Plasma biomarkers predict Alzheimer’s disease before clinical onset in Chinese cohorts

Alzheimer's disease (AD) is a leading cause of dementia, imposing a significant global burden. Current treatments lack the ability to prevent or significantly slow AD progression. The neuropathological changes in AD precede clinical symptoms by decades, highlighting the potential benefits of early intervention. Early detection during the preclinical phase is crucial, requiring a shift from symptom-based diagnosis to a biological definition of AD. The National Institute on Aging-Alzheimer's Association (NIA-AA) framework emphasizes the amyloid, tau, and neurodegeneration (ATN) system, traditionally assessed through invasive CSF sampling or expensive imaging techniques. The need for readily available, cost-effective, and minimally invasive biomarkers has led to the exploration of blood-based biomarkers, such as plasma Aβ42, p-tau181, and NfL, as potential predictors of AD. While their effectiveness in other populations has been studied, their predictive value in Chinese populations remains under-explored. This study aims to evaluate the performance of these plasma biomarkers in predicting AD in Chinese cohorts, potentially paving the way for early intervention strategies.

Existing literature demonstrates the potential of plasma Aβ42, p-tau181, and NfL as AD biomarkers. Studies using insurance databases have shown the ability of these biomarkers to distinguish AD from healthy controls or individuals with other neurodegenerative diseases. Previous research has indicated associations between these plasma biomarkers and various AD-related phenotypes, including cognitive decline, brain atrophy, and amyloid/tau burden. However, research across diverse ethnic and racial groups, including Chinese populations, is crucial, as genetic and environmental factors may influence biomarker performance. While some studies have investigated these biomarkers in Chinese populations, a comprehensive longitudinal analysis with a focus on preclinical AD is lacking.

This study employed a two-cohort design. Cohort 1 comprised 126 participants with preclinical AD and 123 controls, followed for 8–10 years. Cohort 2 included 15 familial AD mutation carriers and 52 non-carriers from the familial Alzheimer’s Disease Network. Plasma Aβ42, p-tau181, and NfL levels were measured at baseline and follow-up in Cohort 1. Correlation analyses were conducted to assess the relationship between plasma and CSF biomarkers. Logistic regression models were developed to determine the ability of plasma biomarkers to discriminate between preclinical AD and control groups. The performance of these models was evaluated using receiver operating characteristic (ROC) curve analyses, calculating the area under the curve (AUC). Cohort 2 served as a replication cohort to validate the findings from Cohort 1. Statistical analyses included two-sided Student's t-tests, linear regression, and DeLong tests.

In Cohort 1, plasma Aβ42 levels were significantly lower, while p-tau181 and NfL levels were significantly higher in participants with preclinical AD compared to controls. Plasma biomarkers showed significant correlations with their CSF counterparts. A logistic regression model combining plasma Aβ42, p-tau181, and NfL effectively discriminated between preclinical AD and controls, yielding AUCs of 0.78 at baseline and 0.90 at follow-up. The addition of APOE ε4 status improved prediction accuracy at baseline but not at follow-up. Cohort 2 replicated the key findings, demonstrating that the combination of plasma Aβ42, p-tau181, and NfL achieved an AUC of 0.79 in differentiating mutation carriers (preclinical AD) from non-carriers. Individual biomarkers showed lower predictive accuracy than the combined model. While Aβ40 and total tau showed less consistent patterns and didn't contribute significantly to the predictive model.

This study's findings provide strong evidence for the utility of plasma Aβ42, p-tau181, and NfL as predictors of AD in Chinese populations. The significant correlations between plasma and CSF biomarkers support the use of plasma as a less invasive alternative for AD assessment. The high AUC values obtained from the combined biomarker model demonstrate the model's strong discriminative ability in identifying individuals at risk of developing AD years before clinical onset. The replication of these findings in a separate cohort strengthens the generalizability of the results. These results align with previous research demonstrating the predictive value of plasma biomarkers for AD in other populations, extending these findings to a significant demographic group. The findings have implications for early detection and intervention strategies for AD in the Chinese population.

This study demonstrates the effectiveness of a combined plasma biomarker model (Aβ42, p-tau181, and NfL) in predicting Alzheimer's disease in Chinese populations up to 8 years before clinical onset. The findings support the use of this non-invasive approach for early detection and potential intervention. Future research should focus on larger, more diverse cohorts to further validate these findings and explore the clinical implications of these biomarkers in a real-world setting, also investigating the reasons for inconsistent findings for Aβ40 and total tau.

The relatively smaller sample size, despite being significant, could limit the generalizability of the findings to the broader Chinese population. While the study included both a longitudinal cohort and a replication cohort, the inclusion of additional cohorts with different genetic backgrounds might strengthen the findings. The cross-sectional nature of cohort 2 might affect the comparison with the results from longitudinal cohort 1. Differences in assay methods across studies might influence the comparison of findings. Further research with larger and more diverse cohorts is needed to fully validate these results and investigate potential interactions among biomarkers and other clinical variables