There is a particular kind of scientific argument that works by rearranging the evidence until the answer comes out the way you wanted it to. Take a pool of trials. Mix the ones that failed, the ones that were stopped for futility, and the ones that succeeded. Analyse the lot as if they were the same thing. Conclude that the field should move on. It is a tidy trick. It is also, increasingly, the house style in one corner of the Alzheimer’s debate.
The argument runs roughly as follows: the first generation of anti-amyloid treatments is a disappointment, therefore the amyloid hypothesis is a dead end, therefore science should pivot. Each step sounds reasonable in isolation. Stitched together they form something rather less reasonable.
What the evidence actually shows
The first generation of disease-modifying treatments does not cure Alzheimer’s disease. Nobody involved in developing, approving or prescribing them has ever claimed it does. What they do is slow the progression of the disease. And they do it in a population the earlier trials largely did not have: people confirmed by testing to have the biology (amyloid) the drug is designed to target.
For a long time, trials enrolled participants on the basis of clinical assessment — people who looked, on examination, to have Alzheimer’s. Some did. Some, it turned out, did not. When you test a drug designed to clear amyloid in a group of people who may or may not have amyloid pathology, you are testing two things at once: whether the drug works, and whether you have the right patients. Modern trials do not make that mistake. They confirm the biology, and they identify people earlier in the disease trajectory. In those trials, anti-amyloid treatments slow the progression of the disease. That is not a small thing. It is what a first generation looks like.
The argument that the field needs to move on from amyloid also misreads where the field is. Walk into any major conference in the last three years and the talk is of combination therapies: anti-amyloid, anti-tau, anti-inflammatory, necroptosis blockers. Look at the Alzheimer's pipeline and the diversity of mechanisms of action is striking. Look at early-stage venture capital and the point is underlined again. The field has not been stuck on amyloid. Perhaps what is needed now is for people to move on from talking about moving on from amyloid.
The word “modest”
Clinicians describing treatment effects have settled on the word “modest”. It is doing a lot of work. The word carries a subjective judgement dressed as a clinical observation — the judgement that the benefit is not much worth having.
One of my closest friends died of bowel cancer less than three months after diagnosis, six weeks short of Christmas. She wanted one last Christmas with her children. Six weeks would not have been modest. It would have been everything.
Clinicians should describe treatments accurately. Patients, and the people who love them, should decide what is worth it. The field is entitled to want treatments that work better, cost less, are easier to deliver, and carry fewer risks. Those are all fair things to want, and all things that are being worked on. None of them changes the core finding that treatments are slowing the disease trajectory. The question of how meaningful that is does not belong to the person writing the review.
The cost of closing the door early
There is a real debate to be had about cost, access, delivery and risk-benefit — a debate the WDC has convened, contributed to, and will keep pressing. That debate is live and serious, and it is happening across every health system currently weighing these treatments.
What I find harder to understand is the impulse to foreclose the scientific question before the science is done. Declaring a therapeutic avenue closed, on the strength of a methodology that pools the failures with the successes, is not caution. It is a different kind of overreach. Over-egging the pudding, one might say.
The four-minute mile
When Paavo Nurmi ran a 4:10 mile, a lot of people said no human would ever get below four minutes. Ten years later Hägg and Andersson brought it to 4:01, and still people said the barrier would hold. Ten years after that Roger Bannister did it. Over 1,700 people have done it since.
The useful question about anti-amyloid treatments is not whether the first generation is the endpoint. It is whether we are running toward a four-minute mile or a biological limit. Trials in earlier populations, trials using brain-shuttle technology, trials of new molecules — these are not a field giving up. They are a field asking the right question, which is how much further this can go.
We do not yet know the answer. That is why we run the trials.