Early Alzheimer's disease progressed with high variability and this may influence whether treatment effects truly are detected in clinical trials, a simulation study suggested.
Individual trajectories on commonly used cognitive measures -- the Mini-Mental State Examination (), the Clinical Dementia Rating Sum of Boxes (), and the Alzheimer's Disease Assessment Scale Cognitive () -- differed greatly among individuals with prodromal and mild Alzheimer's, even among seemingly homogenous groups, reported Roos Jutten, PhD, of Amsterdam UMC in the Netherlands, and co-authors.
As a result, the 95% range of group differences observed on cognitive measures over 18 months was broad, spanning from 0.35 points improvement to 0.35 points decline on the CDR-SB, for example, they wrote in .
"Alzheimer's disease patients show substantial variation in how fast or slow their cognitive functions decline over time," Jutten told app. "This heterogeneity in disease progression is a problem for clinical trials, which want to show that a drug or treatment can slow or halt cognitive decline."
"We investigated how natural variability in disease progression may lead to significant group differences that are not due to the therapy administered, but instead reflect patient differences in their rate of decline," she said.
"We also studied ways to reduce variability in rates of decline, by looking only at subgroups that had increased risk for fast progression, such as APOE4 carriers," Jutten continued. "An unexpected outcome of our study is it is even more difficult to find treatment effects in these high-risk groups, because they show more within-group variability in decline instead of less."
The findings highlight an important challenge of trial design, noted Rachel Buckley, PhD, of Massachusetts General Hospital in Boston, and David Knopman, MD, of Mayo Clinic in Rochester, Minnesota, in an .
"In the absence of a widely accepted therapeutic success threshold to serve as a starting point, what is a minimal therapeutic benefit that is clinically meaningful?" Buckley and Knopman asked.
"Clinical meaningfulness has been an elusive concept in the world of Alzheimer's disease therapeutics," they added. "The description by Jutten et al. of the 95% confidence interval for random variation in slope of several widely used outcome measures provides a rational estimate of a floor for what clinically meaningful therapeutic benefit should look like."
In their study, Jutten and co-authors identified 302 participants from the Alzheimer's Disease Neuroimaging Initiative () with an abnormal amyloid PET scan, diagnosis of mild cognitive impairment or dementia, baseline MMSE score of 24 or higher, global Clinical Dementia Rating score of 0.5, and at least one follow-up cognitive assessment. These were the baseline characteristics of the trials, they noted.
Mean age of the sample was 73, 44% were women, mean education was 16.1 years, and 69% were APOE4 carriers.
The researchers simulated a clinical trial by randomly assigning individuals to "placebo" and "treatment" groups (even though no treatment was actually given), using longitudinal cognitive data to estimate change from baseline over 18 months. This gave them the variability in effect sizes between a placebo and treatment group if heterogeneity in progression wasn't accounted for. They repeated this simulation 10,000 times to determine the 95% range of effect sizes.
They compared this to effect sizes that have been reported for recent trials of several anti-amyloid drugs with similar inclusion criteria and found that, with the exception of the aducanumab EMERGE trial, the treatment effects fell within the 95% boundaries -- in other words, within the range that could be expected if there were no treatment effect.
"This suggests that, even though within some trials, differences between placebo and treatment groups were statistically significant, the possibility cannot be excluded that those differences were actually due to oversampling of fast decliners in the placebo group or oversampling of slow decliners in the treated group," Jutten and co-authors wrote.
"We further showed that, when repeating our simulation for separate risk factors associated with disease progression, a positive APOE4 status and baseline abnormal total tau levels were associated with steeper cognitive decline at a group level, but also with greater variability in progression," they added. "This resulted in even broader ranges of effect sizes in these high-risk groups on all outcome measures (e.g., ±0.70 points for the CDR-SB in those with baseline abnormal tau)."
A run-in period that captures an individual's cognitive trajectory, using biomarkers as primary endpoints, or larger sample sizes may solve the problem, but these solutions have limitations.
"Overly large sample sizes invite the prospect of detecting therapeutic effects that are statistically significant but clinically trivial," Buckley and Knopman noted. "While designing a better outcome measure using criteria to exclude persons who are not likely to decline or using other strategies to minimize measurement error, therapeutics with larger effect sizes are what the field ultimately needs."
The study had several limitations, Jutten and colleagues acknowledged. The simulation was based on ADNI participants, a relatively highly educated sample with little racial and ethnic diversity. Some participants in ADNI had 6-month follow-up data, which may have led to less decline and less variability in change over time. Selection criteria and follow-up timeframe were based on the aducanumab EMERGE and ENGAGE studies, and there may be small differences with criteria used in other trials.
Disclosures
Alzheimer Center Amsterdam, Amsterdam UMC is supported by Alzheimer Nederland and Stichting VUmc funds.
Data collection and sharing for this project was funded by the Alzheimer's Disease Neuroimaging Initiative (ADNI), which receives support from the National Institutes of Health and Department of Defense.
The researchers reported no disclosures relevant to this manuscript.
Buckley reported no disclosures relevant to the manuscript. Knopman serves on a Data Safety Monitoring Board for the Dominantly Inherited Alzheimer Network study. He also serves on a Data Safety Monitoring Board for Biogen but receives no personal compensation. He is an investigator in clinical trials sponsored by Biogen, Lilly Pharmaceuticals, and the University of Southern California. He serves as a consultant for Roche, Samus Therapeutics, Third Rock, and Alzeca Biosciences but receives no personal compensation.
Primary Source
Neurology
Jutten RJ, et al "Finding treatment effects in Alzheimer trials in the face of disease progression heterogeneity" Neurology 2021; DOI: 10.1212/WNL.0000000000012022.
Secondary Source
Neurology
Buckley RF, Knopman DS "Cognitive heterogeneity in Alzheimer clinical trials: harnessing noise to achieve meaningfulness" Neurology 2021; DOI: 10.1212/WNL.0000000000012027.