Racial differences in AD CSF biomarkers in persons with MCI: Implications and insights

Abstract

AbstractBackgroundThe implications for potential racial differences in tau CSF biomarkers for Alzheimer's disease (AD) and mild cognitive impairment (MCI) are unclear for standard cutoffs used for determining enrollment criteria into novel clinical trials, and insight into interaction with markers of non‐biological determinants.MethodSetting & Participants‐ Data presented were obtained from baseline assessment of a case control study conducted from March 1, 2016 through January 31, 2019. Participants were 50 years or older adult volunteers recruited from Atlanta, Georgia into the Brain Stress Hypertension and Aging Research Program (B‐SHARP, PI: Hajjar R01AG051633) at the Emory Goizueta Alzheimer’s Disease Research Center. Exposures‐ Self‐reported race and cognitive status per modified Petersen criteria. Outcome measures‐ CSF levels of Aβ1‐42, tau, P‐tau181, ratio of tau or P‐tau181 to Aβ1‐42, and hippocampal volume.Result362 study participants were analyzed (mean [SD] age, 65.6 [7.9] years of age), where 152 (42.0%) were African American, 230 (63.5%) were women, and 189 (52.2%) had MCI. After adjusting for multiple determinants, including age, sex, cognitive performance, body mass index (BMI), education, and family history of AD, results suggest Black Americans compared to white individuals with MCI had lower mean (SE) levels of CSF tau (52.4 [5.9] vs. 78.98 [5.02] pg/mL; P=.001) and P‐tau181 (15.42 [2.06] vs. 25.24 [1.75] pg/mL; P=.001). After adjusting for intracranial volume and other covariates in the total sample hippocampal volumes did not differ by race. However, using ROC analyses, AUC comparisons suggest that Aβ1‐42 (0.632 vs 0.779; P=.02) and P‐tau181 (0.638 vs 0.768;P=.03) had lower specificity and sensitivity in Black Americans to distinguish between MCI and normal cognition.ConclusionWe recommend caution in using current biomarker cutoffs to determine enrollment criteria for potential Black American study cohorts. While we hypothesized that these differences could be explained by increased vascular contributors, our supposition was not supported; after adjusting for hypertension and diabetes, racial differences yet remained. Further investigation should clarify the impact of biological factors like vascular disease, interaction with more explicit and well‐defined environmental determinants like education quality and quantity, and their life‐course occurrence.