After months of uncertainty, the hammer fell for Biogen: In a draft decision released January 11, the Centers for Medicare and Medicaid Services proposed to cover the anti-amyloid antibody Aduhelm only in the context of clinical trials approved by the CMS itself or by the National Institutes of Health. Many observers had predicted the CMS might choose this route, called coverage with evidence development (CED), but the draft requirements are more stringent than expected. They include only randomized controlled trials conducted in hospital-based outpatient settings, and require patient diversity reflecting that of the U.S. population diagnosed with Alzheimer’s. Another surprise was that this decision applies to the whole class of anti-amyloid monoclonal antibodies, giving it far-reaching implications.

  • In a draft decision, the CMS proposed covering Aduhelm only in clinical trials.
  • Most Alzheimer’s researchers support the proposal.
  • Some challenged its practicality.

The announcement triggered pushback from Biogen, its partner Eisai, and from advocacy groups. The Alzheimer’s Association and UsAgainstAlzheimer’s issued statements criticizing the decision, as did the Global Alzheimer’s Platform. These groups announced campaigns to flood the CMS during the 30-day comment period with demands to change its proposal (Politico article). 

The manufacturers of the other anti-amyloid antibodies in advanced clinical trials, Eli Lilly and Roche, have yet to weigh in publicly. The Centers’ final decision is expected by April 11.

Alzheimer’s researchers largely agreed with the current proposal. Comments from more than 40 scientists, most expressed publicly and some off-record, ran 2:1 in favor. Most who supported it said the CMS struck a reasonable compromise between access to a new drug and patient safety, especially in light of the scant efficacy data for aducanumab. “The CMS decision could be seen as a corrective on the FDA’s original approval,” said Ian Grant at Northwestern University, Chicago, expressing a common view. “It sends a message to companies that trials need to show clear benefit before drugs will be covered.”

Other scientists questioned whether the CMS proposal is workable. “No one besides Biogen is going to have the incentive or money to run another sufficiently large, randomized, placebo-controlled Phase 3 study of aducanumab, and Biogen is already doing this,” said Erik Musiek at Washington University in St. Louis (full comment below).

Some noted the diversity requirement sets a high bar that no other AD drug development program has met, and predicted the decision would be counterproductive. “The proposed National Coverage Determination (NCD) … is essentially a non-coverage decision that restricts access and stifles innovation,” said Stephen Salloway at Butler Hospital in Providence, Rhode Island (full comment below).

The U.S. Congress has weighed in. The chairs of two House committees—Energy and Commerce, and Oversight and Reform—on January 13 sent a nine-page letter to the CMS administrator, Chiquita Brooks-LaSure, commending the draft decision.

Reasonable and Necessary? The Center Has Its Doubts
Before it pays for a new drug, the CMS must determine that it is “reasonable and necessary” for treatment of the given illness. For aducanumab, the CMS found insufficient evidence for this determination, but acknowledged the high unmet need for Alzheimer’s treatments and urgency for patients. “Because of the early but promising evidence and the immense burden of this devastating disease on the Medicare population, we are proposing CED to support rigorous trials to answer whether anti-amyloid monoclonal antibodies improve health outcomes for patients,” the agency wrote.

Many Alzheimer’s researchers think this was the best solution in a difficult situation. Jason Karlawish at the University of Pennsylvania, Philadelphia, noted that the FDA’s use of accelerated approvals creates a flaw in the regulatory process, potentially compelling Medicare to pay for ineffective treatments. “The magnitude of Aduhelm’s costs and harms and uncertainties turned this flaw into a chasm … CED is a best effort to repair it,” he wrote to Alzforum. Eric Ross at Massachusetts General Hospital, Boston, concurred. “I think they’ve made the right decision from a cost-effectiveness standpoint,” he wrote (full comments below).

In addition, aducanumab treatment frequently causes the brain bleeds and edema known as ARIA (Oct 2021 news; Dec 2021 news). Many clinicians fear this would be less well managed in clinical practice than in trials. “I applaud the CMS in restricting the use of the drug. This class of drugs needs to be rolled out in a controlled environment, first and foremost for patient safety,” Ron Petersen at the Mayo Clinic in Rochester, Minnesota, told Alzforum.

Daniel Gibbs, a retired neurologist at Oregon Health and Science University, Portland, participated in the ENGAGE Phase 3 trial of aducanumab and survived severe ARIA that required hospitalization. “Those potentially life-threatening side effects should not be taken lightly. The CMS decision offers a reasonable path forward,” he wrote (comment below).

But How Will This Work?
Nonetheless, many AD researchers wondered how feasible these CED trials would be, given the logistical challenges of assembling a large nationwide cohort that would use a standardized protocol. “It is very difficult to understand which groups of clinicians are experienced, available, and ready to organize and write these multi-institution protocols,” Dennis Selkoe at Brigham and Women’s Hospital, Boston, wrote to Alzforum (comment below).

Others questioned the ethics of asking some patients to take a placebo control when there is an FDA-approved drug. The CMS does not explicitly call for a placebo control in its draft decision, instead asking for “an appropriate control representing the standard of care.” This could mean comparing patients on aducanumab to those who choose not to take the drug, for example.

“It is not possible to include a placebo control in CED since patients are liable for co-pays,” noted Gil Rabinovici at the University of California, San Francisco, who runs the IDEAS CED study for amyloid PET imaging (Aug 2020 news). Rabinovici believes the aducanumab CED trial could follow a similar large national model. Others, including Salloway and Selkoe, suggested that a prospective registry design would work better than randomized trials.

Another hurdle is that the draft decision specifies the CED trial should demonstrate a clinically meaningful benefit in cognition and function, defining this as a 1-2 point benefit on the CDR-SB, or 3-5 points on the Functional Activities Questionnaire (Andrews et al., 2019). Pierre Tariot at the Banner Alzheimer Institute in Phoenix noted that a benefit of this size may be unattainable when treating people with minimal cognitive deficits. “Is the CMS indicating that a prevention trial … would not be eligible for approval?” he asked (full comment below).

Researchers agreed that these and other issues, including how much evidence will be required for full approval, need to be clarified in the CMS’ final decision. Mark McClellan at the Duke-Margolis Center for Health Policy in Durham, North Carolina, helped develop the CED process when he ran the CMS from 2004-2006. He told Alzforum that the agency has previously revised draft coverage decisions based on feedback. “We’re not done with this NCD process,” he said.

Others see potential for innovative CED trial designs. Statistician Kert Viele at Berry Consultants in Austin, Texas, suggested enrolling 10,000 or more patients in a huge platform trial, with streamlined assessments and minimal requirements that would allow patients in remote areas to participate. Because of the size, only a small fraction would need to receive placebo (InvesBrain news story). Maria Glymour at the University of California, San Francisco, also thinks pragmatic trial designs may hold the answer. “I hope we as an Alzheimer's research community can band together to ensure rapid fielding of large, inclusive randomized trials,” she wrote (comment below).

What Are the Implications for Patients?
If finalized, the decision will further restrict access to aducanumab, already scarce due to its high price and clinicians’ doubt about its efficacy. While some regional Medicare offices have been covering the treatment, the NCD would halt reimbursements outside of clinical trials. Some private insurers have been covering as well, but most typically follow the CMS’ lead. For their part, state Medicaid programs would still be on the hook, as they are obligated to cover any FDA-approved drug. However, Medicaid administrators can, and often do, impose usage restrictions to try to limit costs, McClellan noted.

Some patients who have been trying to get the drug will have to wait longer. Marwan Sabbagh at the Barrow Neurological Institute in Phoenix told Alzforum that he has several patients who have been on hold, waiting for a green light from the institute’s Pharmacy & Therapeutics committee to start the drug. Now, because of the CMS ruling, the committee has delayed its own decision. “This has real practice implications,” Sabbagh said.

High Bar for Diversity
Regarding the CMS requirement that CED trial participants reflect the diversity of the U.S. population of AD patients, researchers agreed on its importance, but said that, in practice, meeting it will likely be impossible in the near future. “The CMS’s diversity requirement is as strict as it can be, in that they ask for full sociodemographic representation. This is flatly unattainable, because our national health systems aren’t structured to give equitable care to begin with,” Lon Schneider at the University of Southern California, Los Angeles, wrote to Alzforum (comment below).

Jeffrey Cummings at the University of Nevada, Las Vegas, agreed. “We need to do better, but we cannot achieve this suddenly; it is a long-term trust-building enterprise,” he wrote (comment below). McClellan said that the CMS itself could help foster diversity by providing incentives such as bonus payments for Medicare providers to enroll patients in trials, as the agency currently does for COVID research.

Alzheimer's patients in the U.S. are diverse in many ways. William Hu at Rutgers Biomedical and Health Sciences in Newark, New Jersey, noted that older people with poor vision and hearing, and those who do not speak English, also tend to be excluded from clinical trials. People who live in rural areas far from medical centers are at a disadvantage as well. In addition, because AD trials are burdensome for participants, requiring many procedures and regular travel, they are biased toward more educated and well-off patients, noted Todd Golde at the University of Florida, Gainesville (comments below).

Sudha Seshadri, University of Texas Health Science Center, San Antonio, stressed the importance of including people with health issues besides AD in the CED trial population. “We need to expand to study persons with mixed etiology dementias, particularly co-existing vascular cognitive impairment and dementia; studying a small, healthy, easy-to-study subset and then pushing the drug at marketing to a larger group is neither scientifically sound nor wise, ethical policy,” she wrote (comment below).

Because of the diversity requirement, all ongoing trials of anti-amyloid antibodies, including Biogen’s planned confirmatory trial for aducanumab that is due to start this spring, likely will not qualify for the CMS coverage, McClellan noted. Notably, that trial has already been approved by the FDA as meeting its post-marketing RCT requirements. “If this decision stands, Biogen may have to go back to the drawing board,” he told Alzforum. “In the future, it will be helpful for the CMS and FDA to share information and work together to nail down these issues beforehand, so we can design trials efficiently and make sure we’re getting the most out of patients’ willingness to participate in research.”

The Biden administration’s nominee to lead FDA, Robert Califf, also thinks coordination between the agencies could be improved. In remarks at a recent regulatory science summit, he likened the agencies’ relationship to a relay race in which FDA needs to more smoothly pass off the baton to the CMS (Endpoints story).

Caution Extends Worldwide
The picture looks broadly similar overseas, where European and Japanese regulatory authorities have both pushed the pause button on the aducanumab licensing applications before them. Bruno Imbimbo at Chiesi Pharmaceuticals in Parma, Italy, called the CMS decision “scientifically logical and clinically prudent” for requiring clear evidence of efficacy and acceptable safety. “I believe EMA will follow this same approach with other apparently more promising monoclonal antibodies,” Imbimbo wrote (comment below).

In Japan, the regulatory agency PMDA chose to suspend the decision on aducanumab until more efficacy and safety data are available, noted Takeshi Iwatsubo at the University of Tokyo. “I believe the approval review process by PMDA, which should strictly be based on scientific evaluation, will not be affected by the CMS decision in the U.S.,” he added (comment below).

Judged As A Class
Perhaps the most controversial aspect of the draft decision was how broad a brush it applied. Many AD researchers argued that each anti-amyloid antibody should be assessed on its own merits after completing Phase 3 trials. Both gantenerumab and lecanemab are expected to have Phase 3 data this year, with donanemab not far behind.

“Making a blanket coverage decision for this class of anti-amyloid antibodies prior to them having undergone and completed FDA review does not make sense,” wrote David Holtzman at Washington University School of Medicine in St. Louis. Eric Siemers of Siemers Integration LLC agreed. “The notion that all monoclonal antibodies used for the treatment of AD should be considered as a class is uninformed,” Siemers wrote (comments below).

Those familiar with the regulatory process expect the CMS to tweak its decision in the near future, depending on the data from upcoming trials. “Time, new science, and new trial results will change the therapeutic landscape,” Schneider said. Philip Scheltens at VU University, Amsterdam, wondered if a CED trial will even get off the ground. “Given the prospect of seeing three other drugs in the same class reading out in one to two years’ time, it is not very likely that such a trial with Aduhelm will be initiated,” he wrote (comment below).

Steven Pearson, who leads the Institute for Clinical and Economic Review (ICER) in Boston, believes the CMS’ stance may discourage drug companies from applying for accelerated approval in the future. “The big question is whether some of the developers will now view it as preferable to accelerate their clinical studies and seek FDA approval using the regular pathway, when we can all actually know whether the drug works,” he wrote (comment below).

Rachelle Doody of Roche said that is her company’s priority. “We are highly focused on completing our ongoing Phase 3 trials this year to really test the question of whether lowering fibrillar amyloid leads to clinical benefit. We hope to have clear answers soon,” she told Alzforum (comment below).

Bart De Strooper at the Dementia Research Institute, London, worried that the CMS decision could have a chilling effect on other companies. “I hope that, as a field, we do not lose momentum again,” he wrote (comment below). John Hardy at University College London struck a similar note. “Uncertainty is part of science, but having so much bureaucratic uncertainty in addition cannot be helpful,” he told Alzforum.—Madolyn Bowman Rogers


  1. We discussed this at length as a group, so I can summarize our collective thoughts on this announcement.

    It's worth noting that we were surprised by the original decision by the FDA to approve aducanumab in the first place. We had had some hope that there was going to be some disclosure of new evidence supporting the medication's use that went beyond what had been presented previously from their two incomplete Phase 3 trials. We had known for quite a while that this drug (along with multiple other anti-amyloid antibodies) was effective at removing amyloid from the brain, but there had never been clear evidence that this translated to a clinical benefit to the patient. So the decision for accelerated approval based on the assumption that removing amyloid should have a clinical benefit was not backed by clear data and, if anything, prior evidence seemed to show there wasn't a strong reason to expect clinical improvement from the removal of amyloid alone.

    The approval likely hurt the science of studying this class of drugs, i.e., monoclonal anti-amyloid antibodies, since it provided incentive for other companies to apply for accelerated approval based on the same rationale, without having to first show clear benefit in a large, Phase 3, randomized controlled trial. 

    The CMS decision could be seen as a corrective on the FDA's original approval decision, since it applies to all drugs in the same class. It's likely sending a message to companies that trials need to be completed that actually show clear benefit before the drugs will be covered for all patients.

    The decision further creates this incentive by saying that the CMS will cover the cost of these drugs in the trials, which would certainly reduce the overall cost of running such trials for the companies. For this reason, I think the CMS decision is a fair compromise to encourage better scientific evidence. It would only be an unfair restriction if there was clearly reason to suspect that the medication had a significant clinical benefit or the potential to be a cure. No one is arguing that the drug falls into the latter category and, as I've said, the former remains unproven. 

    Though aducanumab is clearly not a cure, some may argue that even the potential for a slowing of the progression of disease symptoms is worth a high cost. Especially when considering an entitlement program like Medicare, I think this is short-sighted. While a truly significant reduction in progression of disease may, in fact, be worth a high cost, the fact remains that AD is a progressive disease that eventually necessitates significant care that places incredible burdens (financial, mental, emotional, social) on family, caregivers, and society as a whole. In the absence of a cure or a truly significant slowing of disease progression, resources would be much better spent supporting the actual on-the-ground care that our patients will require.

    As for diversity, this is something that should be a requirement for all clinical trials. For true effectiveness of an intervention to be known, the trial population needs to be representative of the population as a whole. For far too long, the vast majority of trials have recruited primarily white people of higher socioeconomic status. This is not a failing of non-white communities for being less engaged, but a failing of the research community to take appropriate steps to meet these communities where they are at and understand their needs and the often systemic reasons for lack of engagement. Better efforts have been made in recent years to improve representation of these groups in clinical research, but there's still a long way to go.

    As for the question about access, my understanding is that this will likely be more limited than where things currently stand for amyloid PET. For these scans, I do not believe there is a requirement that they only be covered as part of an RCT. It's not an entirely fair comparison, though, as the PET scans are not unproven therapeutic interventions but rather clearly validated diagnostic tools.

  2. To me, the logic behind this proposed NCD is confusing. No one besides Biogen is likely to have the incentive or money to run another sufficiently large, randomized, placebo-controlled Phase 3 study of aducanumab. And Biogen is already doing this. Now, it may be that the CMS will end up paying for Biogen’s study. Thus, the proposed NCD would not increase patient access to the drug, and could shift the cost of this study from Biogen to the American public. It also creates the possibility that many small trials could pop up, all funded by CMS, with different endpoints and methodologies, some more rigorous than others, and likely coming to differing conclusions. This will just muddy the waters. If the CMS simply declined to pay for aducanumab, it would force Biogen to do the Phase 3 trial themselves (and incentivize getting it done quickly), which seems like a better alternative.

    I expected that the CMS would either decline coverage (thus forcing Biogen to complete another trial, as above), or would approve a non-controlled CDE trial which would require appropriate entry criteria (CDR 0.5, evidence of amyloid, etc.), as well as some consistent, predefined longitudinal data collection on safety and clinical function (semiannual CDR, basic cognitive testing, etc.). The problem with the latter option is that there is no placebo group, but it would still provide data on real-world ARIA risks, and clinical data could be compared to historical data (from the numerous longitudinal ADRC datasets, ADNI, etc.). If there was a striking effect on cognition as we got out past 18 months, it would be detectable.

    This would be more like the IDEAS trial for amyloid imaging, where physicians can order the scan and report some outcome data, without a control group. This approach would also limit the use of the drug to centers with capacity for detailed follow-up, and provide some useful data, while still making the drug available while confirmatory clinical trials progress. The NCD proposed by the CMS is not really an approval, but rather a denial of coverage in disguise. However, it creates more future problems than a straightforward denial.

    From a patient-care perspective, this NCD would force memory centers to create their own trial protocols, which is extremely arduous and requires staffing that most do not have. But why would they, when Biogen, Lilly, and Eisai are all recruiting for their studies? The truth is that with this arrangement, wealthy patients will still be able to pay for the drug out of pocket, while it will be inaccessible to nearly everyone else. These trials will take years, so I expect that situation to persist. It makes things very difficult for doctors and patients.

  3. The intent of the FDA accelerated approval for aducanumab, and potentially for other anti-amyloid antibodies, is to provide eligible patients access to medication that may benefit them while mandating confirmatory research and to stimulate investment and innovation in AD drug development. The proposed CMS NCD for anti-amyloid monoclonal antibodies is essentially a non-coverage decision that restricts access and stifles innovation.

    CMS has provided coverage for more than 200 cancer drugs that have received accelerated approval without an NCD, and the current proposal reveals a distinct bias against AD and related disorders. The FDA and CMS appear to be at odds regarding AD diagnosis and treatment, and patients are caught in the cross-hairs.

    Four tracers for amyloid and tau PET have received FDA approval for the diagnosis of AD, at great sacrifice from terminally ill patients, but they are not used in clinical practice because the CMS has denied coverage. If this proposal is enacted, the same is likely to be true for currently approved and future anti-amyloid monoclonal antibodies. I suspect the main driver for the CMS, our national provider for the elderly and disabled, is to save money while withholding care from patients who desperately need it.

    The proposal emphasizes treatment of underrepresented populations, which is essential, but this is just window-dressing and unrealistic as currently proposed. A new placebo-controlled trial, which would burden Medicare beneficiaries with exorbitant co-pays, is not needed as Biogen is preparing to launch and fund such a trial as a condition of the accelerated approval.

    The CMS should consider an NCD model that provides broad access and collects data on safety and efficacy of these drugs for eligible patients through a CED-based on an open-label, prospective, longitudinal registry design.

  4. The CMS did the right thing.

    To be sure, details need to be worked out, such as how to cover the costs of a placebo and any co-pay for it; and can tau scan results be part of eligibility criteria? Also, the CMS ought to clarify how it will revisit this decision when new data about the class of drugs are available. For example, a company publishes a study that shows clinical benefit—hallelujah! Surely, the CMS will review that evidence.

    Why was this the right decision? The FDA approval of aducanumab has forced America to confront a vexing dilemma. Do we put into practice a drug whose benefit to patients isn't established, is available to millions, is costly, and has documented risks?

    This dilemma is the result of a flaw in our drug development and health care systems. We expect FDA to approve medicines as safe and effective. We expect the CMS to pay for these medicines. Payment follows approval as summer follows spring. 

    Accelerated approval, however, creates a serious flaw in this system. The FDA can send on to Medicare an ineffective medication, and Medicare has to pay for it. It’s a flaw we’ve tolerated. Because, for example, the relationship between drug-biomarker-clinical endpoint is widely accepted and supported by compelling data. 

    The magnitude of Aduhelm’s costs and harms and uncertainties are stunning. Its lackluster reception at major medical centers, insurers, European and Japanese drug regulators all show the data are not compelling. This isn’t the drug to take a financial gamble on.

    This crack in our system has expanded into a chasm. It broke the order apart. The CMS's Coverage with Evidence development is a best effort to repair it.

  5. The CMS does not explicitly consider cost data when they make a coverage determination like this. But I think they've made the right decision from a cost-effectiveness standpoint. We now have multiple studies which have found aducanumab is priced far higher than its clinical benefits warrant.

    The anti-amyloid antibodies coming down the pipeline have the potential to be more effective than aducanumab, and depending on how they're priced, it's quite possible they could prove to be cost-effective.

    We'll have a much better sense of this as more Phase 3 results become available in the next few years. If one of those trials reveals the kind of robust efficacy that everyone's hoping for, I can imagine this NCD might need to be revised quite quickly!

  6. In my opinion, the FDA should not have approved aducanumab on the basis of the results of the EMERGE and ENGAGE trials. Both studies showed significant lowering of brain Aβ, but they had divergent results on cognition. It is certainly possible that aducanumab and other anti-amyloid monoclonal antibodies will be proven effective, particularly in certain populations such as APOE-4 carriers. We just don’t have enough data yet to make that determination.

    What about safety? Although ARIA side effects occurred in 41.3 percent of subjects (Salloway et al., 2022), they were usually asymptomatic or mild, with less than 2 percent considered severe. However, severe side effects can be life-threatening. I personally required two days of ICU care for severe ARIA after just four doses of aducanumab in the ENGAGE open-label extension (VandeVrede et al., 2020).  

    If aducanumab were an effective drug, a 2 percent severe complication rate might be acceptable. But in the absence of evidence of efficacy, those potential life-threatening side effects should not be taken lightly. The CMS decision offers a reasonable path forward, restricting payment for aducanumab to subjects participating in approved clinical trials. We need unequivocal evidence of efficacy.


    . Amyloid-Related Imaging Abnormalities in 2 Phase 3 Studies Evaluating Aducanumab in Patients With Early Alzheimer Disease. JAMA Neurol. 2022 Jan 1;79(1):13-21. PubMed.

    . Symptomatic amyloid-related imaging abnormalities in an APOE ε4/ε4 patient treated with aducanumab. Alzheimers Dement (Amst). 2020;12(1):e12101. Epub 2020 Oct 9 PubMed. Correction.

  7. To my reading, the CMS decision not to provide Medicare coverage for current and future FDA-approved amyloid antibodies outside of new randomized clinical trials (RCT) that they approve is ill-conceived, impractical for helping AD patients, based on a flawed summary of basic and clinical evidence on amyloid lowering, and inconsistent with the equitable patient access they rightly seek. Their “new-RCT” CED mechanism involves a complex and difficult pathway that would lead to Medicare drug coverage for only a small portion of the county’s estimated 2 million aMCI/mild AD patients eligible for aducanumab based on the FDA-approved label. There are many practical questions not yet addressed by what the CMS released on January 11, among them the following:

    How would the new RCTs be designed and executed? Since randomization is required, what will be the agents: aducanumab versus placebo? If so, then the CMS is proposing to redo the two Phase 3 RCTs already done in ~3,300 such AD patients but now not sponsored by Biogen but by “hospital outpatient” clinicians, potentially with NIH support. Such an undertaking is enormously costly, labor-intensive, and complex. And an academic RCT of aducanumab versus another (unapproved) Aβ antibody is impractical.

    Who should design and sponsor these CMS-adjudicated RCTs? They write that they require multiple hospital-based trial centers to collaborate. It is very difficult to understand which groups of clinicians are experienced, available, and ready to organize and write these multi-institution protocols, submit them to the CMS, react to the CMS critiques, list them at, and then execute and analyze them.

    How does the CMS envision achieving “the diversity of patients included in each trial … representative of the national population diagnosed with AD”? This has been a huge challenge in AD trials to date. Moreover, the CMS’s insistence that all trials must be “conducted in hospital-based outpatient settings” will make it harder, not easier, to attract AD patients from underserved minority communities that are understandably skeptical of care at large academic medical centers.

    Why is the CMS encouraging the use of taxpayer dollars (NIH) to sponsor new RCTs for a therapeutic class that is the most advanced in industry-sponsored development? Shouldn’t the NIH fund trials of novel approaches (whether to amyloid, tau, or another) rather than diverting precious dollars to additional RCTs on amyloid antibodies? The CMS’s strategy leads to more spending on one approach and less diversity.

    The CMS requires that health outcomes be “clinically meaningful” but provides no guidance on how this elusive attribute is to be defined and measured. Instruments quantifying clinical meaningfulness from the perspective of patients and their families are not yet widely validated.

    What is the legal and practical relationship of the CMS and the FDA regarding the regulation and adjudication of these RCTs? This RCT requirement by the CMS for coverage of Aduhelm and all follow-on amyloid antibodies seems to undercut the FDA’s accelerated-approval mechanism, which has been used for years to introduce some early drugs for life-threatening diseases lacking alternatives.

    Who compiled the CMS’s summary of amyloid pathobiology in AD? There are numerous inaccuracies and misunderstandings among the statements and references they cite. For example, it is untrue that the ability to make a “clinical diagnosis of AD is poor.” Early non-amnestic symptoms in clinical AD are not “rare.” ASCVD/diabetes/obesity are not major risk factors for AD per se. The fact that Aβ monomers are normally produced is not a reason to conclude that lowering amyloid plaques and oligomers interferes with a critical normal function and could be hazardous. And citing that “Aβ protects the brain from infections [and] repairs leaks in the blood–brain barrier” is not widely validated and not a reason to avoid clearing oligomeric Aβ assemblies that have been shown in myriad reports to result from mutant presenilins or mutant APP or Down’s syndrome or ApoE4 and induce progressive synaptic dysfunction. All amyloids arise from normal proteins; clearing those amyloids helps patients.

    The CMS could have reasonably required patient registries that can collect detailed information about benefits and adverse events by those who treat aMCI/mild AD with amyloid antibodies. Our field is very anxious to rigorously validate approved AD therapies, but the new-RCT mechanism the CMS proposes is unworkable and will markedly limit the breadth of patients who might benefit from these agents.

  8. It’s important to recognize that this is a draft decision, and there may still be changes when the final decision is rendered in April. With that caveat, I view the current CMS position as a potential middle ground that could allow the field to move forward, together. If the CED program is large and national in scope, similar to IDEAS, it would enable access to patients while addressing key questions regarding efficacy and safety.

    I do not think CED can, or should, supplant the confirmatory placebo-controlled, randomized clinical trial required by the FDA. Biogen has already pledged to launch the confirmatory trial in May 2022. Furthermore, it is not possible to include a placebo control in CED since patients are liable for co-pays. How can one ask patients receiving placebo to endure co-pays for the drug, safety MRIs, etc.?

    However, there are other key questions that can be addressed within the construct of the National Coverage Decision without requiring a placebo control.

    Finally, I would expect the CMS to revise this decision and provide full coverage for any anti-Aβ monoclonal antibody that receives full FDA approval based on evidence of clinical benefit.

  9. I can only imagine how complex the decision-making process was, and I do have a few questions.

    I agree with the mandate for including persons representative of society. What are the criteria for establishing sufficiency in this regard?

    Who will determine whether there is a “statistically significant and clinically meaningful difference in decline in cognition and function”? Does the CMS mean to say that if the FDA determines that there is a “statistically significant and clinically meaningful difference in decline in cognition and function” but the CMS disagrees, there will not be coverage?

    If a trial is done in persons with minimal manifestations of Alzheimer’s disease, it may not be realistic to expect a “statistically significant and clinically meaningful difference in … function.” What then?

    Is the CMS indicating that a preclinical trial, commonly referred to as a “prevention” trial—that is, in cognitively unimpaired persons at high imminent risk for developing symptoms of Alzheimer’s by virtue of risk factors such as age, genetic susceptibility, and/or biomarker findings—would not be eligible for CMS approval (as a trial)? Several such trials are being conducted, including Trailblazer 3 and AHEAD. What about coverage if their results are “positive”?

  10. The CMS's proposed decision strikes me as wise and indicative of their interest in covering medications that help people. Their decision proposes a structure that will rapidly deliver rigorous evidence on whether medications such as Aduhelm have any benefits to outweigh their documented risks. I was especially pleased by their requirement that the studies be inclusive of the people most affected by Alzheimer's disease.

    It is disappointing that advocacy groups such as the Alzheimer's Association are speaking out against a CMS policy that prioritizes patients over pharma profits. The trials Biogen reported in their FDA filing for Aduhelm were simply inconclusive about whether there is any cognitive benefit, but those trials did conclusively show there is potential harm from this medication.

    I cannot understand how honest brokers can look at that evidence and conclude it justifies broad clinical use. It is especially disturbing given that the trials showed adverse events were common even in the trial participants, who were carefully selected to be at low risk of adverse effects. It is also shocking that anyone would consider it acceptable to adopt potentially harmful medications when only six (6) black patients received the drug as part of the trials.  We need more trials to know if these medications are safe or effective.

    I see nothing in the CMS’s proposed decision that limits how many patients can receive treatment. Many providers and patients have been justifiably hesitant to consider using this medication with such uncertain evidence on safety and benefits. By adopting Coverage with Evidence Development, the CMS will get more evidence on whether anti-amyloid therapies are beneficial. In the long run, this will improve access to the medications if they are shown to be beneficial. 

    I've repeatedly seen individual clinicians quoted as saying they witnessed patients improve in their own studies of anti-amyloid agents. When physicians make such claims, they undermine their own credibility and suggest they were biased in favor of believing the medication worked. Even if the most optimistic claims from Biogen about Aduhelm were true, an individual physician who started with a neutral opinion on whether the medication helped would be unlikely to care for enough patients to detect that any difference in outcomes among treated patients was not just due to chance. We are all desperate for effective treatments. I fear these clinicians are seeing what they hope for, not what's really there.

    I hope we as an Alzheimer's research community can band together to ensure rapid fielding of large, inclusive randomized trials. I advocate for pragmatic trial designs because I believe they will enable faster progress with more diverse participant bases. Regardless, the most important goal for all of us should be learning what works for people affected by AD. 

  11. The basis for aducanumab’s approval was an unvalidated surrogate biomarker and not FDA’s standard low bar of substantial evidence of effectiveness in clinical trials. The FDA does not view clinical benefit as necessary for marketing approval on the accelerated pathway. Indeed, the demonstration of clinical benefit would earn regular approval for a treatment. All the FDA needs in order to grant accelerated approval is to view a biomarker endpoint as “reasonably likely to predict clinical benefit.” “Reasonably likely,” however, is essentially a guess or opinion and is hardly compelling. All accelerated approval means here is, “We approved aducanumab because it reduces plaques.”

    This can’t sit well with the CMS or any health insurance company that reimburses based on whether a treatment is “reasonable and necessary,” which basically means clinical benefit outweighs harm, and with improved health outcomes. The CMS sees aducanumab as lacking sufficient evidence for effectiveness and so as not reasonable and customary treatment.

    Medicare, a taxpayer-funded, socialized medical insurance organization, should not pay for ineffective treatments. FDA essentially dropped an unfinished, undertested, potentially unsafe product, without evidence for clinical benefit, on the CMS’s doorstep, as if to say, “You deal with it, CMS."

    The CMS should have decided to not provide coverage and told Biogen to come back when it could show aducanumab is effective. Instead, they are offering a CED program that will pay for all needed effectiveness trials, including Biogen’s FDA-required trial, and any trials the NIH might approve and otherwise fund. The CMS is essentially sponsoring an RFA for aducanumab trials and becoming its sole funder. This is not a better decision, but it at least allows some access to aducanumab for people with early stage Alzheimer’s who are believers or uncertain about aducanumab’s effectiveness and provides for many clinical trials.

    There are numerous interesting considerations in executing these CED trials. First and foremost is Biogen’s buy-in. Will it agree to partner, manufacture, and sell product? How many trials, and for how many people, will the CMS fund? Given how slight sales are now, will there be enough trials participants to keep Biogen in business? Will Biogen be able to sell aducanumab off-Medicare to Medicare beneficiaries who might be willing to pay? Will every eligible person be allowed to enter a trial?

    We struggle for diverse representation in all Alzheimer’s trials, not to mention providing fair and equitable care to all people generally. The CMS’s diversity requirement is as strict as it can be, in that it asks for full sociodemographic representation. This is flatly unattainable because our national health systems aren’t structured to give equitable care to begin with. Also, eligibility criteria for nearly all Alzheimer-related trials are inherently discriminatory such that any given trials sample is a healthier, wealthier, more educated, and whiter population from the get-go. Some adjustments will have to be made.

    As a practical matter, the CMS’s CED coverage decision applies only to aducanumab and will have no effect on other amyloid-fibril antibodies. Time, new science, and new trials results will change the therapeutic landscape. The lecanemab and gantenerumab Phase 3 trials results, due to read out around September 2022, will give answers that will make this CMS coverage decision moot: One or both antibodies will meet the “substantial evidence for effectiveness” standard, or both won’t. If the former, then that treatment will get regular approval, the CMS will provide coverage, and the aducanumab CED studies will be abandoned if they had even started. If the latter, then we'd have the untenable circumstance of three antibodies with six negative or truncated Phase 2 and 3 trials between them, all having received accelerated approval because of plaque reduction, and all lacking evidence for clinical benefit. If this should occur, then all these treatments, accelerated approval and CED program notwithstanding, will collapse under their own weighty lack of evidence. Only donanemab will be left standing, also with accelerated approval, as it awaits its Phase 3 results in March 2023. 

  12. From my perspective, the CMS made a poor decision. Aduhelm was approved under accelerated conditions to allow patient access to treatment while additional data are gathered. This is how it is done with cancer therapeutics and how it should have been done here. This decision discriminates against patients with early Alzheimer’s compared to other patient groups. This decision deprives patients with early Alzheimer’s of the opportunity to engage in informed discussions about whether the benefit/harm balance makes Aduhelm the right choice for them.

    A particularly egregious aspect of the decision is that it applies to all anti-amyloid antibodies, including those for which efficacy and safety data are not yet available. This could limit patient access to treatment far into the future. The decision will likely discourage innovation and investment in drug development for Alzheimer’s disease.

    Regarding the CMS's diversity requirement, it has proven very difficult to recruit minority members to trials. We need to do better, but we cannot achieve this suddenly; it is a long-term trust-building enterprise. This requirement will slow the trial and delay access to treatment. More importantly, limiting Aduhelm to trials deprives minority, rural-dwelling, and other underserved populations from having access to Aduhelm. The treatment will be available only to those with access to trial sites.

    With this decision, Aduhelm availability will be limited to the sample size needed to show a drug-placebo difference in a well-powered trial. That is likely to be fewer than 5,000.

    I have provided consultation to Biogen on Aduhelm and other treatments.

  13. This is an interesting end-run strategy via another regulatory body that, in essence, “walks back” a very flawed FDA decision on Aduhelm. I think the CMS position reflects the overwhelming majority opinion of KOLs in the field.

    It is disturbing that the Alzheimer’s Association advocates against this stance. Everyone needs to follow the science here, which does not support sufficient clinical efficacy, or certainly sufficient benefit over risk and cost.

    There is apparently guidance for the FDA to consider diversity in clinical trials in its decisions on marketing approval, but I am unaware (and would love to be informed) as to how often this is actually followed. Dr. Manly was one of the first to call this out for the Aduhelm trials and deserves a lot of credit. Alzforum digging into that larger issue (how often this impacts FDA approvals and CMS guidance) would inform the field.

    The issue of diversity of trial participants is something we all are aware of, but it remains a huge challenge. My clinical colleagues in South Florida might be able to give a more firsthand experience on the challenges of enrolling diverse participants into AD trials. However, one thing I have not heard stated much for these immunotherapy trials is that these are very burdensome trials for the participants. Infusions, scans, long assessments, etc. … This adds an additional selection/enrollment bias in terms of educational attainment.

    I have not seen the Aduhelm data for this trial on educational attainment, but pretty much all the previous immunotherapy trials had a pretty big bias toward high levels of educational attainment. Subjecting yourself to this kind of therapy requires quite a bit of faith and hope in medicine, and there will be manifold reasons why someone with a higher educational level might be willing to take this calculated risk. 

    The larger question on this class of drugs is trickier. Ultimately, the merits of each drug need to be assessed with respect to efficacy over risk in the pivotal Phase 3 studies, without slicing and dicing the data. We can only hope that the data for those is more clear-cut in favor of efficacy. I think we can all read between the lines as to the message the CMS is sort of sending to the FDA.

  14. I think CMS made the right call here. We need to devote the money and effort required to ensure adequate recruitment of the minoritized. It must be done, and I believe since there are now—for the first time in clinical trials of AD—real consequences for not doing so, it will be done.

    I think we also need to expand to study persons with mixed-etiology dementias, particularly coexisting VCID. Studying a small, healthy, easy-to-study subset and then pushing the drug at marketing to a larger group is neither scientifically sound nor wise, ethical policy.

  15. We heard with surprise the news that the CMS plans to restrict coverage of aducanumab in the U.S. under its "coverage with evidence development" (CED) paradigm only, namely to the participants in a randomized clinical trial approved by the CMS or the NIH. I have three major concerns with this plan.

    Firstly, evidence development should primarily be conducted by well-controlled randomized clinical trials. The scientific decision for approval, and that for coverage, should basically be discussed separately, even if the CED paradigm is applied because of the lack of definitive evidence for efficacy. For aducanumab, the EMERGE/ENGAGE results are being publicized, which have been judged by FDA as being likely to predict clinical benefit. Although coverage should be strictly limited to those who meet the inclusion and exclusion criteria of the Phase 3/4 trials, it would not be reasonable that coverage be limited only to the participants of RCTs including placebo administration to this devastating progressive disease.

    Secondly, it may be quite tough to maintain the quality of trials under CED, and it may take considerable time to complete CED trials that target enrolling patients reflecting the diversity of the U.S. population. Lastly, it may not be rational that the CMS makes this determination for the whole class of drugs, each of which is likely to have a different profile than aducanumab.

    The Japanese regulatory agency PMDA and the Ministry of Health, Labor, and Welfare have discussed the approval of aducanumab in December 2021, reaching the conclusion to suspend their decision until additional data supporting the efficacy and safety of aducanumab are obtained. Other amyloid-removing antibody drugs await application for approval pending completion of their global Phase 3 trials in late 2022-2023. I believe the approval review process by PMDA, which should be based strictly on scientific evaluation, will be unaffected by the CMS decision in the U.S. Hopefully, discussions on the risk-benefit balance of the anti-Aβ antibody drugs will be further deepened with the results of the ongoing Phase 3 trials, and that will lead to better solutions on the CMS coverage decision for this class of drugs.

    Tremendous progress in therapeutic research of AD has accumulated evidence supporting the efficacy of disease-modifying therapies (DMTs) at the symptomatic, early AD stage. Now is the time to bring innovations in prevention and treatment of AD, and we have to further explore different types of DMTs indications as well, e.g., widening the application to asymptomatic, preclinical stages of AD.

  16. The plans of the CMS to restrict access to aducanumab only to people with early stages of Alzheimer’s disease involved in controlled clinical trials come as no surprise. In agreement with the FDA’s expert panel, and the European and Japanese regulatory agencies, the CMS believes that there is no convincing evidence that aducanumab produces clinically significant benefits in the cognition and function of AD patients. In agreement with many AD experts, the CMS does not agree with FDA that there is convincing evidence that monoclonal antibodies which have been shown to significantly reduce Aβ burden in the brain are likely to prospectively produce beneficial clinical effects in AD patients. The CMS pointed out that aducanumab treatment may produce significant adverse events including ARIA, which in some cases have been considered of serious medical concern.

    CSM believes that these adverse events are more closely monitored and treated in the context of a controlled clinical trial compared to general practice. The CMS has established strict criteria for controlled clinical trials with aducanumab, including Aβ-PET evidence of brain amyloid pathology, and will cover the cost of one Aβ PET scan. The CMS opinion concerns other monoclonal antibodies, such as donanemab, lecanemab, and gantenerumab, which have been shown to significantly reduce brain amyloid burden in AD patients but have not yet produced definitive evidence of cognitive and functional efficacy.

    It seems to me that the CMS has done the work that was expected to be done by the FDA during the evaluation of aducanumab. The CMS has come up with a scientifically logical and clinically prudent proposal that respects the expectations of patients and their caregivers and patients’ safety. The CMS has outlined clear boundaries in advance for the evaluation of other monoclonal antibodies: It will not cover the costs of drugs that do not demonstrate clinically significant cognitive and functional efficacy in well-controlled clinical trials and that do not have an acceptable safety profile. I believe EMA will follow this same approach with other apparently more promising monoclonal antibodies.

  17. In commenting on the first set of statements in the draft decision, it states that, “The CMS proposes to cover FDA-approved monoclonal antibodies directed against amyloid for the treatment of Alzheimer’s disease (AD) under Coverage with Evidence Development (CED) in the CMS-approved randomized controlled trials that satisfy the coverage criteria specified in Section C below, and in trials supported by the National Institutes of Health (NIH). All trials must be conducted in a hospital-based outpatient setting.”

    Given that, in the case of aducanumab, it was not clear from the two Phase 3 studies whether there was a clinical benefit; on the surface, this decision could perhaps make sense in regard to that particular drug.

    However, it is unclear to me from the draft decision what the CMS is deciding about the other anti-amyloid antibodies. I think making a blanket coverage decision for this class of anti-amyloid antibodies prior to them having undergone and completed FDA review does not make sense, at least as written. For example, if the FDA reviews, along with their advisers, other antibodies and finds that in a clinical trial(s), an anti-amyloid antibody both removed amyloid and had both a significant and what is interpreted as a clinically meaningful effect, would that fall under the new CED and require another CMS-approved trial?

    At minimum, the CMS needs to address these points and not apply the criteria they are proposing to any anti-amyloid antibody approved by the FDA if the approval is based on both a biomarker change (e.g., amyloid removal) and a statistical as well as what is interpreted as a clinically meaningful change as much as that is possible in the length of the trial that is conducted.

  18. The major question for the CMS regarded reimbursement for aducanumab; however, the regulatory pathway and questions surrounding efficacy are clearly linked to that consideration. The clinical review of the aducanumab BLA from the Office of Neuroscience at FDA was quite positive, while the review from the Office of Biostatistics was quite negative. This disparity arose from the fact that of three studies containing reasonable amounts of efficacy data—studies 103, 301, and 302—the clinical review based their opinion only on studies 103 and 302, while the statistical review was based only on studies 301 and 302. The Office of Clinical Pharmacology did fortunately consider efficacy data from all three trials, and their conclusion was that “the OCP review team supports regulatory approval of aducanumab.” They further state that, “The evidence of effectiveness is based on the totality of evidence from clinical trials 302, 301, and 103.”

    The accelerated approval pathway used for aducanumab is based on a biomarker, with an obligation to show clinical efficacy with an additional study. This regulatory pathway is used frequently in oncology and has been used in other disease states such as HIV and multiple sclerosis. The standard that must be reached is that the biomarker must be “reasonably likely” to predict a clinical effect. The “reasonably likely” determination in this case, that reduction in plaque load on PET scans is predictive of clinical efficacy, is not taken from aducanumab data alone. Two other antibodies, donanemab and lecanemab, have very similar characteristics in that they cause substantial reduction in plaque based on amyloid PET scans. Published data from a Phase 2 study of donanemab (Mintun et al., 2021) showed a statistically significant drug effect on the iADRS, a clinical composite scale that incorporates both cognition and function. Similarly, Phase 2 data from lecanemab (Swanson et al., 2021) showed consistent dose-dependent effects on clinical measures. Thus, the determination of “reasonably likely” is based not only from aducanumab data, but also from data on two other monoclonal antibodies with similar mechanisms of action. 

    The adage that FDA “ignored” the Advisory Committee which met in November 2020 has received much attention. The fact that the donanemab NEJM paper was published in March 2021 is rarely noted. The Advisory Committee had only data from aducanumab and lecanemab to consider; thus, the threshold of “reasonably likely” was not reached in November 2020. Following publication of the Phase 2 donanemab data in NEJM, given data from aducanumab, lecanemab, and donanemab, the threshold of “reasonably likely” was achieved in the view of FDA. Thus, the conclusions reached by the Advisory Committee in November 2020 and the conclusions reached by FDA in June 2021 were different, because the June 2021 conclusion was based on additional data that were not available in November 2020. 

    The CMS has apparently chosen to take a narrow view of the data and the field, and has only considered Phase 3 aducanumab data. Such a narrow view is unfortunate. An important learning is that very substantial reductions in amyloid plaque are likely required to slow disease progression, likely because amyloid plaques are a reservoir of more toxic species of Aβ such as oligomers. Bapineuzumab resulted in only small reductions in amyloid plaque and had no effect on disease progression (Vandenberghe et al., 2016). Gantenerumab, with the doses used in the SCarlet RoAD and Marguerite RoAD studies, resulted in modest reductions in amyloid plaque and the studies were stopped for futility (Ostrowitzki et al., 2018). These are important learnings for the field.

    Additionally, the NCD memo published by the CMS on January 11 states, “This NCD addresses anti-amyloid mAbs as a class since the drugs have a similar function of reducing amyloid in the brain.” This statement is incorrect since solanezumab, which targets Aβ monomers, does not lower amyloid plaque and is not associated with ARIA (Honig et al., 2018). Thus, the notion that all monoclonal antibodies used for the treatment of AD should be considered as a class is uninformed. 

    The CMS may disagree with FDA and conclude that substantial plaque reduction is not “reasonably likely” to predict slowing of disease progression based on clinical measures. While the CMS memo lists other clinical trials in AD, the data in these other studies were apparently not considered to any degree. The CMS should, however, consider all the data available in the field, rather than a limited analysis of data restricted only to Phase 3 studies of aducanumab.


    . Donanemab in Early Alzheimer's Disease. N Engl J Med. 2021 May 6;384(18):1691-1704. Epub 2021 Mar 13 PubMed.

    . A randomized, double-blind, phase 2b proof-of-concept clinical trial in early Alzheimer's disease with lecanemab, an anti-Aβ protofibril antibody. Alzheimers Res Ther. 2021 Apr 17;13(1):80. PubMed. Correction.

    . Bapineuzumab for mild to moderate Alzheimer's disease in two global, randomized, phase 3 trials. Alzheimers Res Ther. 2016 May 12;8(1):18. PubMed.

    . Correction to: A phase III randomized trial of gantenerumab in prodromal Alzheimer's disease. Alzheimers Res Ther. 2018 Sep 27;10(1):99. PubMed.

    . Trial of Solanezumab for Mild Dementia Due to Alzheimer's Disease. N Engl J Med. 2018 Jan 25;378(4):321-330. PubMed.

  19. With the caveat that the U.S. system is quite different from what we have in The Netherlands, I had expected some form of regulation and narrowing of the coverage to a specific population.

    This decision, however, goes one step further and is, in fact, putting the CMS in the position of the FDA, not only regarding Aduhelm, but especially regarding the other anti-amyloid drugs that are still under study and have not even been submitted to the FDA for approval.

    For Aduhelm, it seems the CMS is creating more inequality rather than solving it. The drug is on the market, and those who can afford it can be treated, while those who cannot afford it have to be put into a clinical trial with a 50 percent chance of receiving placebo. The CMS will cover the costs of the drug and one amyloid PET scan, but who will set up and pay for the clinical trial, which needs to be lengthy in order to show the required cognitive and functional benefit?

    Given the prospect of seeing three other drugs in the same class reading out in one to two years’ time, it is not very likely that such a trial with Aduhelm will be initiated. Besides, the FDA already ordered Biogen to initiate a trial.

    As for Europe, the EMA decision was expected since they put much weight on clinical (and functional) outcomes. Reimbursement is in the hands of each individual country, hence may differ within Europe.

    I do wonder if this all, or in part, could have been prevented had Biogen priced Aduhelm more reasonably. They have really shot themselves in the foot, and cannot say weren't warned.

  20. There was no one perfect decision open to the CMS, but I believe their evidence review team followed the evidence and looked at it as objectively as they could. The narrative of their own evaluation of the evidence is clear: They found serious uncertainties about the evidence on the benefits of this treatment, whereas its risks, even in the short term, are notable. Having all patients who get coverage participate in trials that will help us learn whether the benefits do outweigh those risks seems like a principled, reasonable decision.

    Of course, that means some patients who would like the treatment today will have to wait until the trials become available, and these patients might slip past the early stage of AD for which the treatment may work. Even people who enroll in the trials will be randomized and may get a placebo instead of this treatment. And there will be higher burdens to seek enrollment in these studies for people who live far from academic health centers. Those are the potential downsides. But they are part of a tradeoff needed to limit the potential harm to tens or hundreds of thousands of people in the country who might otherwise get a treatment that creates real harm.

    I am also glad that the CMS is proposing to mandate that these clinical trials take the effort to enroll representative populations, including black patients, who were almost nonexistent in the company’s trials. I view that as an important precedent that will help mitigate some of the barriers to enrollment that people from diverse communities have experienced in the past.

    Regarding how this might this affect other anti-amyloid antibodies in development, the big question is whether some of the developers will now view it as preferable to accelerate their clinical studies and seek FDA approval using the regular pathway when we can all actually know whether the drug works. The financial benefits of getting to the market as early as possible may mean that drug developers continue to opt for the accelerated approval pathway based only on amyloid reduction. But do patients really need a choice among multiple drugs that reduce amyloid without knowing whether any of them help people? Better in my mind for patients, and for Medicare, if companies focus on doing the studies needed to demonstrate whether the drugs actually work or not.

    Regarding Biogen's recent decision to cut the price of aducanumab in half, lowering out-of-pocket expenses for patients is a very important goal, as is helping to keep Medicare and private insurance premiums affordable. But half off a grossly inflated price is not a bargain. Half off for a drug whose benefits at best—using the one positive study—would support a price closer to $5,000, is not a bargain. This price drop signaled that Biogen knows it got the launch price wrong. But the $28,000 price, and how we got there, does not serve as a good model for the way pricing should be done to incentivize important advances in care that come with the evidence we need to know we are doing right by patients and their families. 

  21. These are difficult times for everyone in the field, especially patients, caregivers, and clinicians.

    What I can say right now is that Roche/Genentech supports broad and equitable access to anti-amyloid therapies for all appropriate patients with Alzheimer’s disease. Our detailed position on the U.S. Centers for Medicare & Medicaid Services’ National Coverage Determination (NCD) proposal will be available in a few weeks when we submit our comments to the CMS during the 30-day public comment period. We are still working on this with a team of people, including those who have a public policy perspective within the company.

    Meanwhile, we are highly focused on completing our ongoing Phase 3 trials this year to really test the question of whether lowering fibrillar amyloid leads to clinical benefit. We hope to have clear answers soon.

  22. It is clear that the CMS do not believe that sufficient data are yet available to establish the clinical meaningfulness of aducanumab therapy in Alzheimer’s disease. Thus, their proposed stance is that they will cover aducanumab therapy within the Coverage with Evidence Development (CED) scheme in CMS-approved randomized controlled trials. How much this will influence the decision processes in other companies regarding anti-amyloid therapies and Alzheimer’s therapy in general is difficult to evaluate. I hope that, as a field, we do not lose momentum again. 

    With regard to anti-amyloid therapy, the apparent paradox between the genetics and basic research—which demonstrate unequivocally a driving or triggering role of amyloid pathology—and the clinical outcome of the anti-amyloid trials remains unresolved. In an upcoming review in Nature Reviews Drug Discovery, Eric Karran and I will offer a perspective on the extant data from those anti-Aβ antibodies that have been in, or are entering, Phase 3 clinical trials. In general, anti-amyloid trials have lacked a precise therapeutic hypothesis, which makes the interpretation of their results complicated.

  23. The CMS's plan to restrict coverage for aduhelm under its “coverage with evidence development” was the right decision in light of the inconsistent efficacy data, safety concerns, and lack of diversity of participants enrolled in the ENGAGE and EMERGE Phase 3 trials. It was the right decision given the FDA’s approval of the medication that was based on the hopeful but not definitive data that came from those trials. And it was the right decision given the disconnect between the characteristics of the people studied in the safety and efficacy trials and those to whom the drug is marketed. The decision, which applies to the entire class of anti-amyloid therapy, broadcasts an important message that is obvious but often obscured: Medications need to be effective and safe.

    The decision, or “The Decision” (as I suspect it will be referred to), naturally has generated a lot of discussion about demographic representation in clinical trials and about its potential differential impact across race and ethnicity groups. I was profoundly disappointed by the statement released by the Alzheimer’s Association that framed the CMS memo as discriminatory against individuals from minoritized communities because of their lack of “access to research institutions.” This, of course, is a specious argument, unreasonably putting the “blame” for a lack of diversity in our clinical trials on minoritized groups because they simply can’t get to research institutions. Rather, many or most research institutions are situated in diverse areas yet fail to engage with the communities they ostensibly serve, failing to build the trust, the careful community outreach, and the introspection required for diverse and representative trial participant composition.

    Big pharma seems to have unlimited resources for drug development, lobbying, and marketing. I am hopeful that the CMS decision communicates a message that inspires the industry to use some of its resources toward improving the structural and institutional barriers that have restricted the representation of trial participants.

  24. I agree with others about following the science. The decision by the CMS regarding plans to restrict coverage only to people enrolled in a randomized clinical trial approved by the CMS or NIH is reasonable, not only for aducanumab, but I think also for this whole class of drugs. Perhaps an underlying reason for this requirement may be the recognition that amyloid–based therapy does require a much longer trial than is traditionally possible for a pharmaceutical company to afford (i.e., the trials should be at least three years in duration, rather than 18 months). So by paying for the drug in these extended trials, the government would be subsidizing pharma to complete a trial that has a better chance of showing efficacy.

    With respect to the diversity issue, everyone tries to meet these recruitment criteria for minorities, for any drug, but the truth is that few actually succeed in recruiting even 15-20 percent of minorities. The challenges are not only in recruitment, but once recruited, the bigger challenge is in retention—especially for a trial which is quite burdensome for any elderly person and their (often) elderly caregivers to undertake.

    For minorities, the financial issues involved (time away from work and transportation) are also a major deterrents for recruitment and retention. What usually happens is that efficacy and safety of the drug is demonstrated in an overwhelmingly nonminority patient population, but subsequent post-marketing studies, focusing largely on safety issues, have a larger proportion of minorities.

  25. I am very disappointed by this decision. It is essentially denying coverage to aducanumab and other drugs in the class. I agree with the FDA decision that dramatic reduction of brain amyloid, as seen with aducanumab treatment, is reasonably likely to result in clinical benefit; this view is supported by clinical trial results with donanemab, lecanemab, and aducanumab.

    If the CMS decision holds, this benefit will be restricted to the few individuals who can afford to pay the cost of treatment or enroll in randomized trials.

  26. The CMS correctly analyzed the problem, and they made the right decision. From page 17 of the CMS document, this statement is irrefutable: “Our conclusion is that Biogen’s secondary analysis cannot overturn, definitively confirm, or substitute for, the RCT evidence. With conflicting results from two RCTs (EMERGE and ENGAGE), and a secondary analysis that did not resolve the difference between the two RCTs, CMS believes that the available evidence is insufficient to establish that the treatment is reasonable and necessary under section 1862(a)(1)(A) of the Social Security Act.”

    The fact is that both the FDA and the CMS agreed that the evidence for clinical benefit was simply lacking. By recognizing that Medicare patients ought to have access to medications that have benefit for them, the CMS made the right call by requiring a placebo-controlled trial with a clinical outcome that proved or disproved benefit on clinical grounds within a reasonable period of time. For example, within 18 months.

    Given the expense of mounting such a trial, I assume it would have to be funded by the NIH. I may have not read the CMS document carefully enough, but I am unclear whether Biogen could apply. As a consequence, I cannot conceive of what happens next, but it seems likely there will be some considerable lag time before such a trial launches.

    However, launching a new trial of aducanumab (which Biogen could have started now nearly three years ago) needs to be put in the temporal context of the appearance of Phase 3 trial results of lecanemab, donanemab, and gantenerumab on the scene in the next year or so. If one of those drugs were shown to be safer and more effective clinically than aducanumab, it seems to me that interest in aducanumab would plummet.

    Disclosures: I am a site investigator in the aducanumab EMBARK study sponsored by Biogen and in the A4 study sponsored by Lilly and USC. I am on a DSMB for the DIAN-TU and receive compensation from Washington University.

  27. The CMS proposal adds to the growing controversy by restricting aducanumab’s coverage within clinical trials. This proposal effectively restricts its prescription to people with Alzheimer’s disease enrolled in clinical trials due to the drug’s high cost (even after the recent price reduction), and the proposal extends beyond this one agent to include others not yet out of clinical trials.

    The PET-centric coverage decision also introduces unnecessary heterogeneity in these future trials, as many studies have shown patients with Lewy body disease or frontotemporal lobar degeneration to have amyloid co-pathology but cognitive impairment more attributable to a non-amyloid etiology (in the absence of AD tangles). This has been historically the advantage of CSF-based biomarkers over amyloid PET, and diagnostic accuracy in this instance seems paramount if the goal is to develop more evidence.

    At the same time, clinical trial populations in the U.S. are largely white, had above-average education, live in urban centers, and have relatively easy access to health care. Restricting this drug’s coverage to clinical trial settings has the potential of worsening existing disparities, as existing clinical trial infrastructures are highly unlikely to increase their enrollment of underrepresented patients by multiple folds during, or on the heels of, the COVID-19 pandemic. Health disparities extend beyond racial lines: People who have low vision, hearing, English proficiency, or resources develop Alzheimer’s disease but are historically excluded from clinical trials. Exceptions must be considered for them to receive these new treatments.

    I also wish to point out that inclusion of diverse patients in clinical trials according to census or known risk levels will be inadequate to determine the drug’s efficacy in these populations, which will require overrepresentation to achieve the necessary statistical power. If a clinical trial setting is the way forward, it will be essential to require much more than proportionate numbers (e.g., 50 percent) of AD patients traditionally underrepresented in clinical trials.

  28. For better or worse, when the FDA allowed Biogen to market Aduhelm, it let the genie out of the bottle and forced stakeholders to navigate a world with a lot of unanswered questions. Patients and physicians have struggled to consider whether the upsides of treatment outweigh the downsides. After all, we can’t really perform the usual risk-to-benefit ratio calculation because although we do know there are risks, we don’t know if there are clinical benefits. Some would argue clinical benefits were convincingly demonstrated, even if they were down in the weeds, while others argue just as strongly that clinical benefits were not demonstrated.

    Medicare has had to struggle with the possibility that they might divert a reasonable chunk of the nation’s wealth to a drug that may not help people. Biogen apparently struggled with how much they should be charging for their product. The bottom line is that 10 years from now, if Aduhelm is found to benefit people, none of us wants to look back and wish we had prescribed it more. This also works the other way. If 10 years from now, Aduhelm is found not to benefit people, none of us wants to look back and wish we hadn’t prescribed it. It is an unpleasant situation to be in, isn’t it? It would have been better if the crucial question of efficacy had been resolved before unleashing Aduhelm.

    So now the CMS decision comes down, and it seems to try to limit liability, at least to the nation’s finances. It’s the same basic fear, that 10 years from now we don’t want to come off as having been too supportive of a drug that didn’t work, or not supportive enough of a drug that did work.

    We shouldn’t be in this position. We should simply know if it works or not. Maybe over the next four to nine years we’ll get our answer. In the meantime, I’d say the CMS decision is reasonable. Will it slow down the adoption of Aduhelm in routine clinical practice? That seems likely. Will it slow down the entry into the clinical arena of other monoclonals? I would guess yes, if those monoclonals have not been shown to benefit people clinically. If they are convincingly shown to help people, then maybe not. Will this decision make the drug less accessible to those who can’t afford it? That seems likely.

    These considerations do not negate the big, overarching question: Does Aduhelm clinically help people? Now it looks like Medicare will support trials designed to address this question. To me, this seems like a compromise.

  29. The CMS has made the right call in proposing coverage with evidence development. As the controversy surrounding aducanumab’s approval shows, many people remain unconvinced that amyloid clearance is a useful surrogate endpoint for the clinical outcome patients, families, and clinicians care most about: protecting and preserving cognitive function. Moreover, while there is uncertainty about benefits, there are clear risks.

    Further research is clearly necessary on these grounds alone. Additionally, given that aducanumab trials were not at all representative, requiring that studies reflect the diversity of the U.S. population is essential from a scientific and ethical perspective.

    [Editor’s note: For a recent viewpoint on FDA Drug Approval and the Ethics of Desperation, see Largent et al., 2021].


    . FDA Drug Approval and the Ethics of Desperation. JAMA Intern Med. 2021 Dec 1;181(12):1555-1556. PubMed.

  30. The field is in a tough spot. It’s difficult to criticize the decision by CMS, even though it is largely unprecedented. The evidence supportive of a clinically meaningful benefit simply isn’t there yet. What was really needed was another trial of aducanumab before approval by the FDA (if an additional trial was positive). Then the traditional evidentiary standard might have been achieved and perhaps the CMS decision would have been different.

    Regarding the requirement of coverage only in controlled studies, I think this is an important and appropriate component of the requirement. Uncontrolled studies are at too great a risk of bias. The field is urgently in need of answers to important questions about this class of medications. Whether it is for aducanumab or another agent, convincing demonstration of clinical benefit is the only path to these drugs having the impact sponsors desire. 

    That said, some of the critiques of this decision aren’t off-base. The decision will limit the use of approved treatments (for now, just aducanumab) and certainly risks exacerbating disparities in AD treatment use. That of course, at least for aducanumab, is access to a drug that has important safety risks and lacks solid evidence to support benefit, but this does speak to the challenges that are stemming from the FDA decision.

    Formularies are rejecting aducanumab, and that is taking the drug out of the hands of the dementia experts who are best positioned to use it appropriately. I have real concerns about this aspect of the current situation.

    Hopefully the Phase 3 and 4 trials can get done efficiently and resolve some of these disputes and controversies, and help us continue the remarkable progress we’ve made as a field toward meaningful solutions for patients and families dealing with symptomatic Alzheimer’s disease. 

  31. In its decision, the CMS requires that “The diversity of patients included in each trial must be representative of the national population diagnosed with AD.” Ensuring that Alzheimer’s clinical trials recruit from a diverse population of patients is critical. Unless this principle is followed, the generalizability of findings from trials recruiting predominantly from one homogenous group of patients is limited. In the case of aducanumab, this is even more important due to the high incidence of brain swelling (ARIA-E), an adverse event that occurred in 35 percent of patients receiving the high dose, which is approved as the dose on the prescribing label of the drug. The risk of ARIA-E is more than two times greater in APOE4 carriers, and in those who have brain microbleeds at baseline.

    Given that African Americans are 1.4 times more likely than Caucasians to carry the APOE4 allele, and considering that they are also at higher risk for cerebrovascular disease, it is important that this group of patients is adequately represented and studied in these trials to fully assess the adverse effects and safety of the drug. African Americans are also disproportionately affected by AD. Despite this, it is striking that the two Phase 3 trials of aducanumab that recruited more than 3,200 patients, only randomized six black patients to the high-dose arm.

    The CMS mandate to ensure “diversity of patients” in the next generation of aducanumab trials is therefore an appropriate requirement. Given that brain swelling and microbleeds are a common feature of this class of drugs, it is equally appropriate that this requirement extends to other amyloid-lowering antibodies that are likely to come up for FDA approval over the next one to two years.

    Regarding the feasibility of the diversity requirement, there are many reasons for the challenges in recruiting minority participants into Alzheimer’s clinical trials. Not all of them can be overcome immediately. This will take a concerted effort over time and will need to rely upon approaches at the community level that enhance trust in the medical establishment and reinforce the importance of participating in clinical research. Community outreach and enhancing awareness among minority patients/caregivers must be a priority, not only during post-approval marketing of a drug but also at the early stages of designing clinical trials.

  32. I am very happy with, and in favor of, the CMS coverage decision on Aduhelm. It goes a long way toward addressing the FDA's wrong-headed decision to approve Aduhelm using the accelerated approval pathway. In my opinion, the FDA decision is unsupportable from both scientific and clinical perspectives, in that the reliance on amyloid reduction as a surrogate marker for treatment benefit is not currently supported by studies in the literature, and the Biogen clinical trials do not provide convincing evidence that Aduhelm produces clinical benefit in terms of cognition or daily function, while there is clear evidence that Aduhelm treatment is associated with significant risk of dangerous side-effects.

    I am in favor of the CMS imposing this coverage decision on the entire class of amyloid-directed monoclonal antibodies, since the same lack of evidence for efficacy in the literature currently applies to all of these drugs. I am also in favor of the requirement to increase the number of individuals from minority populations on whom the treatments are tested, given the potential for greater side effects among those with, for instance, more cardiovascular and cerebrovascular disease, as is the case, on average, among African American older adults.

  33. I think the CMS made the right decision, except it might have been better to ask the company to run additional, required, randomized controlled trials. A limited number of patients will be able to access the drug under these conditions. Perhaps that is a good thing until the outstanding questions are addressed. Remember, this is not a completely safe or inexpensive drug.

    Whether you think the amyloid hypothesis is right or not, we all agree the only way to test it is through rigorous clinical trials. I have always supported such trials, because the scientific method teaches us that only through “experimental medicine” can a hypothesis be tested.

    At the same time, there is a sense that amyloid plaques outside of neurons reflect the true fire burning within neurons (Petsko and Small, 2020). In this articulation of the amyloid hypothesis, clearing amyloid is like clearing smoke. It will not extinguish the fire.

    A re-examination of APP and the PSENS in IPSC in neurons shows that they cause traffic jams before they coax neurons to secrete amyloid. This agrees with the genetics of late-onset AD, which implicates the neurons' trafficking machinery. It also agrees with molecular mining of hippocampal neurons, where the disease begins, and which also implicates the trafficking machinery. I believe this triangulation might point to medical true north.

    It is very possible that clearing amyloid will have some mild benefits. And in the absence of anything better, that is better than nothing. But it suggests that we should start expecting only mild benefits, and not be surprised that an amyloid-clearing drug has never shown anything but. Perhaps this should be incorporated into the statistical assumptions on effect size.

    How the CMS decision might affect other anti-amyloid antibodies in development is a complicated question. I wonder if might affect the FDA's stance toward amyloid as a surrogate marker for accelerated approval. I think there might be better surrogates, perhaps tau. We desperately need a surrogate marker to expedite these very expensive clinical trials.


    . Endosomal recycling reconciles the Alzheimer's disease paradox. Sci Transl Med. 2020 Dec 2;12(572) PubMed.

  34. Overall, I think the CMS made a reasonable decision amidst a controversial approval by the FDA. The evidence for benefit is weak and the evidence for harm from ARIAs is convincing. The decision for Coverage with Evidence Determination creates a setting where clinical trials for this class of anti-amyloid therapy can be performed. It will take clinical trials to develop convincing evidence for this approach one way or the other. To date, none of the material that is available to me, anyway, suggests any meaningful benefit, in terms of clinically important difference. On the contrary, if the CMS had decided otherwise, then widespread use of this class of medications outside clinical trials would quite likely have more apparent harm.

    The key question (and raised in the material cited by the CMS) is the mixed nature of neurodegenerative processes found in persons with Alzheimer’s disease and later-onset AD. It will take the kind of research that is possible under Coverage with Evidence Determination to get an answer to this general question. Does removal of brain amyloid improve brain health, minimize progression?

    As many have written, to date, after years of sincere efforts through multiple clinical trials to “prove” that intervening in the amyloid hypothesis cascade helps patients, we still lack that understanding. It would help if the aducanumab trials were published in peer-reviewed journals, which, as noted in the CMS material, has not happened to date.

  35. The requirement to enroll from traditionally underrepresented populations is an important step forward, not just for aducanumab but for amyloid-modifying treatments that will follow. I expect that new approaches and procedures will be used to include traditionally underrepresented populations, which may guide the way for future amyloid modifying treatments.

  36. I believe that it will be important for Medicare to pay for aducanumab, particularly for those who participated in its clinical trials. That is only fair to those who participated. In the future, I hope that Medicare will pay for FDA-approved treatments for Alzheimer’s disease.

  37. CMS spends many millions of dollars per year on other FDA approved drugs for which the supportive data are far weaker than Aduhelm. I am speaking specifically of drugs such as Aricept and Namenda, which can run thousands of dollars per year in costs per patient. If Aduhelm is cut off for CMS funding, then those other medications also should be cut off until more clinical trials are conducted. Then we leave Alzheimer's patients with zero options aside from a handshake and certain death.

    As an alternative to a handshake and a death sentence, I recommend providing CMS funding for Aduhelm for patients who have already started taking it, and for all those in a clinical registry such as iCare, so that their progress can be assessed. In parallel, a larger Phase III study can be considered, but a randomized controlled trial is neither ethical or practical at this point.

    There needs to be good clinical judgment applied, as with any other medication, and obviously I do not propose indiscriminate prescribing of Aduhelm to all patients with Alzheimer's disease. My opinion is that those with pronounced brain atrophy, MMSE or MoCA scores less than 10, and CDRS over 2 are unlikely to benefit. But selected patients with MCI and mild dementia should not be short-changed.

    The patients I have selected for treatment with Aduhelm are generally young and most have a history of familial autosomal-dominant Alzheimer's disease, currently with MCI. We know that these patients are doomed to get Alzheimer's at an early age, and my hope is to buy them some time before they become more symptomatic. In the meanwhile, other drug options are in the pipeline and can be tried later. But at this time, Aduhelm is the only disease-modifying drug for Alzheimer's disease.

  38. CMS is essentially conducting its clinical trial. The drug is efficacious at eliminating senile plaques as a biomarker. However, since amyloid deposition can predate degeneration and dementia by twenty years, the importance of the lesion in the clinical stages of the disease is unclear. Given that the drug is likely to need supervised treatment with a high risk of side effects and is very expensive, it is important to show functional efficacy. No patient cares that they carry a heavy amyloid load; they only want to function normally. Hopefully, the placebos are appropriately selected, and the treatment outcomes are well documented across the spectrum of caregivers.

Make a Comment

To make a comment you must login or register.


Therapeutics Citations

  1. Aduhelm
  2. Gantenerumab
  3. Leqembi
  4. Donanemab

News Citations

  1. Aduhelm Administration Remains a Trickle, ARIA a Concern
  2. Aduhelm Phase 3 Data: ARIA Is Common, Sometimes Serious
  3. IDEAS Finds Small Drop in Hospitalizations, Missing Goal

Paper Citations

  1. . Disease severity and minimal clinically important differences in clinical outcome assessments for Alzheimer's disease clinical trials. Alzheimers Dement (N Y). 2019;5:354-363. Epub 2019 Aug 2 PubMed.

External Citations

  1. draft decision
  2. coverage with evidence development
  3. Biogen
  4. Eisai
  5. Alzheimer’s Association
  6. UsAgainstAlzheimer’s
  7. Global Alzheimer’s Platform
  8. Politico article
  9. letter
  10. InvesBrain news story
  11. Endpoints story

Further Reading