A Case of False Hope from FDA

Alzheimer’s disease, the most common form of neurodegenerative disease among the elderly, is a devastating illness with a very poor prognosis. A person diagnosed with Alzheimer’s disease (AD) at age 65 has a median survival expectancy of 8.3 years. More than six million Americans are living with AD now and dementia is the cause of death in about a third of seniors.

         A few drugs have been approved that have very modest effects on slowing the relentless loss of memory and cognitive abilities that AD causes. These drugs, with brand names like Aricept and Namenda, do not cure AD. They may slow cognitive decline, but their positive effects are often so small as to go unnoticed by caretakers. There is a palpable feeling of desperation for an effective treatment that one senses among patients with AD and their loved ones.

         It was that feeling of desperation that may be behind the curious decision of the U.S. Food and Drug Administration (FDA) to approve a new medication, aducanumab (brand name: Aduhelm), for the treatment of AD even though it is entirely unclear it works. Despite an FDA advisory board voting 10 – 0 against approving Aduhelm, physicians can now prescribe monthly intravenous infusions of Aduhelm at a whopping cost of $56,000 a year.

Antibodies Against Plaques

         In fact, aducanumab is an approach to treating AD that has failed on multiple previous attempts at drug development. Alois Alzheimer noted in the seminal paper he presented in 1906 that a patient who had died with “presenile dementia” had two distinctive features in her brain when he performed an autopsy. One of these is called amyloid plaques and the other neurofibrillary tangles, which together remain the hallmarks of the pathological diagnosis of AD to this day. Understandably, scientists have focused a great deal of attention on the plaques and tangles that can be seen with an ordinary light microscope in the brain of someone who has died from AD. The amyloid plaques in particular attracted a great deal of attention. Scientists learned that they are made of a protein called beta-amyloid that in turn comes from the breakdown of another protein, called amyloid precursor protein. Amyloid plaques appear to form early in the course of AD, and it was reasonable to speculate that they could even be the driving force behind the illness.

Deposits between neurons of plaques composed of the toxic protein beta-amyloid are a diagnostic feature of Alzheimer’s disease (source: Shutterstock).

         A novel strategy was devised to try to rid the brain of amyloid plaques: give the patient antibodies manufactured to attack and destroy the protein beta-amyloid. When Critica President Jack Gorman first attended a lecture in which that strategy was described, he was among the first to wonder if the immune system would attack the anti-amyloid antibodies themselves; after all, even though beta-amyloid is a toxic protein it is still a naturally occurring protein made by the patient’s brain, he pointed out at the time, and we don’t usually try to immunize people against their own proteins. Although not much was made of the concern back then, antibodies directed against anti-amyloid antibodies given in an attempt to treat AD have been a significant safety challenge. We’ll come back to this point a bit later.

         It is very hard to study putative drugs for AD because we do not have very good animal models for the disease. No other species besides humans develops AD naturally, although it is possible that this is because no other species lives as long as we do. Dolphins and chimpanzees might get AD-like pathology if they lived long enough. It is possible to insert an abnormal human gene into mice that causes the production of the beta-amyloid protein. These transgenic mice develop memory problems, and it turns out that anti-amyloid antibodies rid the experimental mouse brain of amyloid plaques and restore memory function to them. So that looks like a reasonable animal model to test the hypothesis that giving anti-amyloid antibodies can have a positive impact on the course of Alzheimer’s dementia.

         Except that mice are not humans and the mouse models for AD are all laboratory creations, not examples of a naturally occurring disease process. Drugs that rid the transgenic mouse brain of plaques may even do so in people with Alzheimer’s disease, but that doesn’t mean they will necessarily have a positive impact on the human disease. In fact, every attempt to develop a drug that attacks amyloid plaques until recently has failed. So convincing are these failures, that it has led many scientists who study AD to doubt whether amyloid plaques are as central to the disease as once thought. Rather than causing AD, the plaques may form as a reaction to some other disease process, for example. Amyloid plaques are an attractive subject to study because they are so easily found in the brains of people with AD, but some scientists believe it is time to move on from the beta-amyloid hypothesis for Alzheimer’s diseases.

Intervening Early

         A major advance in studying AD was the introduction of positron emission tomography (PET) brain imaging of amyloid plaques. A radiotracer is injected into the patient and PET scan detectors pick up where it attaches in the brain, thus identifying plaques. Now scientists can identify patients with a significant amount of amyloid plaque burden at very early stages of the illness. They can also see if an anti-amyloid drug actually gets rid of the plaques in the living human subjects.

         This is what researchers studying the drug aducanumab, which is made by the Biogen company, and another anti-amyloid antibody drug for which there was some evidence of efficacy published last may in the New England Journal of Medicine,  donanemab, which is made by the Eli Lilly Company, did. In those studies, patients first underwent beta-amyloid PET scanning to ensure they had a significant accumulation of amyloid plaques, then received the anti-amyloid antibody drugs. For aducanumab (Aduhelm), this is where the controversy begins because in 2019 Biogen and a Japanese pharmaceutical company it is working on Aduhelm with announced that they were stopping two clinical trials because interim analysis of the data that had been so far collected indicated that the drug was not working any better than placebo. Then, in a somewhat startling reversal, the companies later said that on the basis of a new analysis of the data from the stopped clinical trials they had detected an effect of Aduhelm in slowing cognitive decline in patients who received the higher of two doses of the drug compared to placebo. The difference was a 22% reduction in cognitive decline.

         One caution has been raised about abandoning the beta-amyloid hypothesis: plaques probably form very early in the course of the disease, before it is clinically apparent that someone has symptoms. It could be that by that time it is too late to have a meaningful impact on the disease. What would be ideal would be to know that someone has begun to accumulate amyloid plaques before they start to lose memory and other cognitive functions and to be able to intervene at that point.

         FDA usually requires evidence that a new drug is effective in at least two large, also called phase three, clinical trials. Before beginning one of these trials, the sponsor (in this case the drug companies involved) must state their intended main outcome measure. This is done to prevent cherry-picking the data; analyzing data multiple times with multiple different outcome measures can lead, for statistical reasons, to statistically significant findings that are actually flukes. Most experts thought the FDA would require Biogen to perform at least one more phase three clinical trial with the higher dose to prove the drug works better than placebo. The FDA did note that Aduhelm was associated with a decrease in plaques. That is interesting, but it is still unclear if reducing the number of plaques in someone’s brain actually results in an improvement in cognitive function. 

Alzheimer’s disease is a devastating neurodegenerative illness for which there is no cure. It robs its victims of their memory and other cognitive functions; available drugs offer very modest slowing of cognitive decline (image: Shutterstock).

         But it didn’t. Despite the overwhelmingly negative opinion of its own external advisors, FDA granted approval for Aduhelm even though there is very little reason to be confident at this point it actually works. We have one study in which a reanalysis of the data suggests it might be effective and another study in which it did not separate from placebo. That is very thin evidence upon which to hang a new drug approval.

One might say, as the FDA probably considered, that anything that might have the slightest chance of working in a disease for which the outcome is always devastating and there is no cure should be made available to patients. But there are at least three objections to that argument. First, as has been the case with all the attempts at designing anti-amyloid antibody drugs to date, Aduhelm has adverse side effects that are probably at least in part due to anti-drug antibodies mentioned earlier that the brain itself produces. About one-third of patients in the clinical trials developed brain swelling and between 17 and 19% had small brain bleeds. 

Second, as mentioned earlier, the drug is extremely expensive. Because most people with Alzheimer’s disease are over age 65, Medicare will be asked to pay for Aduhelm, thus adding another significant financial burden to taxpayer supported healthcare.

Third, the argument that a lesser standard of evidence can be used for a drug that attempts to treat a very serious illness for which there are very limited existing treatment options may seem humane, but it actually undercuts our entire system of rational drug development. Not only should a drug work in order to garner FDA approval, its benefits should outweigh risks. In the case of Aduhelm we have the barest evidence of benefit and significant evidence of risk. Three members of the advisory board that voted against approval of Aduhelm have resigned their positions in protest.

         Our system of drug approvals has worked remarkably well. It is rare that the FDA approves a drug that subsequently proved harmful; instances in which that has happened, like Vioxx, come to mind so quickly because they are unusual. Patient advocacy groups and pharmaceutical companies complain all the time that the drug approval process is too slow and thus deprives people of the chance to take advantage of treatment breakthroughs, but it simply takes a long time to put together enough evidence to convince experts that a new drug is both safe and effective. As the editor of the Washington Post recently wrote in its discussion of Aduhelm “In the U.S. health-care system, patients’ interests chronically come behind the interests of big business.”

         It is a fine thing to want to give hope to people suffering with Alzheimer’s disease and to their caregivers. What we don’t want to do, however, is undermine our regulatory procedures and give people false hope. The FDA is requiring that Biogen perform another large-scale study of Aduhelm and perhaps that will yield the so far elusive evidence that the drug works. As things stand now, however, we believe the FDA acted impulsively and unscientifically to approve a drug for which there is minimal evidence of benefit and significant evidence of risk. Healthcare dollars are precious and should never be wasted on ineffective treatments.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s