In exchange for the financial costs and risk of side effects, we expect prescription drugs to do something useful. It may be reducing the likelihood of death, eliminating an infection, or improving the quality of our life in some way. But the expected benefits should be understood and they should be meaningful. Depending on the condition being treated, measuring “clinically meaningful” endpoints can be time consuming and difficult, but they are the only way we can determine whether a drug is actually delivering a real benefit – or not.
Sometimes researchers choose to measure “surrogates”, instead of clinically meaningful endpoints. Surrogates are not themselves clinically meaningful or relevant to patients, but are assumed to be a proxy or substitute for the real thing. While some surrogates are good predictors of relevant endpoints, some have proven over time to be useless. A cautionary tale comes from the Cardiac Arrhythmia Suppression Trial which studied drugs that did a good job at improving a surrogate marker (premature heart beats) but were eventually found to increase the risk of death, not reduce it. Another example might be cancer drugs that improve a surrogate such as “progression free survival” which is the time until the cancer reappears based on imaging or laboratory measures, but subsequently are found to have no effect on when you actually die (overall survival). That is, they improved the surrogate, but not a measure that is relevant to cancer patients. The main reason to measure surrogates is because of convenience. If you are studying a drug or other intervention when the relevant endpoint can take years to appear, a surrogate, if it correlates well, can reduce the time needed to complete a clinical trial. But it’s risky.
Given the mixed history with surrogates, there is tremendous hesitation among science-based health professionals to accept surrogate markers as relevant when those measures have not themselves been proven to correlate with endpoints that are actually meaningful to patients and their families. Whether or not regulators like the FDA should accept surrogates in clinical trials as evidence of efficacy is a point of controversy and debate. Regulators must balance what may be promising preliminary evidence, and the potential for a drug to offer some benefit, against the possibility that the drug is actually ineffective or even harmful.
All of this gives some context for reaction to the FDA’s approval this week of aducanumab, the first drug in decades approved to treat Alzheimer’s disease. Aducanumab failed in clinical trials – yet it was approved for sale. Why?
The challenge of Alzheimer’s disease
Alzheimer’s disease (AD) is a neurological disease that is the most common cause of dementia. AD tends to be a disease of the elderly that commonly presents initially as memory impairment (particularly of recent events) and progresses to include greater deficits including inability to problem solve and impaired judgement. Overtime, AD can include psychiatric and other behavioural symptoms. Progressive disease is inevitable, but the rate of decline can range from a few years to decades. AD is common, and affects an estimated 6 million Americans, with that prevalence expected to increase based on population trends.
How AD develops is not known, but it produces changes in the brain that are destructive, progressive, and irreversible. A common feature of AD is the accumulation of protein called “beta amyloid” in the form of plaques that are thought to interfere with signalling and eventually cause cell death. Another feature is called “tau tangles,” where tau protein fibers twist and collapse, which are also thought to cause cell death.
Despite decades of research there has been little progress in developing medications or therapies that can improve symptoms or slow the progression of the disease. A class of drugs called cholinesterase inhibitors provide modest benefit for some of the symptoms of the disease, but do not affect its progressive nature. Memantine, another drug, works differently from the cholinesterase inhibitors and may provide some benefit in moderate AD. However, there is no treatment that has been shown to have any meaningful effect on the disease itself.
The unimpressive history of aducanumab
For over twenty years, researchers have pursued treatments against beta amyloid after trials in mice showed promising results. No treatments have ever shown benefit in phase 3 clinical trials, however. Biogen has been studying aducanumab, an antibody which targets beta amyloid, for years. Two large Phase 3 clinical trials called ENGAGE and EMERGE were established to evaluate the efficacy of the drug in patients with mild AD. These were 18 month trials that measured the effect of aducanumab in early AD. Patients were randomized 1:1:1 to low-dose aducanumab, high-dose aducanumab, or placebo, and the drug was given as an intravenous infusion every 4 weeks. The primary endpoint measured was the Clinical Dementia Rating Scale Sum of Boxes test (CDR-SB), which can assess the severity of Alzheimer’s dementia. In March 2019 both ENGAGE and EMERGE were discontinued after an independent data safety monitoring board concluded that the trials were unlikely to meet their primary endpoint. That is, in the language of clinical trials, they were discontinued “for futility”. Given no drug with a similar mechanism of action had ever been shown to be successful, this was not surprising. However, in analyses done after the trials were stopped (i.e., a “post-hoc” analysis), the manufacturer suddenly claimed that the drug was in fact effective, and it intended to pursue regulatory approval. (Post-hoc analysis can be important to generate hypotheses but are the statistical equivalent of shooting at a wall and painting the bullseye afterwards.) In short, the manufacturer claimed that after the futility analysis, additional data was made available that provided sufficient evidence of efficacy to merit moving forward with seeking approval. Notably, Biogen did not launch another trial to actually confirm their post-trial hypothesis. It filed for FDA approval.
It should be noted that the drug is not free of harms. It can cause swelling of the brain called amyloid-related imaging abnormalities (ARIA), common to this category of drugs, which can cause symptoms like headache.
Aducanumab was reviewed in November 2020 at an FDA advisory committee meeting. The briefing documents are here if you wish to peruse them. Among those documents included endorsements from FDA staff members supporting the manufacturer’s contention that the drug is effective. The report also contained comments from other FDA staff members criticizing the efficacy statements and the manufacturer’s reliance on post-hoc analyses, which were (justifiably) described as biased. Ultimately the panel votes on the evidence were strongly negative. While the FDA is not required to follow its advisory committee recommendations, these recommendations are typically important.
On June 8, 2021 the FDA granted accelerated approval for aducanumab, now branded Aduhelm, while requiring Biogen to do a confirmatory clinical trial over the next nine years. Incredibly, it did not even limit approval to patients similar to those in the clinical trial, who had mild disease and confirmed amyloid plaques. Instead, it put no restrictions on its use. There is no evidence whatsoever to suggest that aducanumab is effective in patients with advanced Alzheimer’s disease, but the drug can now be used in those patients.
In the meantime, Biogen can start marketing this product immediately, while clinicians, patients, and families are left wondering if the drug actually works at all. Notably the FDA approval states that the basis of approval is not a clinically meaningful endpoint – it was ultimately approved based on a surrogate :
“approval is based on a surrogate or intermediate clinical endpoint (in this case reduction of amyloid plaque in the brain).” Additionally, “drug companies are required to conduct post-approval studies to verify the anticipated clinical benefit. These studies are known as phase 4 confirmatory trials. If the confirmatory trial does not verify the drug’s anticipated clinical benefit, FDA has regulatory procedures in place that could lead to removing the drug from the market.”
So for the next several years, this drug is expected to be used in millions of Alzheimer’s patients, generating tens of billions of dollars in revenue (STAT says it may be the most lucrative drug product in history), without any clear evidence that the drug is actually doing anything beneficial at all.
A travesty of a mockery of a sham of an approval
FDA drug approval is a worldwide gold-standard that establishes that a drug is effective and has an acceptable safety profile. With aducanumab, the FDA has approved a drug that failed its Phase 3 clinical trials, and its own advisory committee recommended against approval. And the basis of approval was a surrogate endpoint that has not been shown to correlate with any clinically meaningful outcome for patients. With this approval, it may have done serious damage to its own reputation as a regulator.
Alzheimer’s is a devastating disease and there will be immediate demand for a drug which offers even a glimmer of hope. But is hope worth the $56,000 per patient, per year, price tag? Is the risk of brain swelling worth $56,000 per year, if the drug provides no benefit? A drug that meaningfully affects the course of Alzheimer’s disease justifiably should be a multi-billion dollar drug. But Biogen hasn’t shown this. Maybe we will find out in 2030, if Biogen completes the required trial, and after billions of dollars have been spent on treatments. Alzheimer’s patients and their families deserve better.