← All articles
An empty hospital PET scanner room seen from the doorway, circular gantry, cool blue-grey light

A tau antibody missed its main Alzheimer's score

For anyone following Alzheimer's research: what a phase 2 trial found on the score clinicians use to track daily function, and which questions that result leaves open.

Lonevi

The number in this trial is the Clinical Dementia Rating Sum of Boxes, usually shortened to CDR-SB. A higher total means more difficulty with memory and with ordinary tasks such as handling money, chores, and getting around. In a phase 2 study published in The Lancet Neurology and indexed on PubMed on 10 November 2026, that total rose by about two points over 80 weeks in the antibody groups and in the placebo group.

This page is information for readers. It does not replace a consultation with a doctor.

Who took part, and what was compared?

UCB funded a double-blind, placebo-controlled trial, registered as NCT04867616 and marked complete. The registry lists primary completion on 24 May 2024 and study completion on 1 August 2025. Adults aged 50 to 80 years with prodromal-to-mild Alzheimer's disease were randomly assigned, one to one to one, to bepranemab at the higher of two intravenous doses, to the lower dose, or to placebo. The registry's short title describes the same trial as mild cognitive impairment or mild Alzheimer's disease. Assignment was stratified by diagnosis, by whether the person carried an apolipoprotein E ε4 allele, and by use of symptomatic treatment.

Recruitment ran across 101 outpatient sites in eight European countries, Canada, and the United States. Between 24 August 2021 and 25 October 2022, 466 people were randomly assigned at 85 of those sites: 154 to the lower dose, 155 to the higher dose, and 157 to placebo. Median age was 72 years (interquartile range 66 to 76). Of the 466, 259 (56%) were women and 207 (44%) were men.

Doses were given by vein every four weeks through week 76. The prespecified main comparison was the change in CDR-SB from baseline to week 80. The efficacy figures below come from the full analysis set: 460 people who received at least part of a dose and had at least one later efficacy measurement (152, 152, and 156). Missing values were not filled in.

Bepranemab is an antibody directed at the middle region of tau, the protein that forms tangles inside neurons in Alzheimer's disease. The investigators write that earlier tau antibodies, aimed mostly at the N-terminal end of the protein, had not slowed tau progression or cognitive decline. That sentence is their framing of prior work. The numbers in this article come from this trial.

How far apart were the groups on the main score?

The least-squares mean change in CDR-SB at week 80 was 2.15 points in the higher-dose group, 1.85 in the lower-dose group, and 1.81 on placebo. A rise is a worsening on this scale. Compared with placebo, the higher dose differed by 0.34 points (95% confidence interval −0.09 to 0.78, p = 0.12). The lower dose differed by 0.04 points (95% confidence interval −0.40 to 0.47, p = 0.87). Both 95% intervals include zero, so the trial does not show a separation from placebo on the main score. The point estimate for the higher dose sits slightly on the side of a larger rise, not a smaller one. The authors state that the primary endpoint was not met.

The paper also prints an 80% confidence interval for the higher-dose difference, 0.06 to 0.63, which lies entirely above zero. That interval is narrower by design. It is not the threshold the authors use when they say the primary endpoint was not met.

What was recorded about harms?

Among people who received at least part of a dose, treatment-emergent adverse events were reported in 139 (91%) on the lower dose, 140 (91%) on the higher dose, and 142 (90%) on placebo. The investigators report no treatment-related neuroinflammatory signal on clinical review or on brain MRI. Adverse events leading to death occurred in four people (1%) in the antibody groups combined and in six people (4%) on placebo. None of those deaths was considered related to the trial treatment.

Similar rates are not a finding about years of later use, and they say nothing about people who were outside this trial's age range or disease stage.

What did the investigators say is still untested?

They describe differences favouring bepranemab on a tau PET scan and on the 14-item cognitive subscale of the Alzheimer's Disease Assessment Scale as hypothesis-generating. They say those signals should be tested in new trials powered for the scan and stratified by the amount of tau at the start and by apolipoprotein E ε4 status. The abstract does not give the size of those differences. Without a number, a reader cannot weigh that signal. This trial also does not speak for moderate or severe Alzheimer's disease, for people younger than 50 or older than 80, or for prevention in people who do not already have prodromal or mild disease. It is not a regulatory approval.

A null result on CDR-SB does not show that tau is a useless target. It also does not show that this antibody slows day-to-day decline. Those are different claims, and this paper does not close either one.

What should you do with a result like this?

If a clinician mentions an anti-tau antibody, the useful question is which scale the decision rests on, and whether that scale is the one on which this trial did not separate the groups. Blood tests that estimate who may later develop dementia answer a different question. We wrote about one such cohort separately, in Blood tests for dementia caught fewer than 6 in 10 future cases. Nothing here is a reason to start, stop, or change a medicine on your own. Discuss the paper with a doctor who knows the person in front of them.

Sources

Barton ME, Byrnes W, van den Steen B, et al. Safety, tolerability, and efficacy of bepranemab in people living with prodromal-to-mild Alzheimer's disease: a randomised, double-blind, placebo-controlled, phase 2 trial. The Lancet Neurology. 2026. PMID 43008775. DOI 10.1016/S1474-4422(26)00357-100357-1).

ClinicalTrials.gov. NCT04867616, bepranemab (UCB0107) in prodromal to mild Alzheimer's disease. Sponsor: UCB Biopharma SRL. Status: completed. Registry record.

What this piece relies on

Cards

Articles in this section are educational and are not medical advice, a diagnosis, or a prescription. Consult a qualified professional before acting on anything you read here.

Read next