The Amyloid Hypothesis Is Crumbling, and Alzheimer's Research Must Reckon With the Fallout

For nearly four decades, the dominant theory of Alzheimer's disease has been deceptively elegant: a sticky protein called amyloid-beta accumulates in the brain, forms plaques, and triggers the neurological collapse that robs millions of people of their memories and identities. The hypothesis was compelling enough to organize an entire scientific field around it, attracting billions of dollars in funding and generating hundreds of clinical trials. The problem is that the drugs built on this premise have, with rare and modest exceptions, failed patients in trial after trial, and the scientific community is now grappling with what that failure actually means.

The amyloid cascade hypothesis was first formally articulated in 1992 by John Hardy and David Allsop, drawing on observations of patients with Down syndrome, who develop Alzheimer's pathology at unusually high rates, and on rare genetic mutations that cause early-onset disease. The logic was clean: if amyloid plaques are present in every Alzheimer's brain, and if certain genetic variants accelerate plaque formation, then clearing the plaques should slow or stop the disease. Pharmaceutical companies poured resources into that bet. Between 1998 and 2017, roughly 146 Alzheimer's drug trials failed, many of them targeting amyloid. The drugs could often clear the plaques. The patients did not get better.

The cracks in the hypothesis widened considerably in 2022, when Science magazine reported concerns about a foundational 2006 paper by Sylvain Lesné, which had been widely cited as key evidence for a specific form of amyloid called Aβ*56. Neuroscientist Matthew Schrag raised questions about image manipulation in the paper, and subsequent investigations by Vanderbilt University and the National Institutes of Health found the concerns credible enough to warrant serious scrutiny. The paper has since been retracted. While the amyloid hypothesis does not stand or fall on a single study, the episode exposed how much of the field's confidence rested on a surprisingly narrow evidentiary base, and how institutional incentives, funding structures, and academic prestige had reinforced that confidence long past the point where skepticism should have flourished.

This is not simply a story about scientific fraud or even scientific error. It is a story about how a paradigm, once entrenched, develops its own immune system. Researchers who questioned amyloid's centrality found it harder to secure grants. Journal editors and peer reviewers, themselves trained within the paradigm, were less receptive to heterodox findings. The feedback loop was self-reinforcing: amyloid research produced publications, publications attracted funding, funding produced more amyloid research. Alternative hypotheses, including those centered on tau protein tangles, neuroinflammation, mitochondrial dysfunction, and the role of the brain's glymphatic clearance system, were explored, but rarely with the same institutional momentum.

The picture is not entirely bleak. Two drugs, lecanemab and donanemab, have shown statistically significant slowing of cognitive decline in early-stage patients, lending partial vindication to the amyloid approach. But the effect sizes are modest, the side effects, including brain swelling and microbleeds, are real, and the treatments require early diagnosis at a stage when most patients are not yet in clinical care. The drugs work, a little, in a narrow window, for patients who can access expensive infusion therapies. That is a long way from a solution to a disease affecting more than 6.7 million Americans, a number projected to nearly double by 2060 as the population ages.

The more consequential second-order effect of this reckoning may be structural rather than pharmacological. If funding agencies, universities, and journals begin genuinely rewarding scientific heterodoxy, the next generation of Alzheimer's researchers might pursue the disease's complexity with fewer inherited assumptions. The glymphatic system, which clears metabolic waste from the brain during sleep, has attracted growing attention as a potential contributor to amyloid accumulation in the first place, suggesting that the protein may be a symptom of a deeper dysfunction rather than its root cause. That reframing, if it holds, would shift the therapeutic target from clearance to prevention, and from the neurology clinic to the sleep lab and the primary care office.

Alzheimer's research is not starting over. But it is, slowly and painfully, learning to question the map it has been using. The patients waiting for a treatment that works cannot afford for that lesson to take another four decades to absorb.