Designing Alzheimer’s Disease Clinical Trials For Success
Alzheimer’s disease (AD) is and will remain one of the most unyielding public health crises of the 21st century. Aging global populations translate to a constantly growing prevalence of a disease with no cure and an immense cost of care. In the United States alone, recent estimates1 suggest that 28 million people will develop AD by midcentury. Approximately 25% of the annual Medicare budget will be spent treating the disease – a projected $330B in 2040 – with hundreds of billions more in indirect costs.
Alzheimer’s disease (AD) is and will remain one of the most unyielding public health crises of the 21st century. Aging global populations translate to a constantly growing prevalence of a disease with no cure and an immense cost of care. In the United States alone, recent estimates1 suggest that 28 million people will develop AD by midcentury. Approximately 25% of the annual Medicare budget will be spent treating the disease – a projected $330B in 2040 – with hundreds of billions more in indirect costs.
An effective disease-modifying treatment for AD would dramatically improve this prognosis, and it is no surprise that pharmaceutical companies are willing to invest millions in Alzheimer’s R&D efforts. However, without careful clinical trial design, even the most promising drugs can fail. Indeed this has been the fate of the vast majority of AD therapies over the past decade. Of 83 Phase III AD clinical trials conducted in the past decade, not one met its primary endpoint(s). In fact, since 2000, only one new compound for treating AD successfully made it to market – memantine (Namenda). This added to a lineup of symptomatic treatments that do nothing to affect the underlying pathology of AD - namely, the abnormal buildup of two proteins in the brain, amyloid and tau.
Of the many compounds that failed in Phase III trials, two targeted AD pathology directly -- bapineuzemab (Elan/JNJ & Pfizer) and solanezumab (Lilly). Both are monoclonal antibodies that facilitate clearance of amyloid from the brain, which scientists believe could reduce the impact of the hallmark amyloid plaques of AD on brain function. Both underwent massive Phase III clinical trials between 2008 and 2012, which cost hundreds of millions of dollars. Both fell short of meeting their primary efficacy endpoints. However, recently, there has been a resurgence in the number of AD clinical trials based on the hypothesis that historical disappointing trial failures were a result of clinical trial design shortcomings rather than the efficacy of the compounds themselves.
The next wave of AD clinical trials is taking a more targeted approach. Here we review the core considerations of two critical components of Alzheimer’s trial design – sample enrichment and outcomes selection. In particular, we will discuss the role of diagnostic imaging and innovative cognitive and functional measures in cultivating successful AD treatments.
1. Sample Enrichment: Diagnostic Imaging and The Road to Alzheimer’s Prevention
Sample enrichment in Alzheimer’s trials primarily involves a combination of genetic testing and biomarker screening. Genetic testing is being used to identify early stage AD adults. For instance, carriers of the e4 allele of the apolipoprotein E (APOE) gene have a higher risk of developing AD than non-carriers. There are also several established rare autosomal dominant genetic mutations that virtually guarantee the development of AD at an early age. Genetic testing helps identify this critical population for inclusion in clinical trials.
The most important component of sample enrichment in AD trials is biomarker screening, specifically for amyloid and tau. Until recently, the only way to diagnose AD was through post-mortem neuropathological examination. Researchers had no way of determining whether a dementing person was the victim of amyloid and tau (AD) or some other cause of dementia. Even the best clinicians were only able to say that an individual had “dementia due to probable Alzheimer’s disease.” Since the monoclonal antibodies under development for the treatment of AD primarily target amyloid, there was the risk that treated patients with probable AD did not in fact have amyloid pathology, and therefore received a drug with no target. This is precisely what happened in the original solanezumab trial, where over 25% of the treated “AD” patients2 did not have AD at all.
In the past decade, biomarker screening for Alzheimer’s has advanced rapidly. First, there was the introduction of cerebrospinal fluid (CSF) testing for amyloid and tau, a process that involves invasive spinal taps. Then there was the introduction of easier, albeit costly, positron emission tomography (PET) imaging for amyloid, using one of a number of amyloid ligands. The FDA approved the first of these ligands, Amyvid (florbetapir), in 2012, and Lilly now markets the product in the US despite the fact that most US commercial payers and the Centers for Medicare & Medicaid Services (CMS) and most commercial payers do not currently reimburse for commercial use of the test3. Now a number of tau PET ligands are also in development along with other novel biomarker diagnostics techniques like blood/plasma assays and saliva testing4, which will add to the arsenal of diagnostic tests available for AD in the years to come.
As a result of these advancements, it is now possible for drug makers to ensure that their clinical trial samples include only those patients for whom their respective therapies are designed to work. Trials for amyloid-targeting antibodies will enroll only patients with known amyloid pathology and trials for tau-targeting therapies can select for patients with tau tangles. This is critically important as the appearance of amyloid and tau tangles occur at different points over the course of the disease. Tau tangles form much later in the progression of Alzheimer’s disease, at roughly the same time that cognitive deficits begin to appear. This is in contrast to amyloid, which may begin to aggregate in the brain up to 20 years prior to the onset of symptoms. What has typically been described as the beginning of Alzheimer’s disease (i.e., memory and other cognitive deficits severe enough to impair activities of daily living) is actually the end phase of the disease.
This last fact is one that drug companies are keenly aware of – that Alzheimer’s disease itself begins many years before patients show any sign of cognitive decline. In fact, nearly 20% of the healthy adult population over the age of 50 shows evidence of amyloid buildup when screened using the abovementioned PET imaging tests5. Not all of these individuals will go on to develop clinical dementia, but by having amyloid pathology, the risk is far greater. These people may benefit from amyloid-targeting therapies. In contrast, adults without amyloid pathology do not have AD and will not benefit from any amyloid-targeting drug.
In summary, biomarker screening makes it possible to detect the very earliest signs of AD, which experts now call preclinical AD, and drugs such as aducanumab (Biogen), solanezumab, gantenerumab (Roche), crenezumab (Genentech), and MK-8931 (Merck) hold promise not only for treating patients with the early stages of clinical AD, but also for those with preclinical AD. For these individuals, amyloid-targeting treatments may in fact be preventative, for once Alzheimer’s shows itself in the form of dead neurons and cognitive decay, the damage is likely irreversible. For those with preclinical AD, the cognitive damage is not yet done. If these drug candidates prove successful in slowing the progression of clinical AD, as many as 16 million individuals in the US alone in 2015 could benefit from these amyloid-targeting drugs5. If these amyloid-targeting drugs prove to be efficacious in preclinical AD and early AD, then this raises a new core question: how will we pay for these treatments, and the diagnostic imaging required to ensure only adults that are clinically appropriate receive treatment?
2. Outcomes Selection: Innovation in Cognitive and Functional Measures
Strong outcome measures are equally critical, and similarly elusive, for Alzheimer’s clinical trials. A combination of biomarkers (the same PET or CSF tests used for amyloid and tau screening may be used to measure the direct impact of a treatment on its target pathology), along with measures of cognitive and functional decline, make up the core primary and secondary endpoints of AD clinical trials.
Unfortunately, existing measures often fall woefully short, lacking the sensitivity and specificity that such high value trials demand. For example, the Mini Mental State Examination (MMSE), an 11-question screening test scored on a 30-point scale, where anything less than 27 is considered a sign of cognitive impairment. However, this test has been around for 40 years, and many clinicians admonish its simplicity and poor sensitivity and specificity. Other tests such as the Alzheimer’s Disease Assessment Scale-cognitive (ADAS-cog) and Clinical Dementia Rating scale (CDR) are only slightly better. Most alternatives lack the global recognition or standardization needed for a large global trial. Likewise, there are few good standardized measures of function (i.e., ability to perform activities of daily living such as dressing or making food), a matter complicated by the fact that definitions of functional status may vary widely between individuals as well as between geographical regions. Functional “impairment” in the United States may have a very different meaning than in India, for example, and trials with sites in both locations that use a single standardized functional outcome measure as an endpoint may doom themselves to regional confounding.
These issues become compounded in those preclinical AD populations, where the patients are otherwise healthy and cognitively normal. Patient-reported outcome measures, which are sometimes used, and often mandated, for tests of function in clinical AD populations are suboptimal and may in fact produce results entirely contradictory to reality due to the problem of anosagnosia: individuals with dementia are often unaware of their decline. As a result, if a drug is not successful in treating AD, it may falsely demonstrate an improvement in self-reported outcomes.
Current standard cognitive assessments such as the MMSE, ADAS-cog, and others also completely lack the sensitivity needed for preclinical populations due, in part, to “ceiling effects,” meaning that any enrolled individual, treated or not, will likely get a perfect score upon enrollment, at the end of the trial, and for many years thereafter. Instead, trials for preclinical or preventative indications often utilize composite cognitive assessment scales that incorporate key aspects of existing tests designed through statistical modeling to capture early decline. One example is the ADCS-PACC, which combines items from four existing standardized cognitive tests and is a primary endpoint of the solanezumab “A4” trial6 (currently in the process of enrolling 1,000 individuals with preclinical AD, the first trial to test this important population and tackle these issues head-on).
The holy grail of outcomes measures would be a well-validated, globally standardized cognitive and functional assessment battery, designed with the goal of remaining valid across the entire spectrum of Alzheimer’s disease: sensitive to the earliest signs of cognitive decline in preclinical AD and valid through the clinical stages of the disease. Unfortunately, designing such an assessment tool would take many years, and itself require three to five plus years of clinical study.
For the time being, it appears that the industry is learning from its past mistakes. Drug manufacturers know that without proper sample selection and endpoints, trials for clinical AD and preclinical AD will be doomed for failure before they enroll their first patient. Drug manufacturers understand that with appropriate investment of both time and money, innovation in outcome measures, and careful protocol design, these trials may result in FDA approval of treatments that will reduce the burden of Alzheimer’s disease.
Jillian Scaife, Principal, Trinity Partners and Nick Simmons-Stern, Consultant, Trinity Partners
Suggested Links:
[1] http://alz.org/aaic/releases_2015/Mon930ET.asp
[2] http://fortune.com/2015/04/23/alzheimers-drug-research
[3] https://www.cms.gov/medicare-coverage-database/details/nca-decision-memo.aspx?NCAId=265
[5] Trinity estimate based on US Census data and review of available literature
In Focus: Addressing the Health Literacy Roadblock in Patient Recruitment
Published: November 15th 2024 | Updated: November 15th 2024With universal adoption of health literacy best practices slow going over the years, advocates are redefining the term to encompass much more of what health-related communication requires beyond simply words.
