How to Read an Alzheimer Drug Trial: CDR-SB and What Counts as Meaningful

A modern Alzheimer drug trial usually rises or falls on a single number: the gap between the treated group and the placebo group on a scale called the CDR-SB. When lecanemab reported a 0.45-point difference over 18 months, that number was statistically firm and became the headline. Whether 0.45 points is meaningful to a person living with the disease is a separate question, and it is one the field has not fully settled. Learning to hold those two ideas apart is the whole skill of reading these trials.

What the CDR-SB actually measures

The Clinical Dementia Rating Sum of Boxes (CDR-SB) is a clinician-scored instrument. A trained rater interviews the patient and a caregiver, then rates six domains: memory, orientation, judgment and problem-solving, community affairs, home and hobbies, and personal care. Each domain is scored from 0 (normal) to 3 (severe), and the six scores are added, so the total runs from 0 to 18. Higher is worse.

Because it blends thinking and daily function into one figure, the CDR-SB has become the workhorse endpoint for early Alzheimer trials. It captures the thing patients and families care about, which is not a lab value but whether someone can still manage money, keep up a hobby, or dress independently. Untreated, people with early Alzheimer disease tend to accumulate roughly one to two points on this scale per year. That baseline drift is what any drug is measured against.

The 0.45-point result, read carefully

In the CLARITY-AD trial, participants on placebo worsened by an adjusted mean of 1.66 points over 18 months, while those on lecanemab worsened by 1.21 points. The difference was 0.45 points, with a P value below 0.001, as reported in the New England Journal of Medicine. Commentators translated the 27 percent slowing of decline into an estimate of roughly five to six months of delayed progression over the trial period.

Two things are true at once here. The result is statistically robust: it is very unlikely to be a chance finding, and the confidence interval sits well away from zero. But statistical firmness only tells you the effect is real, not that it is large. A big enough trial can detect a genuine but tiny difference with great confidence. That is exactly the tension a careful reader has to sit with.

What "clinically meaningful" is supposed to mean

The concept the field reaches for is the minimal clinically important difference, or MCID: the smallest change that a patient or clinician would actually notice and value. The trouble is that no single, agreed MCID for the CDR-SB exists. A frequently cited benchmark treats about a 1-point change over a year as the floor for a difference to be clinically noticeable. Other work has proposed thresholds around 0.98 points for people at the mild-cognitive-impairment stage and higher for mild dementia. A review in Brain Communications is blunt that there is no standardized approach to defining meaningfulness in dementia trials, and survey estimates for related scales scatter widely.

Set the 0.45-point result against these benchmarks and you can see why the debate is real. An analysis in Scientific Reports pooled anti-amyloid trials and found lecanemab's effect on the CDR-SB corresponded to a standardized mean difference of about -0.19, a small effect size that fell below common thresholds for clinically important change. Meanwhile, an analysis in Alzheimer's Research & Therapy argued the group-level difference was still likely to be tangible, while conceding that validated thresholds simply have not been established. Neither camp is misreading the arithmetic. They are applying different, defensible yardsticks to the same number.

Why group averages hide individual stories

A single mean difference describes the average gap between two groups. It does not tell you that every treated person did 0.45 points better. Within any trial, some participants decline fast and others barely move, on both drug and placebo. A group-level slowing can coexist with many individuals who notice nothing and a few who may benefit more. This is why a difference below a proposed MCID does not automatically mean "no one benefits," and why a difference above it does not guarantee a given patient will.

It also explains why context matters when comparing agents. The Brain Communications review notes that donanemab reported a larger relative slowing on the CDR-SB in one baseline subgroup, and modeling exercises have projected multi-year delays in reaching later disease stages. Those projections are models built on trial data, not measured outcomes, and they carry their own assumptions. Read them as hypotheses about the future, not as observations.

A short checklist for the next headline

When a new Alzheimer readout crosses your feed, a few questions cut through most of the noise. First, what was the absolute difference on the CDR-SB, in points, rather than the percentage slowing alone? A percentage can look dramatic while resting on a fraction of a point. Second, how does that difference compare to any proposed MCID, and does the trial acknowledge that the threshold is contested? Third, is the benefit weighed against harms such as amyloid-related imaging abnormalities and the monitoring they require, a balance both cited reviews stress? Fourth, are long-term claims based on measured data or on simulation? Holding those four questions in mind turns a headline number back into something you can actually judge.

This article is educational and is not medical advice; decisions about any specific treatment belong with a qualified clinician who knows the individual.