The importance of cognitive correlates of plasma amyloid-β 42/40

The landscape of Alzheimer's disease (AD) diagnosis has shifted dramatically with the advent of blood-based biomarkers that are sensitive to plasma amyloid-β (Aβ) and p-tau levels. Commercially available blood tests for plasma Aβ and p-tau biomarkers are much less costly, more accessible, and less invasive for patients than Aβ PET or obtaining cerebrospinal fluid (CSF) markers via lumbar puncture. This increase in accessibility is welcome given that the revised criteria for the diagnosis and staging of AD from the Alzheimer's Association Workgroup defines AD by a biological process, beginning with presence of core biomarkers as measured by positron emission tomography (PET), plasma, or CSF. ¹ In particular, blood-based plasma Aβ markers such as plasma Aβ_42/40 are considered a core diagnostic biomarker, as long as plasma Aβ_42/40 levels are equivalent to approved CSF assays in detecting abnormal amyloid PET in the intended-use population. This represents a breakthrough in the use of plasma biomarkers in diagnosing AD and has been widely lauded for the potential to improve early AD screening efforts.

With this newly proposed classification of AD, clinical symptoms are not a necessary component for a biological AD diagnosis. While many view defining a disease process based on biomarkers as a positive step forward, the reactions to these criteria have been mixed. For example, the American Geriatrics Society raised concerns about the risks of labeling someone as asymptomatic with AD due to the potential unintended personal, social, or fiscal consequences. ² Neuropsychologists have also raised concerns focused on barriers to implementing the recommended Workgroup criteria, including lack of healthcare resources, equity, and costs of biomarker testing. ³ In light of this, it is important to continue investigating how easily obtained clinical measures relate to plasma AD biomarkers for broad clinical utility. Indeed, the 2024 revised Alzheimer's Association Workgroup criteria ended up incorporating a staging system that integrates biological markers with clinical symptoms to accommodate the many complexities of AD.

In a recent edition of Journal of Alzheimer's Disease, Curiel Cid and colleagues (2024) set forth to examine whether emerging blood-based Aβ biomarkers are sensitive to cognitive performance prior to the onset of clinical AD in an effort to merge AD biomarker and clinical correlates. ⁴ Specifically, they examined whether plasma Aβ_42/40 levels were related to semantic intrusion errors (SIEs) on the Loewenstein-Acevedo Scales of Semantic Interference and Learning (LASSI-L) test, a measure of verbal list learning. As defined by Curiel Cid et al., semantic interference errors (measured by SIEs) occur when learning of one item or set of items interferes with the recall of a different item. Proactive semantic interference occurs when previously learned information interferes with the ability to learn something new, whereas retroactive semantic interference is when newly learned information interferes with the recall of previously learned information. Semantic interference likely derives from regional AD neuropathology affecting explicit memory, executive functions, and semantic processing. SIEs are less frequently examined than other subtle cognitive symptoms despite high sensitivity to AD. For instance, SIEs can predict conversion from no cognitive disorder to mild cognitive impairment (MCI) or mild dementia, possibly due to their association with increased Aβ accumulation and entorhinal cortex thinning.^5–7

In this study by Curiel Cid et al., 119 participants were categorized as either having amnestic MCI (aMCI) or as cognitively unimpaired based on neuropsychological tests independent of the LASSI-L. All participants were subgrouped as either Aβ+ or Aβ- on PET imaging. As expected, SIEs differed between the aMCI Aβ+ and Aβ- groups such that the Aβ+ group performed worse on measures of semantic interference. aMCI participants with low plasma Aβ_42/40 ratios had a higher percentage of SIEs on learning trials, indicating behavioral correlates of this blood-based biomarker. This finding provides evidence for the concordance of plasma Aβ_42/40 and Aβ PET in detecting cognitive symptoms in a population at high risk for developing AD (aMCI).

The authors of this study effectively highlight the importance of validating clinical utility of blood-based biomarkers, especially in light of the proposed biomarker-based diagnostic criteria of AD. These types of examinations are necessary in the quest to develop a comprehensive definition of AD that incorporates both biological precursors and clinical consequences. However, in the cited sample, the neuropsychological battery was not comprehensive. It relied heavily on verbal memory measures at the expense of detailed visual memory, complex attention, and executive measures. As such, it is unclear how the LASSI-L might function for diagnostic specificity and relation to plasma Aβ_42/40 ratios when considering both aMCI and non-amnestic MCI. Also, while the results were significant, they were based on differences of only 1 word or 8.2% SIE error rates, which the authors note are smaller effect sizes than seen in other LASSI-L examinations using Aβ PET positivity. ⁷ Finally, there is some evidence that health factors such as chronic kidney disease may impact plasma Aβ_42/40 levels in cognitively unimpaired individuals. ⁸ Without knowing the broader health context of the sample examined in the Curiel Cid. et al., investigation, it is unclear whether these health considerations are of similar concern for people with aMCI and future work may want to examine this.

With these broader considerations, it is encouraging that this work by Curiel Cid and colleagues finds a correlate of plasma Aβ_42/40 with a cognitive marker in people at high risk for AD. Plasma AD biomarkers are still relatively new and measure peripheral levels of Aβ, which do not only emerge from brain-derived sources. Indeed, plasma Aβ_42/40 has a more limited clinical range of detecting Aβ as compared to CSF Aβ_42/40. ⁹ For blood-based AD biomarkers to become clinically useful, they must have high sensitivity and specificity in a broad range of settings. There is significant need to harmonize plasma biomarker cut-off scores so that clinical utility does not rely solely on homogeneous samples derived from convenience-based memory clinics and consider factors that may affect diagnostic cut-off values such as age, comorbidities, or other differences including social determinants of health. Ideally, plasma biomarker assays must be deployable to settings that have minimal access to specialized screening procedures or tertiary care centers. Because of this, the clinical utility of a plasma screening marker cannot be detached from the sensitivity of the test to Aβ PET levels, and future investigations should continue to examine plasma Aβ and p-tau to test sensitivity for behavioral correlates in a diverse range of settings and populations.