Anderson TS, Ayanian JZ, Souza J, Landon BE. Representativeness of Participants Eligible to Be Enrolled in Clinical Trials of Aducanumab for Alzheimer Disease Compared With Medicare Beneficiaries With Alzheimer Disease and Mild Cognitive Impairment. JAMA. 2021 Oct 26;326(16):1627-1629. PubMed.

Recommends

Please login to recommend the paper.

Comments

  1. Following approval of Aduhelm by the FDA, there have been fierce discussions in the field. In my view, one important topic is who would be eligible/who might benefit from Aduhelm. This is clearly only a small subset of all patients with AD, namely those with MCI or mild-stage dementia, with only little comorbidity, and who are willing and able to undergo the treatment regimen.

    In this respect, the letter by Anderson and colleagues is very insightful, as they illustrate that only a small percentage of all AD patients would be eligible for treatment. It would be even more informative if the authors had taken the population of MCI or mild dementia due to AD as a starting point (in the paper, all AD were included). A large proportion of AD patients in the Medicare database have moderate-to-severe AD dementia. It is clear that they will not benefit, and with the narrowed labeling, they are no longer eligible to be treated with Aduhelm. I suspect that the percentage of eligible patients would be somewhat higher if MCI or mild dementia was taken as starting point.

    Of note, there could also be a proportion of patients with early stage AD, particularly the MCI stage that is not known at Medicare. Many individuals with MCI due to AD go undiagnosed, and as such we do not know the true prevalence of AD. Availability of a medication like Aduhelm may change the whole diagnostic landscape, with a far larger number of people seeking a diagnosis in an early stage.

    Given the potential side effects, it is indeed relevant that any factors increasing the risk of side effects be limited. I could imagine that clinicians are reluctant to prescribe Aduhelm to patients with cardiovascular disease or anti-coagulation, particularly in the coming years, while we need to get more insight into the clinical efficacy and occurrence of side effects of the treatment.

    View all comments by Wiesje van der Flier

  2. I reviewed the Anderson et al. JAMA article, and I am not surprised. One of our junior faculty at the Cleveland Clinic did his own analysis using the inclusion/exclusion criteria and came to the same conclusion.

    In fact, when inclusion/exclusion criteria are applied, most patients do not qualify for aducanumab, or any monoclonal antibody for that matter. This is never reflected in the data of randomized controlled trials, because they are excluded before enrollment. RCTs include a very selected population that excludes people with pacemakers, stroke, and a variety of health conditions. Further, the aducanumab study did not assess age >85, which could be many potential patients. Thus, the effects of aducanumab in age >85 are largely unknown.

    If the inclusion/exclusion criteria are actually applied, then the number of people eligible and appropriate for aducanumab are far fewer than the pundits have hyped (“Medicare will be bankrupt!”). The takeaway is that the findings from Anderson et al. are not surprising and should be considered when selecting patients for monoclonal antibody treatments in the future.

    View all comments by Marwan Sabbagh

  3. Indeed, ARIA is a concern. Although asymptomatic in most cases, it may have an aggressive course. I was extremely surprised to see the first label that was proposed and welcomed the narrowing of the population, in line with the appropriate-use criteria suggested by Cummings et al.

    But the total lack of contraindications completely baffled me. Cardiovascular risk factors, such as arterial hypertension, increase the risk of ARIA, as does the use of oral anticoagulants. If I were to prescribe Aduhelm, I would aim for a patient who is relatively “clean” with regard to CV disease and MRI scan, to mitigate the risk of ARIA. Managing ARIA is only possible by repeated MRI scanning and should be done as in the trial.

    Let me just add here that with Aduhelm and other therapeutic antibodies to follow later, we have entered a new era in which treating AD patients becomes more of a precision-medicine approach, in the hands of specialists only.

    View all comments by Philip Scheltens

  4. I am not a trialist, but there must be a balance between sufficient representativeness and adequate homogeneity. Practitioners of evidence-based medicine know that clinical trials provide critically important information, but individual clinicians must make the final decision regarding whether the trial outcomes are applicable to each patient considering the treatment.

    I doubt there will ever a clinical trial that is completely representative of everyone eligible for the drug. For example, people who are deaf or blind can develop Alzheimer’s disease, but are not part of even the most inclusive trials. Inclusion for the sake of representativeness, i.e., enrolling a small number of participants with some of the comorbidities outlined in the article, won’t have sufficient power to address whether a drug is effective in these conditions, and risks not addressing if the drug works in an extremely artificial construct.  

    Many who have commented on the artificial nature of these explanatory clinical trials are essentially reinforcing the call for more pragmatic trials in drug development, that is, trials in the “real world,” which NIH has been advocating over the past years (NIH blog, 2017). Alas, until more centers have the infrastructure to implement pragmatic trials, writing off explanatory trials may be throwing out the baby with the bath water.

    The COVID-19 pandemic has also reminded us that motivated physicians and scientists can design and execute elegant trials to repurpose an FDA-approved drug, or test a drug’s efficacy against another therapy when they don’t agree with the approved treatment paradigms. Given the concerns related to anti-amyloid therapy, I imagine many would prefer such studies over idle chatter.

    In the meantime, patients are faced with real decisions of trying an FDA-approved drug with known limitations and known side effects versus entering other clinical trials of unknown risks and unknown outcomes. The decision to try aducanumab—or not—should eventually rest with each patient and his or her physician.

    View all comments by William Hu

  5. The FDA's Accelerated Approval for Aduhelm relies on the assertion that current amyloid fibril burden status, as determined by PET, is a useful surrogate for a meaningful and beneficial clinical response to amyloid-reducing interventions. Writing in The BMJ Open earlier this year, Ackley and colleagues performed an unbiased computational meta-analysis of all available amyloid fibril burden PET data, together with the corresponding contemporaneous cognitive status. They concluded that no important relationship links the PET data to cognitive status. The Ackley report would appear to undermine the FDA's Accelerated Approval language so directly and so effectively that the validity of the accelerated approval must be considered suspect at the very minimum.

    References:

    Ackley SF, Zimmerman SC, Brenowitz WD, Tchetgen Tchetgen EJ, Gold AL, Manly JJ, Mayeda ER, Filshtein TJ, Power MC, Elahi FM, Brickman AM, Glymour MM. Effect of reductions in amyloid levels on cognitive change in randomized trials: instrumental variable meta-analysis. BMJ. 2021 Feb 25;372:n156. PubMed.

    View all comments by Sam Gandy

Make a Comment

To make a comment you must login or register.