Prevalence of abnormal amyloid was about 10% higher than previously estimated among people with normal cognition, updated data from the Amyloid Biomarkers Study indicated.
Across 19,000 people in 85 cohorts, those without dementia had higher amyloid abnormality prevalence when adjusted data-driven cutoffs of cerebrospinal fluid (CSF) measures were used, suggesting preclinical and prodromal Alzheimer's disease could be more prevalent than once thought, reported Willemijn Jansen, PhD, and Olin Janssen, MSc, of Maastricht University in the Netherlands, and co-authors in JAMA Neurology.
"At age 50, about 18% of persons with normal cognition had abnormal amyloid, increasing to 54% at age 90," Jansen told MedPage Today. "This was 10% higher than we previously estimated when using unbiased CSF cutoffs."
"Overall, the prevalence of amyloid abnormality was 10% higher in all pre-dementia stages," she added. "Also, the prevalence estimated with these new CSF cutoffs was 10% higher as compared to PET estimates of amyloid abnormality. This indicates that CSF amyloid abnormality might be more sensitive to identify persons earlier in the disease process compared with PET."
The methodology to define abnormal amyloid cutoffs differs by center and can influence prevalence estimates, Janssen noted. "It has appeared recently that some CSF values gradually increased over the past 2 decades, indicating that older cutoffs might be too conservative," she told MedPage Today. "Using older available CSF cutoffs that are not corrected for drift in CSF values might lead to an underestimation of amyloid abnormality."
The findings update earlier data from the Amyloid Biomarkers Study, which reported the prevalence of amyloid abnormality in 56 cohorts of people with and without dementia in 2015.
The current analysis extended that work by increasing sample sizes to 19,097 participants in 85 cohorts, including 9,908 people with normal cognition, 1,524 with subjective cognitive decline, 5,405 with mild cognitive impairment, and 2,260 with Alzheimer's dementia. Mean age was 69, and 53% were women. CSF and PET data were collected from 2013 through 2020.
Amyloid measurements were dichotomized as normal or abnormal using cohort-defined cutoffs for CSF or PET. Adjusted data-driven cutoffs for abnormal amyloid were calculated using Gaussian mixture modeling that incorporated within-cohort distributions of continuous amyloid values.
Using cohort-defined cutoffs, amyloid abnormality prevalence mirrored 2015 estimates for people without dementia and was similar across PET- and CSF-based estimates, at 24% for people with normal cognition, 27% for people with subjective cognitive decline, and 51% for people with mild cognitive impairment. For people with clinical Alzheimer's dementia, however, estimates were higher for PET than CSF (87% vs 79%, mean difference 8%, 95% CI 0%-16%, P=0.04).
Adjusted cutoffs for PET amyloid measures were similar to cohort-defined cutoffs. For CSF, adjusted cutoffs resulted in higher amyloid abnormality prevalence than PET-based estimates in people with:
- Normal cognition (mean difference 9%, 95% CI 3%-15%, P=0.004)
- Subjective cognitive decline (mean difference 9%, 95% CI 3%-15%, P=0.005)
- Mild cognitive impairment (mean difference 10%, 95% CI 3%-17%, P=0.004)
Adjusted cutoffs between CSF and PET amyloid were comparable in people with clinical Alzheimer's dementia (mean difference 4%, 95% CI -2% to 9%, P=0.18).
"These prevalence estimates can improve recruitment efficiency for clinical trials that target individuals with biomarker positivity," noted Christina Young, PhD, and Elizabeth Mormino, PhD, both of Stanford University School of Medicine in Palo Alto, California, in an accompanying editorial.
Amyloid alone is not sufficient to predict clinically meaningful Alzheimer's progression, Young and Mormino pointed out. "A primary source of controversy with the approval of aducanumab [Aduhelm] was the ambiguity of the association between treatment and clinical outcomes in the phase III trials," they noted.
"Although long-term cognitive assessments of individuals with amyloid positivity have shown cognitive decline at the group level, there has also been considerable heterogeneity in who experiences decline and how fast decline occurs," the editorialists added. "Characterization of clinically meaningful decline, especially among individuals with amyloid positivity and normal cognition, is an active area of research that is needed to elucidate the risks associated with amyloid positivity."
Race and ethnicity were not included in the data collected, which is a limitation of the study, Young and Mormino observed. "Given this gap, amyloid-positivity prevalence rates estimated by Jansen and colleagues should not be assumed to be reflective of all racial and ethnic groups," they wrote.
In addition, Gaussian mixture modeling could be applied only to a subset of cohorts that provided continuous data, the researchers noted.
Disclosures
This study was funded by Biogen. An employee of Biogen had a role in the analysis plan, review, and revision of the manuscript.
Jansen and Janssen reported no disclosures. Co-authors listed numerous relationships with academic institutions, nonprofit organizations, government agencies, and industry.
Young reported receiving grants from the Alzheimer's Association. Mormino reported relationships with NIH, Eli Lilly, Genentech, Neurotrack, and Roche.
Primary Source
JAMA Neurology
Source Reference: Jansen WJ, et al "Prevalence estimates of amyloid abnormality across the Alzheimer disease clinical spectrum" JAMA Neurol 2022; DOI: 10.1001/jamaneurol.2021.5216.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.