Alzheimer’s treatment is starting earlier and earlier - critical summary review

What Alzheimer's is, and what treating it meant until now

Dementia is a syndrome that progressively erodes memory, thinking, behaviour and the ability to carry out everyday tasks. Alzheimer's disease is its most common form, responsible for something like 60% to 70% of cases, and it is not considered a normal consequence of ageing.

The scale is what makes it hard to ignore. More than 55 million people were living with some form of dementia worldwide in 2020, according to Alzheimer's Disease International, a figure close to the entire population of Italy, which today has around 59 million inhabitants. The projection is 78 million by 2030 and 139 million by 2050.

The early signs usually involve recent memory, the neurosurgeon Luan Messias Magalhães told g1: repeating the same question, struggling with things once done easily, losing words, shifts in mood. What matters, he said, is how much this interferes with routine. Occasionally forgetting where you left a key happens to anyone. Not knowing what the key is for, getting lost on a familiar route, or no longer being able to manage bills you used to manage is another matter.

For decades, the available medicines did nothing about that progression. Tacrine, donepezil, galantamine, rivastigmine and memantine were aimed essentially at symptoms, the geriatrician Lindsey Nakakogue, a medicine professor at PUCPR and scientific director of the Brazilian federation of Alzheimer's associations, told Folha de S.Paulo, the São Paulo daily. Donepezil is distributed by SUS, Brazil's public health system.

That is the backdrop against which this week's news arrives.

A new class of drug, and why the diagnosis has to come early

The new medicines are monoclonal antibodies designed to bind to the beta-amyloid protein and reduce the plaques it forms in the brain, which are associated with the neurodegenerative process. Two have reached patients: lecanemab, from the Japanese company Eisai in partnership with Biogen, and donanemab, from Eli Lilly.

They are not a cure, they do not reverse damage already done, and they are not for everyone. Approvals and trials concentrate on the early symptomatic stages, mild cognitive impairment or mild dementia, with confirmed amyloid pathology, and treatment requires imaging follow-up because of the risk of brain changes when amyloid is removed, the neurologist Marcelo Zalli wrote for Jovem Pan. Reported adverse effects include swelling and small bleeds in areas of the brain.

That restriction is what pushed diagnosis to the centre. Biomarkers in cerebrospinal fluid, amyloid PET scans and, increasingly, blood tests now allow doctors to look for the biology of the disease before dementia sets in. Recognising Alzheimer's only when cognitive loss is advanced, Zalli argued, means losing part of the window in which these therapies can even be considered.

Out of the studies: a private clinic in Sweden, a court injunction in Brazil

On September 22, the Swedish company BioArctic, which invented lecanemab, announced that a first patient in Sweden had begun treatment outside a clinical study, at a private clinic. Lecanemab was approved in the European Union in April 2025, and it is now cleared in 53 countries. The Swedish detail matters: in April 2026, the country's New Therapies Council, which advises on what the public system pays for, issued a negative recommendation on including the drug in publicly funded care. The clinic in Stockholm has been working to offer it anyway.

In Brazil, donanemab was approved by Anvisa, the health regulator, in April 2025 for mild cognitive impairment or mild dementia in early symptomatic Alzheimer's, given as a monthly intravenous infusion. An international phase 3 study with 1,736 patients found less clinical progression among those who received it than among those on placebo. Lecanemab was approved in Brazil last December.

Edna Barbosa, 72, diagnosed at the end of 2022, was scheduled for her fourth donanemab infusion on World Alzheimer's Day, September 21. Her health plan initially refused to cover it. Her husband went to court and obtained an injunction; the first two infusions came out of their own pocket, at a reported cost of about 5,000 reais per vial. "I just don't want to go backwards," she told Folha.

The catch: "small, at best", and 24,000 reais a month

A review published in May in the Cochrane Library, which pools clinical trials, pulled together 17 studies of drugs in this class and found no clinically significant effect. The drugs do reduce amyloid plaques, the analysis concluded, but without relevant impact on patients' ability to carry out daily activities. It described the effects as "small, at best".

Nakakogue is following four patients on donanemab and says all are doing well so far. Trial data, she told Folha, indicate that a share of treated patients do not move to the next stage of the disease for a certain period, though how long that holds is unknown.

Then there is the price. A month of lecanemab, with infusions every 15 days, costs around 8,000 reais before fees and taxes, reaching 11,000 reais with the 18% rate applied in most Brazilian states. A vial of donanemab can run to 6,000 reais, and after the ramp-up phase a patient needs four a month, about 24,000 reais. Blood biomarker tests are already available in Brazil, but are not part of SUS.

The race to know earlier: blood, genes and a new layer in DNA

Two studies published in the same stretch show where diagnosis is heading. In The Lancet Neurology, researchers in the United States combined a blood test measuring the p-tau217 biomarker with a genetic test looking at variants of APOE, the gene most strongly implicated in the disease, to flag people at higher risk of cognitive impairment before symptoms. The method was validated in almost 9,000 study participants followed for up to four years, and still needs more evidence before being rolled out. Blood tests alone, studies indicate, are not precise enough on their own.

Eduardo Ferriolli, a professor and researcher at USP, told Folha there is no reason to rush to laboratories: biological alterations do not necessarily mean a person has dementia, and the currently approved treatments are not indicated for people who have only those alterations without cognitive impairment.

In Science, researchers from Carnegie Mellon, Pittsburgh and Washington reported changes in the three-dimensional architecture of the genome in brain cells affected by Alzheimer's, using post-mortem samples of prefrontal cortex. The study does not show these changes cause the disease and presents no treatment. The next step, the authors say, is to find out which of those regulatory regions play a direct role, and whether any could become a target.

The other side of the advance: cancelled grants in the United States

In August, the Trump administration terminated a 15-year research project in two low-income, majority-Black neighbourhoods of Pittsburgh that had evolved into a study of cognitive health, the Guardian reported. Researchers had just finished the second round of cognitive testing on roughly 600 participants and were preparing the expert review needed to interpret the results. Given 90 days to shut down, more than five months earlier than planned, they no longer have the time or the money to do it. Participants were promised those results when they consented.

The National Institutes of Health said the grant "no longer effectuates the program goals or agency priorities" and that its self-reported measures lacked "concrete, objective, and measurable variables". Over the past two years, more than 5,700 NIH medical research grants have been cancelled, frozen or disrupted, according to the tracker Grant Witness, with more than 4,500 since restored. The University of Pittsburgh is appealing, and researchers there are named plaintiffs in a class action against the NIH.

The gap this touches is specific. Black Americans have nearly double the risk of developing Alzheimer's dementia, according to the Alzheimer's Association, and are far less likely to be included in research: of more than 12,000 brain-imaging studies on the disease, only around 700 report race at all, and those that do average more than 87% white participants. A second cancelled Pittsburgh study, on brain scans and blood biomarkers, had contributed to publications showing that Alzheimer's blood tests can be thrown off by comorbidities such as kidney and vascular disease. Its samples are now at risk of destruction.

What to do with this information

If someone close to you has memory lapses that are interfering with their routine, the moment to seek neurological evaluation and biomarker testing is now, not later. It is the early stage, confirmed by evidence of amyloid, that determines whether lecanemab or donanemab can even be discussed, and that window is narrow. A diagnosis normally starts with a detailed conversation between doctor, patient and family about when symptoms began and how they have progressed, followed by memory testing and blood and imaging exams that also rule out other causes, including vitamin deficiency and depression.

From here, three things are worth following. Whether health plans and SUS start covering these drugs and tests, which today are paid for by court injunction or out of pocket. Whether the combined blood and genetic tests move from validation to clinical use, given that the researchers themselves say more evidence is needed. And whether the funding cuts in the United States reach the early-diagnosis research that eventually arrives in Brazil. Those decisions, more than the next finding in a laboratory, are what will determine who gets access to what already exists.