Circulating Biomarkers of Neurodegeneration and Risk of Cognitive Impairment

Blood-based biomarkers offer a widely available, scalable, and minimally invasive alternative to study neurodegeneration. However, the association between blood-based biomarkers of neurodegeneration and long-term risk of cognitive impairment in a racially diverse cohort remains understudied. The objective was to determine whether baseline biomarkers of neurodegeneration are associated with the development of incident cognitive impairment in a biracial cohort. The Reasons for Geographic and Racial Differences in Stroke cohort study enrolled 30,239 Black and White participants from 2003 to 2007 across the continental United States with ongoing follow-up. In a random sample of participants with no baseline cognitive impairment, plasma neurofilament light chain (NfL), total tau, glial fibrillary acidic protein (GFAP), and ubiquitin carboxy-terminal hydrolase L1 (UCH-L1) were measured. Incident cognitive impairment was defined using the Six-Item Screener (SIS) and a 4-test battery consisting of Word List Learning, Delayed Recall, Animal Fluency, and Letter "F" Fluency. Models were adjusted for demographics, medical history, prebaseline and incident stroke, lifestyle habits, and depressive symptoms. A total of 724 participants had baseline plasma markers of neurodegeneration measured; the mean baseline age was 66.1 years (SD 11.7), 51.7% were female, and 43.6% self-identified as Black. With a mean follow-up of 12.2 (SD 4.8) years, 8.9% (64/724) developed incident cognitive impairment. In fully adjusted models for each biomarker individually, the hazard ratios (HRs) of incident cognitive impairment per SD increment of GFAP (HR 1.71, 95% CI 1.04-2.80) and NfL (HR 1.63, 95% CI 1.08-2.46) were increased. The highest tertile of UCH-L1 was associated with an increased risk (HR 2.42, 95% CI 1.10-5.30) compared with the lowest tertile. No association was observed for total tau. Associations did not differ by race, sex, or age. All interactions with demographics were not significant for the models with incident cognitive impairment as the outcome. Strong associations of baseline circulating biomarkers of neurodegeneration with cognitive impairment over 12.2 years suggest that early pathologic brain changes can be detected noninvasively years before clinical symptoms of cognitive impairment. This relationship persisted after statistical adjustment for other risk factors. Possible clinical and research applications for these biomarkers merit intensive investigation.