150,000 per QALY. ©Institute for Clinical and Economic Review, 2017 Page 102 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Figure F2. Number of Years of Efficacy Decline Duration Scenarios $9,000,000 Base Case $8,000,000 $7,000,000 $6,000,000 $5,000,000 ICER $4,000,000 Teriparatide $3,000,000 Abaloparatide $2,000,000 $1,000,000 $0 0 years 4 years 8 years 12 years 16 years 20 years Rate of Treatment Efficacy Decline Ramp-Up Time for Efficacy of Zoledronic Acid We explored the impact of various assumptions regarding the rate at which zoledronic acid reaches full efficacy in the baseline comparator arm (Figure F3). All three anabolic regimens’ incremental cost-effectiveness ratios improved the longer it took zoledronic acid to reach full efficacy, as expected. However, even with 10 years’ ramp-up time for zoledronic acid, the anabolic agents did not reach the $150,000 per QALY threshold. The following scenario analysis, comparison of anabolics to no treatment, further explains this result. Figure F3. Zoledronic Acid Ramp-Up Time Scenarios $1,000,000 $900,000 $800,000 $700,000 $600,000 ICER $500,000 Teriparatide $400,000 Abaloparatide $300,000 $200,000 $100,000 $0 1 years 2 years 3 years 4 years 5 years 6 years 7 years 8 years 9 years 10 years Ramp-up Time for Zoledronic Acid Efficacy (base case = 0) ©Institute for Clinical and Economic Review, 2017 Page 103 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Excluding Open-Label Teriparatide Data from ACTIVE Trial in NMA RR Estimates The exclusion of the ACTIVE trial’s teriparatide results in the NMA resulted in slightly different relative risk estimates for teriparatide and abaloparatide, presented below. We calculated hip fracture relative risk estimates based on the ratio of hip to non-vertebral fracture relative risks in the HORIZON trial, similar to the base case approach. This scenario resulted in a small decline in incremental QALYs and small increase in cost for abaloparatide, and an increase in incremental QALYs with decreased cost for teriparatide. Table F3. Model Inputs for Scenario Analysis Excluding Open-Label Teriparatide Data from ACTIVE Trial Drug Vertebral Fracture RR Non-Vertebral Fracture RR Zoledronic Acid 5 mg 0.30 (0.24 – 0.37) 0.75 (0.64 – 0.86) Teriparatide 20 mcg 0.15 (0.07 – 0.28) 0.45 (0.23 – 0.81) Abaloparatide 80 mcg 0.13 (0.04 – 0.34) 0.55 (0.31 – 0.95) RR: relative risk Table F4. Results of Scenario Analysis Excluding Open-Label Teriparatide Data Regimen Cost QALYs Life Years Zoledronic Acid $25,465 8.933 12.188 Teriparatide $66,010 9.008 12.196 Abaloparatide $48,183 8.992 12.194 QALY: quality-adjusted life year Table F5. Pairwise Results of Anabolic Drugs Compared to Zoledronic Acid, Excluding Open-Label Teriparatide Data Regimen Incr. Cost Incr. QALYs Incr. LYs ICER vs. Zoledronic Acid Teriparatide $40,545 0.076 0.008 $535,758 Abaloparatide $22,718 0.060 0.006 $380,332 Incr.: incremental, LY: life year, QALY: quality-adjusted life year Including Zoledronic Acid-Associated Infusion Reaction Although we found little evidence of significant differences in adverse event rates between modeled comparators and placebo, multiple clinical stakeholders indicated that infusion reactions following zoledronic acid administration were a potentially significant adverse event that warranted consideration. In this scenario, we used the approach employed by the National Institute for Health and Care Excellence (NICE) in their systematic review and cost-effectiveness analysis of bisphosphonates.23 Briefly, they assumed a disutility of 0.30 for 3 days for flu-like symptoms associated with IV bisphosphonates, which is equivalent to a QALY loss of 0.005. They applied this as a fixed QALY decrement at the start of the model without adjustment for baseline utility. The ©Institute for Clinical and Economic Review, 2017 Page 104 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents rate of influenza-like symptoms was assumed to be the differential rate of pyrexia reported in the HORIZON-PFT study (14%).20 This small disutility for zoledronic acid had little impact on our model results. Table F6. Results of Scenario Analysis Including Zoledronic Acid-Associated Infusion Reaction Cost QALYs Life Years Zoledronic Acid $25,465 8.932 12.188 Teriparatide $68,905 8.979 12.193 Abaloparatide $47,525 8.999 12.195 QALY: quality-adjusted life year Table F7. Pairwise Results of Anabolic Drugs Compared to Zoledronic Acid, Including Zoledronic Acid-Associated Infusion Reaction Incr. Cost Incr. QALYs Incr. LYs ICER Teriparatide $43,440 0.047 0.005 $927,466 Abaloparatide $22,061 0.067 0.007 $330,391 Incr.: incremental, LY: life year, QALY: quality-adjusted life year Baseline Fracture Risk in Higher-Risk Patient Groups who are Intolerant of Zoledronic Acid In this scenario, we set zoledronic acid relative risks estimates to 1, set the zoledronic acid costs to zero, “turned off” post-anabolic zoledronic acid treatment, and modeled a range of increased baseline fracture probabilities to explore the impacts of anabolics on high risk patients who cannot tolerate zoledronic acid either as primary or subsequent therapy. Even at 100% increased baseline fracture probability, the ICERs for abaloparatide and teriparatide did not reach the $150,000 per QALY threshold. ©Institute for Clinical and Economic Review, 2017 Page 105 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Figure F4. Baseline Fracture Risk Scenarios in Higher-Risk Groups Intolerant of Zoledronic Acid $800,000 $700,000 $600,000 $500,000 ICER $400,000 Teriparatide $300,000 Abaloparatide $200,000 $100,000 $0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Increase in Baseline Fracture Risk Inclusion of a Disutility for Morphometric Vertebral Fractures A number of stakeholders indicated that patients who have morphometric fractures experience a small disutility, and that we should include this in our model. In this scenario analysis, we added a morphometric utility multiplier and varied it over a wide range of values, from no disutility (multiplier = 1) down to an extreme value of 0.8. Because this was applied to all comparators, and because the relative risk for vertebral fracture were generally similar among comparators, the differences in utilities were largely “washed out”, and had little impact on ICER results. Figure F5. Inclusion of a Disutility for Morphometric Vertebral Fractures $1,000,000 $900,000 $800,000 $700,000 $600,000 ICER $500,000 Teriparatide $400,000 Abaloparatide $300,000 $200,000 $100,000 $0 0.80 0.82 0.84 0.86 0.88 0.90 0.92 0.94 0.96 0.98 1.00 Morphometric Vertebral Fracture Utility Multiplier ©Institute for Clinical and Economic Review, 2017 Page 106 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Scenario Analysis Using Results of the VERO Trial A number of stakeholders suggested using the results of the VERO trial comparing teriparatide and an oral bisphosphonate (risedronate) in the treatment of patients with severe postmenopausal osteoporosis. The NMA was updated by incorporating the VERO results for vertebral and non- vertebral fractures. Similar to the base case analysis, we imputed the relative risk estimate for hip fractures based on the ratio of hip versus non-vertebral fractures observed in the HORIZON trial. The corresponding relative risk estimates for fractures are shown in Table F8. Table F8. Model Inputs for Scenario Analysis Using Results of the VERO Trial in the NMA Vertebral Fracture Non-vertebral Hip Fracture RR RR Fracture RR Bisphosphonate (risedronate [VERO] + 0.31 0.76 0.60 zoledronic acid [HORIZON]) (0.25 - 0.38) (0.65 - 0.87) (0.42 - 0.83) Teriparatide 20 mcg 0.14 0.56 0.44 (0.10 - 0.21) (0.41 - 0.76) (0.28 - 0.75) Abaloparatide 80 mcg 0.13 0.49 0.39 (0.03 - 0.32) (0.29 - 0.81) (0.17 - 0.74) RR: relative risk The results (Tables F9 and F10) show that using the data from the VERO study does not meaningfully change the model outcomes and produces ICERs that are above the $150,000 per QALY threshold. Table F9. Results of Scenario Analysis Incorporating VERO Data Regimen Cost QALYs Life Years Zoledronic Acid $25,651 8.930 12.188 Teriparatide $67,990 8.990 12.194 Abaloparatide $47,248 9.001 12.195 QALY: quality-adjusted life year Table F10. Pairwise Results of Scenario Analysis Incorporating VERO Data Regimen Incr. Cost Incr. QALYs Incr. LYs ICER vs. Zoledronic Acid Teriparatide $42,339 0.059 0.006 $714,343 Abaloparatide $21,597 0.071 0.008 $304,039 Incr.: incremental, LY: life year, QALY: quality-adjusted life year Scenario Analysis using Fracture Risks from a Cohort of US Administrative Medicare Claims We used the fracture risks reported in a retrospective observational cohort study of US administrative claims data from fee-for-service Medicare beneficiaries (study period: January 1, 2010 to December 31, 2012).107 This study included 103,852 postmenopausal women, with an ©Institute for Clinical and Economic Review, 2017 Page 107 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents average age of 82 years (±8.2 years). A majority had hip, pelvis, or femur fractures (53%), followed by other non-vertebral fractures (28%) and clinical vertebral fractures (19%). This cohort has a lower annual risk for any fracture in this population (8.3%), compared to the baseline fracture risk in our primary analysis (10.2%) in the 80- to 84-year age group in our base-case model. The annual hip and clinical vertebral fracture risks as reported in the study, however, were higher than those for the 80- to 84-year-old age group in the base-case model (i.e., 0.053 vs. 0.023 for hip and 0.041 vs. 0.027 for vertebral). The annual risk for other fracture was markedly lower in the study (i.e. 0.023 in the claims study vs. 0.053 in the base-case model). When we used fracture risk inputs from this study for a cohort of 82-year-old women, the cumulative lifetime hip fracture rate for patients was extremely high. For example, patients treated with abaloparatide or teriparatide had cumulative lifetime hip fracture rates rates of of 23% and 27%, respectively, and patients treated with zoledronic acid had a rate of approximately 33%. The ICERs resulting from this scenario remained above the $150,000 per QALY threshold for both abaloparatide and teriparatide (Tables F11 and F12). Table F11. Results of Scenario Analysis Incorporating Fracture Risks from a Cohort of US Administrative Medicare Claims Regimen Cost QALYs Life Years Zoledronic Acid $26,015 4.612 6.690 Teriparatide $66,561 4.672 6.723 Abaloparatide $45,071 4.712 6.745 QALY: quality-adjusted life year Table F12. Pairwise Results of Scenario Analysis Incorporating Fracture Risks from a Cohort of US Administrative Medicare Claims Regimen Incr. Cost Incr. QALYs Incr. LYs ICER vs. Zoledronic Acid Teriparatide $40,546 0.060 0.033 $677,573 Abaloparatide $19,056 0.099 0.055 $191,637 Incr.: incremental, LY: life year, QALY: quality-adjusted life year ©Institute for Clinical and Economic Review, 2017 Page 108 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Appendix G. Public Comments This section includes summaries of the public comments prepared for the CTAF Public Meeting on June 30, 2016 in Los Angeles, CA. These summaries were prepared by those who delivered the public comments at the meeting and are presented in order of delivery. Three speakers did not submit summaries of their public comments. A video recording of all comments can be found here, beginning at minute 01:18:00. Conflict of interest disclosures are included at the bottom of each statement for each speaker who is not employed by a pharmaceutical manufacturer. John Krege, MD, FAHA, Eli Lilly and Co. Medical Fellow, Eli Lilly and Co. Anabolics are for patients at high risk of fracture and in subsets from the Fracture Prevention Trial (FPT) where placebo patients showed extreme fracture risk, teriparatide patients showed low risk; this has not been reported with antiresorptives. Although ICER costs out 24 months of teriparatide, ICER does not account for additional anabolic efficacy during months 19-24 since the FPT included a median 19 months of treatment. ICER focuses on the pivotal fracture clinical trials for the anabolics and zoledronic acid, but zoledronic acid has not been shown to have fracture superiority over other antiresorptives, has no bone anabolic properties, and evidence does not exist showing efficacy in patients at extremes of fracture risk. ICER includes blinded data except for some teriparatide open- label results; this open-label data is not of high quality according to ICER Appendix Table E4 and should be excluded. The VERO study is the most important study to address ICER’s questions and it has been accepted for publication in the Lancet. VERO was a 24-month, double-blind, randomized trial of post-menopausal women with a severe or two moderate spine fractures. Patients treated with teriparatide had fewer vertebral fractures (56% RRR) and fewer clinical fractures (52% RRR) versus bisphosphonate; these results should be included in the network meta-analysis. Real-world evidence for teriparatide illustrate many important points including fracture risk decreases during ongoing teriparatide treatment and falls further after stopping, and ICER should consider this evidence. Holistically, teriparatide shows marked efficacy benefit versus bisphosphonate treatment in high-risk patients with osteoporosis. Lorraine Fitzpatrick, MD, Radius Health Chief Medical Officer The model proposed by ICER misses several very important considerations related to the disease etiology and progression: ©Institute for Clinical and Economic Review, 2017 Page 109 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents • First, the populations being compared do not have a similar probability of benefiting from treatments being compared. For women at high risk of fracture, many classes of medications provide only a moderate rate of return so anabolics should be compared with anabolics, not with antiresorptives. • Second, ICER has overlooked onset of action – as a result there is an overestimation of the clinical benefit associated with slow onset of action. High-risk patients experience a 5 times elevation in fracture risk during the first year after an incident fracture and as high as 17 times elevation in risk of a hip fracture during the first month after a wrist fracture and may not benefit from treatments with slower onset of action. • Third, the fear of rare but real side effects with the antiresorptive agents such as bisphosphonates and denosumab have limited patient acceptance and compliance. ICER needs to consider the rapid loss of benefit and a rebound in fracture rates following treatment discontinuation. • Lastly, the incremental cost of fractures and downstream effects of fracture were not considered in the model. As noted at ISPOR US 2017 annual meeting, having a single threshold for cost effectiveness evaluation is not sufficient, but instead a broader look at the affordability may be necessary to fully assess the value of the medication. Jorge Arellano, MSc, MPhil, Amgen Executive Director, Global Health Economics Medical science is useful only when it’s done in the service of patients. Value assessments should address patients and science; we believe this ICER assessment misses both. Had ICER put the patient first, it would have recognized the significant value of bone-forming agents in the treatment of osteoporosis. We have four critical areas of concern: 1. The choice of zoledronic acid as the main comparator is a fundamental flaw, as it is a bisphosphonate, not a bone-forming agent. Bisphosphonates slow bone loss, they do not directly build new bone. 2. The urgency to treat these women is ignored by assuming the time to effect for zoledronic acid is as rapid as bone-forming agents. This is in direct conflict with existing clinical evidence and disregards general scientific understanding of how these medicines work. 3. The experience of patients and clinicians is not sufficiently represented in the assessment. The conclusions appear to be based on multiple unrealistic and ungrounded assumptions. ©Institute for Clinical and Economic Review, 2017 Page 110 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents 4. The report underestimates the impact of fractures on mortality and undervalues the importance of reducing the risk of these debilitating events. ICER chose to publish this assessment before key forthcoming evidence is available; ASBMR and other stakeholders call this assessment significantly premature. The fact is, choices made by ICER for this assessment, by design, undervalue bone-forming agents. We ask the panel to consider these critical issues when deliberating on the value of bone-forming agents. And that they implore ICER to acknowledge that this assessment, in its current form, is fundamentally flawed. Benjamin Leder, MD, American Society for Bone and Mineral Research, Massachusetts General Hospital Chair, Professional Practice Committee (ASBMR) We are in the midst of a crisis in the treatment of osteoporosis, a devastating disease that is responsible for more than 1.5 million fractures in the U.S. each year, including 300,000 hip fractures. We are concerned that this ICER report and the timing of its publication may increase the barriers between patients and effective treatments. The current ICER report is premature and relies on only a handful of placebo-controlled trials and no comparative efficacy studies. Moreover, 2 large comparative efficacy trials comparing the anti- fracture benefit of anabolic therapies to bisphosphonates are complete and will be published soon. This report would be much stronger if it included this data. We strongly recommend that ICER postpone the publication of this report until the results of these trials are reported. If ICER proceeds with its publication at this time, we recommend that the report be updated as soon as pending trial results become available. We are also concerned that ICER neglects a wealth of evidence from numerous studies, including comparative efficacy studies that utilize surrogate markers such as bone mineral density. Finally, the report does not recognize the growing body of evidence demonstrating that the sequence in which these anabolic and antiresorptive therapies are administered has a profound effect on efficacy. The ASBMR shares ICER’s and the public’s concern with the high cost of anabolic osteoporosis therapies. We enthusiastically support all efforts to make them more affordable to our patients. Osteoporosis patients deserve unimpeded access to screening, evaluation, and the full array of treatment options. Dr. Leder has received research support (medication supply) from Amgen and Eli Lilly on an investigator-initiated trial. ©Institute for Clinical and Economic Review, 2017 Page 111 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Appendix H. Conflict of Interest Disclosures Tables H1 through H3 contain conflict of interest (COI) disclosures for all participants at the June 30, 2017 Public meeting of the California Technology Assessment Forum. Table H1. ICER Staff and Consultant COI Disclosures Name Organization Disclosures Gregory Guzauskas, MSPH, PhD University of Washington None David Rind, MD, MSc ICER None Steven Pearson, MD, MSc ICER None Matt Seidner, BS ICER None Lotte Steuten, MSc, PhD University of Washington None Patricia Synnott, MALD, MS ICER None Jeffrey Tice, MD University of California, San Francisco None Table H2. CTAF Panel Member COI Disclosures Name Organization Disclosures Ralph G. Brindis, MD, MPH, MACC, FSCAI, FAHA UCSF * Christine Castano, MD HealthCare Partners * Robert Collyar Patient Advocates in Research * Felicia Cohn, PhD Kaiser Permanente * Luanda Grazette, MD, MPH, FACC USC * Kimberly Gregory, MD, MPH Cedars-Sinai Medical Center * Paul Heidenreich, MD, MS (Chair) Stanford University * Jeffrey Hoch, PhD UC Davis * Joy Melnikow, MD, MPH UC Davis * Patricia E. Powers, MPA Center for Healthcare Decisions * Rita F. Redberg, MD, MSc, FACC UCSF * William Remak, BsC MT, BPH Emeritus Board Member, California * Chronic Care Coalition Robert E. Rentschler, MD Beaver Medical Group * Alexander Smith, MD, MPH UCSF * Michael Steinberg, MD UCLA * * No relevant conflicts of interest to disclose, defined as more than $10,000 in healthcare company stock or more than $5,000 in honoraria or consultancies during the previous year from health care manufacturers or insurers. ©Institute for Clinical and Economic Review, 2017 Page 112 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents Table H3. Policy Roundtable Participant Disclosures Name Title Disclosures Victoria Dang, PharmD Director, CDAG Program Performance, UHC Employee and stockholder UnitedHealthcare Medicare and Retirement Matthew Drake, MD, PhD Consultant, Division of Endocrinology, None Department of Medicine; Associate Professor of Medicine, Mayo Clinic Deborah Kado, MD, MS Professor, Department of Family Scientific advisory board: Amgen Medicine and Public Health; (romosozumab), Kalytera Osteoporosis Clinic Director, Department of Medicine; Deputy Director of Clinical Research and Education, Sam and Rose Stein Institute for Research on Aging, University of California, San Diego John Krege, MD, FAHA Medical Fellow, Eli Lilly and Co. Eli Lilly employee and stockholder. Shireen Fatemi, MD Healthy Bones Regional Co-Lead, Kaiser None Permanente Southern California; National Clinical Lead for Osteoporosis, Kaiser Permanente, Assistant Area Medical Director, Kaiser Permanente Panorama City Stuart L. Silverman, MD, Clinical Professor of Medicine, Cedars- Advisory Board, Speaker: Amgen, Lilly, FACP, FACR Sinai Medical Center and UCLA School of Radius Medicine; Medical Director, Consultant: Amgen Osteoporosis Medical Center Clinical Research Grants: Amgen, Lilly, Novartis, Research Center; Member, National Pfizer, Roche Bone Health Alliance Osteoporosis Former Officer: Kalytera Messaging Group Roselyne Smith Patient None Martin Zagari, MD Vice President, Global Health Amgen employee, officer, and Economics, Amgen, Inc. stockholder ©Institute for Clinical and Economic Review, 2017 Page 113 Final Evidence Report – Anabolic Therapies for Osteoporosis Return to Table of Contents