18 Feb 2022

Behold the deluge. The Centers for Medicare and Medicaid Services’ (CMS) proposal last month to restrict Aduhelm coverage to clinical trials stirred up such strong feelings on both sides that the usual 30-day public comment period following every CMS draft decision drew a record 9,956 submissions posted to date, nearly crashing this government website at times and dwarfing the 132 commenters who had weighed in during the agency’s initial feedback period last July (Jan 2022 news). The comment period ran from January 11 to February 10, but CMS continues to process and post submissions that poured in at the end. CMS will issue its final decision in April.

Two letter-writing campaigns—one pro, one con—inflated this number, with thousands of those comments using nearly identical language. Among the unique, often lengthy contributions, healthcare organizations largely commended the CMS proposal, while patient advocacy groups implored the agency to open up coverage for aducanumab and similar anti-amyloid antibodies to come. The issue has become politicized, with 78 Republicans in the U.S. House of Representatives signing a letter to the CMS demanding broader access, and liberal groups urging the agency to resist what they see as pharma raiding Medicare for a drug that does not work. “Alzheimer’s disease has been swept up into the red state versus blue state rhetorical vortex,” Jason Karlawish at the University of Pennsylvania, Philadelphia, told Alzforum (full comment below).

Verbal fireworks aside, Alzheimer’s researchers said the fundamental issues remain. “Despite attempts by some to politicize this debate … the facts are the same now as they were prior to the CMS announcement,” David Knopman at the Mayo Clinic in Rochester, Minnesota, wrote to Alzforum. Ian Grant at Northwestern University, Chicago, agreed. “[The comments] represent a further entrenchment of positions that were brought out by the original FDA approval last June,” Grant wrote (comments below).

AD researchers themselves are divided. They agree that aducanumab’s Phase 3 efficacy data were inadequate but, beyond that, some believe the FDA should revoke approval altogether, some support CMS’ proposal as a reasonable compromise, and some say running a randomized clinical trial for an approved drug would be unworkable. This latter group favors registry-based studies instead, similar to past coverage with evidence development (CED) requirements.

Scientists also agree that each antibody in this class should be evaluated on its own merits. Several researchers challenged CMS to clarify not only how upcoming anti-amyloid antibodies will be treated, but also what would constitute adequate evidence of efficacy, and how the CED process might end. Mark McClellan and colleagues at the Duke-Margolis Center for Health Policy in Durham, North Carolina, published a brief summarizing key questions raised by the draft NCD, and suggesting these be addressed in the final decision.

“Pro” Commenters Focus on Cost, Safety, Efficacy

Perhaps unsurprisingly, healthcare organizations that would foot the bill for aducanumab use strongly supported CMS’ effort to limit cost. These included the Pharmaceutical Care Management Association, which represents the pharmacies that administer most of the prescription drug plans in the U.S., the Academy of Managed Care Pharmacy, a professional association of pharmacists, and Doctors for America, an organization of more than 20,000 physicians. Hundreds of individual commentators reflected this concern, writing to the effect of: “My loved one suffers from Alzheimer’s, but we do not want false hope,” and “Most of us live on fixed incomes; do not raise Medicare premiums over this drug.”

The situation is more complicated for state Medicaid programs, which would find themselves on the hook for payments Medicare declines. In its comment, the National Association of Medicaid Directors asked CMS to allow it to limit coverage. “Just as CMS utilized its clinical judgement to set narrow coverage parameters for Aduhelm, so too should states have the ability to employ sound clinical decision-making in creating medical necessity and prior authorization criteria,” the organization wrote. State Medicaid programs weighed in, as well. The Colorado agency asked CMS to add language to its National Coverage Determination that the CED restrictions would apply to people dually covered by Medicare and Medicaid.

What do neurologists think? Many support the NCD based on concerns about aducanumab’s efficacy and safety, according to a survey conducted by Spherix Global Insights. Just over half of 75 responders agreed with CMS, 15 percent disagreed. About two-thirds said they would be willing to enroll patients in a CMS-approved trial of aducanumab (Endpoints news). 

Relatively few neurologists or other practitioners submitted individual comments to CMS. One, from a group of geriatricians at the University of Connecticut, noted that after the FDA approved aducanumab, “We were put in a position of having to disappoint patients and caregivers. … We applaud the decision of CMS to mandate the collection of more data about the safety and efficacy of aducanumab.”

More than 4,600 comments in support of the decision came from a letter-writing campaign organized by More Perfect Union, a political group affiliated with Senator Bernie Sanders (D-VT). Most of these used boilerplate language accusing the FDA of demonstrating a “stunning disregard for science” in its approval of aducanumab.

“Con” Commenters Push For Access

On the other side of the issue, patient advocacy organizations (PAOs) oppose the CMS’ draft proposal. More than 50 such groups representing numerous diseases banded together to submit a letter demanding that Health and Human Services Secretary Xavier Becerra intervene. They asked CMS to cover the drug for anyone who meets the FDA label requirements. Meanwhile, the Health Equity Collaborative, a coalition of groups concerned with equity in healthcare, submitted a comment asking CMS to allow pragmatic trial designs rather than randomized clinical trials, arguing that this would encourage more participation by underrepresented groups. Many PAOs worry the CMS decision would set a precedent for other drugs, where FDA accelerated approval would no longer be considered sufficient for coverage.

Alzheimer’s advocacy groups such as the Alzheimer’s Association and UsAgainstAlzheimer’s argued the case for broad coverage. This came amid attention on industry funding for such PAOs (for example, see donor list on page 21 of UsAgainstAlzheimer’s 2020 report; Endpoints news). Other groups registering opposition to the NCD included the Society for Women’s Health Research and regional groups such as Montana’s Big Sky Senior Services.

A more surprising campaign against the CMS’ NCD came from Down’s syndrome advocates. They generated more than 3,300 comments, many of them identically worded, demanding aducanumab be made available to this population. People with trisomy 21 accumulate amyloid plaques as they age, because amyloid precursor protein lies on the duplicated chromosome. As do exclusion criteria for most randomized controlled trials in AD, the NCD currently bars people with neurological conditions other than AD from participating in a CMS-approved trial of aducanumab. As is, this would exclude people with Down’s unless a separate trial of aducanumab is performed in this population.

House Republicans led by Cathy Rodgers (R-WA) wrote a letter to HHS asking for broader aducanumab coverage for Medicare beneficiaries, in particular those with Down’s syndrome. Rodgers is the founder and co-chair of the Congressional Down Syndrome Caucus, and mother of a son with the condition.

Eric Siemers of Siemers Integration LLC cautioned that the issue is complicated because people with both Down’s and AD tend to have more vascular amyloid than do other Alzheimer’s patients, potentially raising their odds of developing the brain edema known as ARIA. “The risk-benefit ratio in this population has not been determined. Randomized controlled trials in patients with Down’s Syndrome should be undertaken as quickly as possible,” Siemers wrote to Alzforum (full comment below).

Clinician-scientists have been laying the groundwork for such trials by building trial-ready cohorts from observational studies, and a few such trials have started, for example for the anti-Aβ vaccine ACI-24 (May 2021 news).

RCTs versus Registries

Some experts who object to the draft NCD question whether the proposed randomized clinical trials are feasible. A group of neurologists in Colorado noted that their state has only one site capable of mounting hospital-based AD trials, the University of Colorado Anschutz Medical Campus, and limited personnel to run them. “There are simply no additional resources at this time to accommodate new CMS-directed clinical trials,” Kavita Nair at CU Anschutz and colleagues wrote to CMS.

Paul Aisen, Jeffrey Cummings, Howard Fillit, Marwan Sabbagh, and Dennis Selkoe co-authored a letter to CMS calling such trials unworkable. “We have significant concerns about the viability of executing these CMS RCTs, given open questions about who will organize and fund them and about anticipated challenges in recruiting patients,” they wrote, adding that the proposed trials would duplicate the FDA-mandated trial Biogen is already planning.

“CMS needs to abandon the randomized controlled trial requirement and find a more balanced approach that provides access, generates real-world data on safety and efficacy, and stimulates investment and innovation,” Stephen Salloway at Butler Hospital in Providence, Rhode Island, wrote to Alzforum (full comment below). Salloway and other leading clinicians advocated gathering safety and efficacy evidence via so-called pragmatic trial designs, such as patient registries, Medicare claims data, and electronic health records.

In agreement with this point, the Duke-Margolis brief noted that real-world evidence gleaned from registries and health records would be able to draw from larger and more diverse patient populations than those who participate in trials. “Such large longitudinal post-market studies would likely be more informative than RCTs about some of the important safety questions related to the use of monoclonal antibodies,” McClellan and colleagues wrote.

Support for this view came from Gil Rabinovici at the University of California, San Francisco, who runs an open-label CED registry study for amyloid PET imaging (Aug 2020 news). “[This] would address critical questions regarding clinical response and safety in ‘real world’ specialty practice, as we await complementary results from the pharma-sponsored confirmatory RCT,” he wrote in his CMS comment.

The Alzheimer’s Association is developing one such national registry, ALZ-NET, which will capture data on multiple disease-modifying AD treatments (Nov 2021 news). UsAgainstAlzheimer’s is building a public-private partnership modeled on the COVID-19 Evidence Accelerator. Dubbed the Alzheimer’s Disease Evidence Accelerator, ADEA is to foster data-sharing via a new database called AD Real-World Evidence Learning System, which would “… compare and analyze data from existing sources of real-world data such as electronic health records, insurance claims, Phase IV trials, and registries,” the organization wrote to CMS.

Biogen itself, in a 31-page response to CMS, promoted alternative means of gathering efficacy data, proposing a “three-pronged approach” consisting of their ICARE-AD observational study, a data-sharing network, and Medicare claims (Aug 2021 news).

Filling in Gaps in CMS Ruling

The Duke-Margolis analysis laid out other issues the draft NCD does not address. For example, it asked what evidence would be needed to expand coverage for other anti-amyloid antibodies, what level of cognitive benefit would be considered clinically meaningful, and what would be needed to validate plaque reduction as a surrogate marker. Alzheimer’s researchers agreed that these points ought to be clear in the final ruling.

In particular, the CED process typically does not set an end date, or define clear criteria for success. It can go on for a decade or more. The CED for amyloid PET imaging has been in effect for nearly nine years, with no end in sight (Jul 2013 conference news). The American College of Radiology and the Society of Nuclear Medicine and Molecular Imaging used this CMS open comment period to submit requests that the agency lift the CED requirement for amyloid PET scans and authorize full coverage.

The vagueness of the draft NCD’s requirement for “clinically meaningful” cognitive and functional benefit troubled many, though previous articles have called on the fields’ clinicians, not CMS administrators, to work out such a definition (Liu et al., 2021; Messner et al., 2019). Rabinovici noted that the placebo groups in the aducanumab Phase 3 AD trials had declined by less than two points on the CDR-SB over 18 months. Thus, if a one- or two-point change on the CDR-SB was considered the bar for clinical meaningfulness, as suggested by the draft NCD, then a therapy would need to completely halt clinical progression to meet the standard.

As is, a sponsor cannot know if its antibody would be deemed clinically meaningful if it had met the primary endpoints in pivotal FDA-approved trials, the Duke-Margolis group said.

Others consider the concept itself too vague to yield a sensible bar. “A ‘clinical meaningfulness’ standard cannot possibly be set by academic researchers, the FDA, or CMS, because what is meaningful to one patient will be different to another,” argued Andrew Stern, a neurologist at Brigham and Women’s Hospital, Boston, in a comment to CMS. He thinks a given treatment’s cost-effectiveness would make a better benchmark.

In addition, researchers questioned the CMS’s requirement that CED trials reflect the diversity of the U.S. Alzheimer’s population. No current anti-amyloid antibody trial meets this standard, the Duke-Margolis group noted. Biogen recently announced plans to recruit a more diverse population for its FDA-required confirmatory trial, with at least 18 percent of participants either black or Latino (Endpoints news). This still falls short of the standard set forth in the NCD.

The draft NCD does not specify what level of statistical confidence would be sufficient to show that a therapy worked equally well in particular subgroups determined by race, ethnicity, comorbidities, or other factors. Most drugs currently being used were not specifically studied in every racial and ethnic group for whom they are being prescribed. The CMS draft proposal does not define which subgroups would need such evidence. “Would a therapy with stronger evidence of overall benefit remain uncovered until the diversity conditions are met?” McClellan and colleagues asked.

Some think CMS will adjust these requirements in its final decision. “Some of the detailed criticism of the CED protocol might have an impact, because rearranging eligibility criteria, outcomes, and limits of reimbursement is easily doable,” Lon Schneider at the University of Southern California, Los Angeles, wrote to Alzforum (full comment below). Last December, Schneider joined a call by several scientists for the FDA to pull its approval of aducanumab (RightCare Statement).

Treat Anti-Amyloid Therapy as a Class?

A uniformly criticized aspect of the draft CED is its application to all anti-amyloid antibodies. Rabinovici noted that different antibodies have distinct biological effects. “Given [this], it is likely that different drugs within this class will have different efficacy and safety profiles,” he wrote to CMS. Most AD researchers agree.

In comments to CMS, Rabinovici and many others requested that the agency not prejudge any antibody before its Phase 3 data are available. David Holtzman and colleagues at Washington University, St. Louis, went further. “Future monoclonal antibodies approved by the FDA that show significant clinical benefit should not require CED, but rather enjoy full Medicare coverage,” they wrote in their CMS comment.

Pivotal trials for gantenerumab and lecanemab are expected to read out this year, and donanemab next year.

Unsurprisingly, pharma companies developing such drugs, including Eisai, Roche/Genentech, Eli Lilly, and Acumen Pharmaceuticals, all asked that the CMS judge their drug on its own data. Michael Irizarry at Eisai called the proposed CED overly restrictive. “We stand … in strong opposition to a CED for this class of drugs,” he wrote to Alzforum.

The CMS proposal has prompted Lilly to pump the brakes on its accelerated approval for donanemab. Lilly moved its submission date from the first quarter of 2022 to near the end of the year, with an eye toward bolstering the application with more data on cognitive and clinical benefit (Endpoints news). Roche decided in December 2021, before the CMS proposal, to forgo accelerated approval altogether, and await results of its Phase 3 gantenerumab trials later this year in hopes of applying for full approval. “[This decision] was well-received by the Alzheimer’s community, including clinical trial investigators and advocacy partners,” Rachelle Doody wrote to Alzforum.

Neurologists in general seem to agree with this. In a survey of 209 neurologists conducted by the Global Alzheimer’s Platform (GAP) Foundation, 83 percent thought results from one drug should not determine the fate of an entire class. Some AD researchers think the scientific consensus on this issue may sway CMS. “This could be one area with a relatively high probability that CMS will modify their initial proposed decision,” Siemers wrote to Alzforum. In the end, a loud, messy controversy simply boils down to whether a given medication truly helps patients.

Meanwhile, pressure on Biogen continues to rise. Both the U.S. Federal Trade Commission and the Securities and Exchange Commission earlier this month opened investigations into aducanumab’s approval process, joining investigations already underway by the U.S. House, FDA, and HHS Inspector General (Endpoints news; Aug 2021 news). Congress recently used an unrelated hearing to question Patrizia Cavazzoni, who leads the FDA’s drug evaluation center, on the topic (STAT news). To boot, a Wall Street analyst uncovered three more deaths of people taking aducanumab from the FDA’s Adverse Event Reporting System. It is unclear if any of them are linked to the drug (Endpoints news). These are in addition to the death in Canada reported last fall.—Madolyn Bowman Rogers