Comparing Detection Methods for Alzheimer Disease Biomarkers: CSF vs PET

The prevalence of amyloid abnormality was higher in individuals across the Alzheimer disease (AD) spectrum with cerebrospinal fluid (CSF) testing, according to a study published in JAMA Neurology. The CSF-based and positron emission tomography (PET)-based estimates, however, were similar in those with clinical AD dementia.

In patients with AD, pathologic changes are distinguished by cerebral amyloid aggregation, as observed by the presence of amyloid biomarkers on PET scans or in CSF. Estimating the prevalence of amyloid abnormalities in individuals across the AD clinical spectrum is key to health care planning and the reduction of screening failure rates.

The researchers sought to estimate the prevalence of amyloid abnormality in individuals with normal cognition, subjective cognitive decline, mild cognitive impairment, or clinical AD dementia, as well as to evaluate the potential implications of cutoff methods, biomarker modality (PET or CSF), sex, age, APOE genotype, geographic locale, educational level, and severity of dementia for these assessments.

A cross-sectional, individual-participant, pooled study was conducted among individuals from 85 cohorts of the ongoing, worldwide Amyloid Biomarker Study, with data obtained between January 1, 2013, and December 31, 2020. All of the participants exhibited normal cognition, subjective cognitive decline, mild cognitive impairment, or clinical AD dementia. Normal cognition and subjective cognitive decline were delineated by normal scores on cognitive tests, with subjective cognitive decline defined by the presence of cognitive complaints. Published criteria were used for the diagnosis of mild cognitive impairment and clinical AD dementia.

A total of 19,097 participants were enrolled in the study, 53.1% of whom were women and 46.9% of whom were men. The mean participant age was 69.1±9.8 years. Overall, 8.2% (1571 of 19,097) of the participants underwent both CSF and PET measurements. In total, 53.1% (10,139 of 19,097) of the participants in 50 cohorts received an amyloid-PET measurement (26 quantitative readings, 23 visual readings, and 1 combined reading). Further, 46.9% (8958 of 19,097) of the participants had an amyloid-CSF measurement.

Study results showed that with the use of cohort-provided cutoffs, the prevalence of amyloid abnormalities was similar to estimates from 2015 in individuals without dementia. Additionally, the prevalence was similar across PET-based and CSF-based estimates (24%; 95% CI, 21% to 28%) in individuals with normal cognition; 27%
(95% CI, 21% to 33%) in those with subjective cognitive decline; and 51% (95% CI, 46% to 56%) in participants with mild cognitive impairment. In contrast, in those with clinical AD dementia, the estimates of amyloid abnormalities were significantly higher with PET scans than with CSF analysis (87% vs 79%, respectively; 95% CI, 0% to 16%; P =.04).

Further, adjusted CSF cutoffs were associated with a 10% higher prevalence of amyloid abnormality compared with PET-based estimates among individuals with normal cognition (95% CI, 3% to 15%; P =.004), in those with subjective cognitive decline (95% CI, 3% to 15%; P =.005), and in individuals with mild cognitive impairment
(95% CI, 3% to 17%; P =.004). In those with clinical AD dementia, however, the CSF-based and PET-based estimates of amyloid abnormality were comparable (95% CI, –2% to 9%; P =.18).

Study limitations included the fact that the use of cohort-specific cutoffs to define abnormal amyloid in cohorts for which no data-driven cutoff could be calculated may have resulted in an underestimation of amyloid abnormality in these cohorts. Moreover, the cross-sectional design of the study might have led to an underestimation of amyloid abnormalities as opposed to lifetime risk estimates.

The researchers concluded that the findings from the present study might be useful in “health care planning, providing potential eligible patient population sizes for antiamyloid therapies, and in recruitment strategies for clinical trials.”

Disclosure: Some of the study authors have declared affiliations with biotech, pharmaceutical, and/or device companies. Please see the original reference for a full list of authors’ disclosures. 

Reference  

Jansen WJ, Janssen O, Tijms BM, Vos SJB, Ossenkoppele R, Visser PJ; Amyloid Biomarker Study Group. Prevalence estimates of amyloid abnormality across the Alzheimer disease clinical spectrum. JAMA Neurol. Published online January 31, 2022. doi:10.1001/jamaneurol.2021.5216