Back

Diagnostic performance of Alzheimer's disease blood biomarkers in a Brazilian cohort

Borelli, W. V.; Ferreira, P. C. L.; Brum, W. S.; Ferrari-Souza, J. P.; Carello-Collar, G.; Holz, M.; Tizeli, V.; Strelow, M. Z.; Formoso, C.; Chaves, M. L. F.; Rocha, A.; Aguzzoli, C. S.; Rohden, F.; Souza, D. G.; Schumacher-Schuh, A. F.; Povala, G.; Rosa-Neto, P.; Castilhos, R. M.; Pascoal, T. A.; Zimmer, E. R.

2025-02-24 neurology
10.1101/2025.02.23.24319116 medRxiv
Show abstract

BackgroundBlood-based biomarkers (BBMs) have emerged as promising tools to enhance Alzheimers disease (AD) diagnosis. Despite two-thirds of dementia cases occurring in the Global South, research on BBMs has predominantly focused on populations from the Global North. This geographical disparity hinders our understanding of BBM performance in diverse populations. To address this, we evaluated the diagnostic properties of AD BBMs in a real-world memory clinic from Brazil, one of the largest countries in the Global South. We measured blood and cerebrospinal fluid (CSF) biomarkers - amyloid-{beta} (A{beta})40, A{beta}42, phosphorylated tau (p-tau) 217, neurofilament light (NfL) chain, and glial fibrillary acidic protein (GFAP) - in 59 individuals. Sample comprised 20 cognitively unimpaired (CU) individuals, 22 with AD dementia, and 17 with vascular dementia (VaD). We compared BBM levels across diagnostic groups and assessed their discriminative ability for AD. Notably, individuals with VaD and AD had lower educational levels (6.8{+/-}3.0) compared to CU individuals (61.4{+/-}6.6). Among the BBMs tested, plasma p-tau217 demonstrated the best performance, exhibiting high accuracy in differentiating CU from AD (AUC 0.96) and A{beta} pathology (AUC 0.98). However, the ability of AD BBMs to distinguish between AD and VaD was lower than expected (AUC from 0.52 to 0.79), particularly when compared to studies from the Global North. Our findings highlight the potential utility of BBMs for AD diagnosis in real-world settings within the Global South. However, they also underscore the need for proper implementation and validation of these biomarkers within these populations to ensure accurate and reliable results.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

50% of probability mass above