Endpoints beyond fracture promise to facilitate shorter trials, new osteoporosis research
Osteoporotic fractures represent a costly human and socioeconomic burden. Even with currently available drugs, treatments that more effectively increase bone mineral density, reduce fractures and reverse the disease are needed.
“Osteoporotic fractures, most notably hip fractures, cause substantial morbidity and mortality. In the osteoporosis scientific community, there is concern regarding the plateau or slight rise in hip fracture trends in the U.S., as well as a decline in the diagnosis and treatment of osteoporosis despite the treatment options available,” Theresa Kehoe, MD, director of the division of general endocrinology at the Center for Drug Evaluation and Research at the FDA, told Endocrine Today. “Therefore, there is wide agreement that patients would benefit from new treatment options.”
Despite this need, phase 3 drug trials for osteoporosis therapies have become almost impossible to conduct. Regulatory agencies require a primary outcome of fracture for any phase 3 osteoporosis drug trial. With effective treatments already available, it is not considered ethical to conduct a placebo-controlled drug trial with participants at high fracture risk. Any placebo-controlled trial must be limited to low-risk participants who are less likely to experience a fracture, substantially increasing trial size, duration and cost.
“Given existing drugs, most institutional review boards will now not uniformly approve the use of a placebo control, so then you’re trying to prove noninferiority or superiority to existing treatments, which, with fractures, becomes virtually impossible,” Sundeep Khosla, MD, professor of medicine and physiology at Mayo Clinic, told Endocrine Today. “Continuing to use fracture as an endpoint is essentially going to prohibit any new drugs from coming to market.”
Investigators have now proposed a validated alternative outcome for fracture — change in bone mineral density — to serve as a surrogate endpoint to push forward promising osteoporosis agents that have stalled in the development pipeline. Now the FDA is listening. With funding from the Foundation for the NIH (FNIH) Biomarkers Consortium, Dennis M. Black, PhD, professor of epidemiology and biostatistics at the University of California, San Francisco, and colleagues with the Bone Quality Project are preparing to submit individual, patient-level data from nearly every large osteoporosis trial to the FDA as part of a Qualification Plan. The data from more than 150,000 patients demonstrate that change in BMD is a reliable surrogate for fracture risk reduction, according to Black.
“Instead of a trial with 10,000 people over 5 years, you can conduct a trial with 2,000 people over 2 to 3 years with BMD as a substitute for a hard clinical endpoint,” Black, also an adjunct investigator at Kaiser Permanente Southern California, told Endocrine Today. “We believe the primary mechanism of action of all these bone drugs is through strengthening bone, and that is reflected in bone density.”
Lessons from sodium fluoride
Traditionally, the FDA has required fracture as an endpoint for any new osteoporosis drug trial. The requirement stems from studies conducted in the 1980s assessing the effect of fluoride therapy on fracture rate. Data showed a marked increase in bone density among treated women, but no reduction in fracture risk.
For one such study published in The New England Journal of Medicine, B. Lawrence Riggs, MD, professor of medicine at Mayo Clinic, and colleagues analyzed data from 202 postmenopausal women with osteoporosis and vertebral fractures assigned 75 mg daily sodium fluoride or placebo. Women assigned fluoride experienced a 35% increase in BMD at the lumbar spine (P < .0001), a 12% increase at the femoral neck (P < .0001) and a 10% increase at the femoral trochanter (P < .0001); BMD decreased at the radius by 4% (P < .02). Despite large BMD gains, the number of new fractures was similar between groups; nonvertebral fractures were higher among treated women vs. placebo (72 vs. 24; P < .01).
“Until then, bone density was considered a good surrogate measure,” Khosla said. “The problem was, fluoride led to a marked increase in bone density, but didn’t reduce fracture risk. It had to do with the incorporation of fluoride into bone actually making bone more brittle. The point that is lost though is, given all the preclinical testing that is done now in animal models to make sure a change in bone mass in animal models translates into a change in bone strength, fluoride would never come to clinical trial today. There is now a completely different evaluation process.”
The utility of fracture as an endpoint for osteoporosis trials has been debated repeatedly in the decades since. Nearly 20 years ago, the NIH and American Society for Bone and Mineral Research (ASBMR) convened a meeting to address concerns regarding the appropriateness of placebo-controlled fracture trials in osteoporosis drug development and the use of nonfracture surrogate endpoints. In the end, experts determined that a demonstration of antifracture efficacy would remain “essential” for acceptance of new therapies.
The requirement has been frustrating for trialists and has served to limit new drug research, Khosla said.
“You don’t need to show cardiovascular risk reduction to approve a drug for blood pressure reduction — BP is used as a surrogate,” Khosla said. “For diabetes drugs, they target HbA1c, not diabetic complications. For virtually every chronic disease that we deal with, the FDA has reasonable surrogate markers. Osteoporosis is the outlier, and it is because of that fluoride story.”
Building a case
In 2016, the UCSF group working with Black in conjunction with the FNIH Biomarkers Consortium released the first study results from the Bone Quality Project, a large-scale effort to reanalyze pooled imaging and biochemical data from virtually all large osteoporosis drug trials to allow definition of better clinical endpoints. Unlike a meta-analysis of already-published studies, the project includes raw, patient-level data from more than 175,000 participants from 50 randomized controlled trials. The unusual effort is in part the result of pharmaceutical companies agreeing to share what is typically tightly held information.
“This database is a jewel in the crown,” Gayle Lester, PhD, program director of the National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS) and a former chair of the Bone Quality Project, told Endocrine Today. “It’s really something. It started very small — maybe 10,000 subjects. Now you could use this database for epidemiologic investigations of sensitivities to fracture or risk for fracture or loss of BMD in these populations. I don’t know of any other investigator in the world who could have earned the trust of all of these companies other than Dr. Black. Part of it was he had good relationships with all these companies over the years. He won their trust.”
Using this database, Black and colleagues assessed the association between treatment-related changes in BMD, as measured by DXA, and fracture outcomes, including the proportion of treatment effect explained by BMD changes, across 23 randomized placebo-controlled trials with 91,779 participants. Trials had 1 to 9 years of follow-up and assessed bisphosphonates (n = 12), odanacatib (n = 1), hormone therapy (n = 2), parathyroid hormone receptor agonists (n = 3), denosumab (Prolia, Amgen; n = 1) and selective estrogen receptor modulators (n = 4).
In meta-regression analysis published in August in The Lancet Diabetes & Endocrinology, the researchers demonstrated significant associations between the change in BMD at 24 months and reduction in vertebral, hip and nonvertebral fractures for all sites of BMD measurement, such that greater gains in BMD were associated with larger fracture reductions.
“For example, the study with the smallest difference in total hip BMD (1.3%) was associated with a 23% reduction in vertebral fractures (OR = 0.77; 95% CI, 0.71-0.83), whereas vertebral fractures were reduced by 66% (OR = 0.34; 95% CI, 0.31-0.37) in the study with the largest total hip BMD difference,” the researchers wrote. “For hip fractures, the estimated HRs ranged from no significant fracture reduction for the smallest BMD difference (HR = 1.01; 95% CI, 0.9-1.15) to a 48% reduction in hip fracture (HR = 0.52; 95% CI, 0.46-0.59) for the largest BMD difference.”
Black said the analyses support BMD as a surrogate outcome for fracture outcomes in future randomized trials of new osteoporosis therapies and provide an important demonstration of the value of public access to individual patient data from multiple trials.
“Because we have individual patient data, we can standardize things across studies,” Black said. “We can define fracture, for example, the same way across all studies. If FDA wants us to look at, for example, older people vs. younger people, we can do that. It’s really cool, and that is why FDA is excited, too. In this era of more public availability of data from clinical trials, these kinds of things will become more possible in the future.”
‘It looks like a shortcut’
Experts acknowledge that a barrier to the adoption of BMD as the criteria for registration of new osteoporosis drugs is the need to demonstrate safety. In April 2019, the humanized monoclonal antibody romosozumab-aqqg (Evenity, Amgen) was approved with a boxed warning alerting users that it may increase risks for myocardial infarction, stroke and CV death and that patients should not use the drug if they have had an MI or stroke within the previous year. In 2016, Merck announced it discontinued the development program for the oral cathepsin K inhibitor odanacatib, after an independent adjudication and analysis of major adverse CV events confirmed an increased risk for stroke during the 5-year trial period.
“We were worried about this [proposed change] because when you stick your neck out and say bone density will give us the answer, and you can do a trial in 500 instead of 10,000 people, it looks like a shortcut,” Richard Eastell, MD, FRCP, FRCPath, FMedSci, professor of bone metabolism, head of the academic unit of bone metabolism and director of the Mellanby Centre for Musculoskeletal Research at the University of Sheffield Medical School, United Kingdom, told Endocrine Today. “We’re all so used to trials taking many years with many thousands of patients. And those companies that have taken their drugs through these long, expensive trials might consider this new approach unfair. I can understand these issues.”
Other rare adverse events have been a concern with some osteoporosis treatments; cases of osteonecrosis of the jaw or atypical femur fracture with bisphosphonate use, while uncommon, have put off some patients from initiating a therapy where benefits still largely outweigh any risk.
“One of the great things about having a 5-year trial of 16,000 people is if there are very rare side effects, you pick them up,” Eastell said. “People have now become used to the idea that we conduct these massive trials for many years in osteoporosis, but in no other therapeutic area do you conduct such large trials for such a long time. In the large trial for odanacatib, you picked up a stroke signal. So, people say, ‘Let’s keep doing them,’ but their high costs prohibit this.”
A solution, Lester said, could be that agents demonstrating a reduction in BMD are recommended for approval with long-term, post-marketing follow-up.
“So, a much smaller trial with fewer patients and getting your result and saying, yes, this has a significant impact on BMD,” Lester said. “The inference is that means it is going to reduce fracture risk. The proof in the pudding is long-term follow-up on those subjects.”
Establishing efficacy for hip fractures, in particular, can be difficult, Khosla said.
“Maybe [the FDA] won’t go completely to bone density,” Khosla said. “Maybe they will say, bone density, plus a reduction in vertebral fracture risk. This would still reduce the numbers of subjects needed [for a trial] by a lot. Hopefully, they will move to bone density alone, with vertebral fracture as a secondary endpoint.”
Under the 21st Century Cures Act, biomarker qualification through the FDA involves a three-stage submission process to develop a biomarker for regulatory use. The first stage is a letter of intent to initiate the qualification process of a biomarker for a proposed context of use in drug development. The second stage requires a Qualification Plan, or QP, to define the intended development to generate the necessary supportive data. For the final stage, researchers submit a full qualification package, or FQP, with all accumulated data to support the qualification of the biomarker. The FDA will then determine whether the biomarker is qualified for the proposed context of use.
“The goal is for these biomarkers to be qualified by FDA and then used in future clinical studies for drug development and clinical practice in osteoporosis,” Kehoe said.
Researchers with the Bone Quality Project have so far submitted a letter of intent to the FDA, which was approved, as well as an initial QP, Black said.
Lester called the pace of recent developments gratifying after years of work.
“Maybe within a year from now, there could be preliminary approval from FDA, or actually allowing trialists to go forward [using the surrogate endpoint],” Lester said. “There are companies in the research arena who could be pulled out and put into a larger clinical trial for preliminary approval.”
A new surrogate endpoint approved during the next 2 years could be a game changer for osteoporosis research, Eastell said.
“For trialists, for the companies developing drugs, this suddenly means clinical trials will become affordable,” Eastell said. “Because bone turnover markers are my particular area of interest, I am very much aware of the number of phase 1 and phase 2 trials going on at the moment — and these are for bone drugs that could really be helpful. Most of these are small companies, so they need a big partner. They may not find anyone because there are already several drugs licensed for osteoporosis. The difference this will make is these companies will say, ‘We can do a 600-person trial and that will be sufficient.’ It now becomes feasible to do it.”
The FDA, while keeping patient safety in mind, is also looking to be more innovative in how it assesses drug approvals, Khosla said. The unprecedented speed of vaccine development during the COVID-19 pandemic has stimulated that mindset, he added.
“Some of that culture change related to COVID-19 may well spill over into other diseases, and I hope it does,” Khosla said. “It has become clear that while that fluoride example was real, there are now so many safeguards built into drug development that that will never happen again. FDA is coming around to that view. There is such an enormous amount of data relating changes in bone density to fracture risk. This FNIH effort, led by Dr. Black, is a huge catalyst for this. It is a true big data effort. It’s compelling.”