By The Numbers - part the twenty fourth

Some oddities in the numbers from last night's lecanemab show. If that means nothing to you: lecanemab is a new monoclonal antibody (MAB) said to remove Aβ from the brains of Alzheimer's patients. The 'amyloid hypothesis' asserts that Aβ accumulation in the brain is the cause of Alzheimer's. That assertion has never been proven, alas. Some in the field assert quite the opposite: that Aβ is a reaction by the immune system to the actual driver (to date unknown) of Alzheimer's.

Aduhelm was, with great controversy, approved by FDA (reportedly over the objections of its scientists and overtly by outside reviewers) last year. It is also a MAB, and is alleged to work by clearing Aβ from the brain. IIRC, Aduhelm was the first anti-Aβ (aka, disease modifying compound) to be approved. All previous therapies have been symptom reduction.

Now comes lecanemab. Two, to my reading, contradictory statements appear in the report.
First:

The researchers found that participants in both groups had a "clinical dementia rating" or CDR-SB score of about 3.2 at the start of the trial. Such a score is consistent with early Alzheimer's disease, with a higher number associated with more cognitive impairment. By 18 months, the CDR-SB score went up 1.21 points in the lecanemab group, compared with 1.66 in the placebo group.

That difference is characterized in the report as a 27% improvement with lecanemab (.45/1.66). Where things get weird is
Second:

Lecanemab, a monoclonal antibody, works by binding to amyloid beta, a hallmark of the degenerative brain disorder. At the start of the study, the participants' average amyloid level was 77.92 centiloids in the lecanemab group and 75.03 centiloids in the placebo group. By 18 months, the average amyloid level dropped 55.48 centiloids in the lecanemab group and went up 3.64 centiloids in the placebo group, the researchers found.

So, nearly 20 times greater scavenging of Aβ by lecanemab leads to a 27% improvement in decline? Color me skeptical. It would seem logical that if Aβ is driving Alzheimer's, then such a massive reduction in it should lead to a hell of a lot more benefit. Time will tell if this abyss is made a big deal of by the medical and pharma fraternity.