Plasma Alzheimer’s biomarkers tracked different clinical signals in Latino older adults
Blood-based biomarkers are moving closer to everyday Alzheimer’s disease care, but much of the evidence still comes from cohorts that do not reflect the populations most likely to face barriers to diagnosis. In the Boston Latino Aging Study, investigators examined whether plasma markers of amyloid, tau, and neurodegeneration were linked to cognition and depressive symptoms in Spanish-speaking Latino older adults without dementia.
The study centered on Alzheimer’s disease blood-based biomarkers that are increasingly discussed as tools for early risk characterization: plasma Aβ42/40, brain-derived phosphorylated tau 217, neurofilament light chain, and glial fibrillary acidic protein. Those markers were measured with NULISA in 122 adults aged 55 years and older. Cognitive performance was assessed with the Latin American Spanish version of the Preclinical Alzheimer’s Cognitive Composite, LAS-PACC5, and depressive symptoms were measured with the Geriatric Depression Scale. Regression models adjusted for age, sex, and education.
Even in this relatively small baseline sample, biomarker patterns separated into clinically meaningful directions. Participants with cognitive impairment had higher p-tau217 and GFAP than cognitively normal participants. Higher p-tau217 and GFAP were also associated with poorer LAS-PACC5 performance. Higher NfL aligned with greater depressive symptoms. Aβ42/40 did not show significant associations with either cognition or mood.
Tau and astroglial injury carried the strongest cognitive signal
The most direct cognitive associations fell to p-tau217 and GFAP. In practical terms, that pattern fits with the biology of Alzheimer’s disease: phosphorylated tau is tied to pathologic tau accumulation, while GFAP reflects astroglial activation and injury responses in the central nervous system. When both rise alongside poorer performance on a sensitive preclinical cognitive composite, the signal points toward active neurobiological change rather than nonspecific aging alone.
The finding matters because the study focused on Spanish-speaking Latino older adults, a group for whom biomarker data remain limited despite growing need. Cognitive screening tools and biomarker interpretation can behave differently across language and cultural contexts, so a Spanish-language cognitive composite anchored the analysis in a more appropriate framework for this population. The association between elevated p-tau217 and worse LAS-PACC5 performance adds to the case that plasma tau measures may have real utility outside the cohorts that have dominated biomarker development.
GFAP followed the same direction. Elevated glial fibrillary acidic protein, paired with poorer cognitive performance, suggests that astrocytic response may be part of the biochemical signature accompanying early symptomatic change. The result does not prove causality, and it does not establish diagnosis. It does, however, suggest that blood markers capturing both tau pathology and glial reactivity may be more informative than a single biomarker viewed in isolation.
A different biomarker signal emerged for depressive symptoms
NfL behaved differently. Higher neurofilament light chain was associated with greater depressive symptoms, while the amyloid ratio remained uninformative in this sample. NfL is widely viewed as a marker of axonal injury and broader neurodegeneration, so its relationship with mood symptoms may reflect a shared burden of brain vulnerability rather than a mechanism specific to Alzheimer’s pathology.
The mood finding is clinically interesting because depressive symptoms are common in older adults and often complicate cognitive assessment. In memory-clinic settings, depression can blur the boundary between subjective cognitive concerns and measurable impairment. A biomarker linked to depressive symptoms may therefore carry value beyond dementia risk stratification alone, especially if future work confirms whether NfL adds information about neuropsychiatric burden, underlying neural injury, or both.
Here, too, the data remain associative. The study did not claim that NfL causes depressive symptoms, and it did not establish whether the biomarker tracks symptom severity, chronicity, or a broader neurodegenerative process. Still, the dissociation is notable: p-tau217 and GFAP mapped most clearly to cognition, while NfL aligned with mood.
Why the null amyloid result still matters
Aβ42/40 did not show significant associations with cognition or depressive symptoms in this baseline analysis. In a field where amyloid has long dominated the discussion around Alzheimer’s disease, a null result can be as informative as a positive one. It suggests that in this cohort, plasma amyloid ratio alone was not the marker most closely tied to cross-sectional function.
That does not diminish amyloid’s role in disease biology. Instead, it reinforces the idea that biomarker performance depends on the clinical question being asked. If the question is who already shows measurable cognitive differences at baseline, tau- and glial-related markers may offer a sharper readout than amyloid ratio in some populations. If the question is broader preclinical disease staging, amyloid may still contribute in ways not captured by this single analysis.
The cohort’s design leaves those distinctions open. Baseline observations cannot establish trajectory, and they cannot determine whether biomarker elevations precede cognitive change, follow it, or track a parallel process. Longitudinal follow-up will be necessary to know whether p-tau217, GFAP, and NfL predict future decline, stabilize over time, or respond to other clinical variables.
A Latino cohort adds needed representation to biomarker research
The study’s real contribution may be representational as much as biological. Latino older adults remain underrepresented in many Alzheimer’s biomarker studies, yet they face substantial dementia burden and persistent structural barriers to timely care. A cohort drawn from the Boston Latino Aging Study helps narrow that gap by examining plasma markers in Spanish-speaking participants with a culturally and linguistically appropriate cognitive measure.
The sample size was modest, with 122 adults, and the analysis was cross-sectional. Those limits matter when trying to generalize to the broader U.S. Latino population or to Latino communities across different countries of origin, educational backgrounds, and health systems. The work also examined a specific set of plasma markers and one mood instrument, so it does not capture the full complexity of neuropsychiatric symptoms in aging.
Even so, the findings support a simple clinical direction: different plasma biomarkers may be linked to different domains of brain health. p-tau217 and GFAP appear closer to cognition, while NfL may relate more closely to depressive symptoms in this cohort. The next step is longitudinal validation to determine whether these blood measures can help identify who is most likely to decline, who needs closer monitoring, and which biomarker combinations will be most useful in Latino older adults before dementia is present.