Statistical Flaw May Have Overstated Alzheimer's Drug Benefits in Key Trial
A new analysis finds a popular statistical technique inflated amyloid-cognition links 29-fold, casting doubt on some Alzheimer's drug findings.
Summary
Researchers are questioning a statistical method used to support amyloid-targeting Alzheimer's drugs. The technique, called quantile aggregation, groups trial participants into buckets and averages their data — a process that can make weak relationships look dramatically stronger. A JAMA Neurology study found this method inflated the apparent link between amyloid reduction and cognitive benefit by 29 times compared to individual-level analysis. The findings raise concerns about a secondary analysis of donanemab (Kisunla), an Eli Lilly drug, which used this approach to claim lower amyloid correlated with slower cognitive decline. Critics argue the method obscures variability and undermines the causal logic that randomized controlled trials are designed to provide.
Detailed Summary
A new statistical critique published in JAMA Neurology is raising uncomfortable questions about how the benefits of Alzheimer's drugs have been measured and communicated. The study targets a technique called quantile aggregation — a method that divides trial participants into groups, averages outcomes within each group, and then looks for patterns. On the surface it sounds reasonable, but researchers found it can dramatically distort results.
Led by Sarah Ackley, PhD, of Brown University School of Public Health, the team used both simulated data and publicly available trial data to test the method. In simulated trials, individual-level analysis showed only a weak association between post-treatment amyloid levels and cognitive change (R² = 0.03). Apply quantile aggregation, and that figure ballooned to R² = 0.87 — a near-perfect apparent relationship. That represents a 29-fold inflation in the strength of the association.
The real-world test used data from the A4 study, a phase III trial of solanezumab that showed no cognitive benefit. Individual-level analysis again showed a weak amyloid-cognition link (R² = 0.04). Quantile aggregation produced a near-perfect association. A drug that failed still looked effective under the flawed lens.
The concern is directly relevant to donanemab (Kisunla), an Eli Lilly drug approved for early Alzheimer's. A secondary analysis of the TRAILBLAZER-ALZ 2 trial used quantile aggregation to argue that lower post-treatment amyloid correlated with slower clinical progression. Ackley argues this analysis mixed placebo and treatment participants, erasing the randomization that makes causal conclusions possible.
For health-conscious readers, this matters because it introduces meaningful uncertainty into current Alzheimer's treatment narratives. Cognitive benefits of amyloid-targeting drugs have been modest in trials. This critique suggests some reported benefits may be statistical artifacts rather than true drug effects, warranting caution before strong clinical conclusions are drawn.
Key Findings
- Quantile aggregation inflated apparent amyloid-cognition associations 29-fold compared to individual-level analysis
- Even a failed Alzheimer's drug (solanezumab) appeared highly effective when this statistical method was applied
- The technique was used in a key secondary analysis supporting donanemab (Kisunla) approval
- Mixing placebo and treatment groups in the analysis undermines causal interpretation from randomized trials
- Real cognitive benefits of amyloid-targeting Alzheimer's drugs remain modest and short-term in current evidence
Methodology
This is a news report summarizing a JAMA Neurology research letter by Brown University researchers. The evidence basis includes simulated trial data and publicly available A4 study data. MedPage Today is a credible peer-reviewed medical news outlet; the underlying letter is peer-reviewed.
Study Limitations
The article was cut off before reporting the full A4 quantile aggregation R² result. The critique targets a secondary analysis, not the primary trial powering donanemab's approval. Eli Lilly's response or any rebuttal from original analysis authors was not included.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.