Racial and Ethnic Differences in Plasma Biomarker Eligibility in a Preclinical Alzheimer’s Disease Trial

Abstract

AbstractBackgroundIn Alzheimer’s disease (AD) trials, differential screen failure due to cognitive and biomarker requirements may contribute to underrepresentation of racially and ethnically minoritized groups. The AHEAD 3‐45 Study (NCT04468659) is an ongoing program testing lecanemab at the stage of preclinical AD that utilizes plasma biomarker prescreening, acquired before cognitive, clinical, and medical history eligibility assessments, to enrich for participants likely to qualify based on amyloid PET eligibility criteria.MethodsWe examined the frequency of plasma amyloid biomarker eligibility among racial and ethnic groups in the AHEAD Study. We assigned participants ages 55‐80 to mutually exclusive groups: Hispanic Black (HB), Hispanic White (HW), Non‐Hispanic Asian (NHA), Non‐Hispanic Black (NHB), and Non‐Hispanic White (NHW). We used univariate logistic regression models to explore group differences in screen failure rates as determined by an algorithm that includes the plasma Aβ 42/40 ratio, age, and APOE status. The algorithm indicates an adequately high probability of elevated brain amyloid (>20 centiloids). We further explored whether APOE ε4 status (carrier vs non‐carrier) contributed to group differences.ResultsAmong 4274 participants undergoing plasma screening, 59 (1.4%) were HB, 622 (14.6%) were HW, 74 (1.73%) were NHA, 329 (7.7%) were NHB, and 3190 (74.6%) were NHW. Screen failure rates were 86% for HB, 76% for HW; 80% for NHA; 77% for NHB, and 62% for NHW. Using NHW participants as a reference group, we observed increased odds of screen failure among all other groups (HB OR = 4.0 95% CI 2.0, 9.1; HW OR = 2.0 95% CI 1.6, 2.4; NHA OR = 2.5 95% CI 1.4, 4.5; NHB OR = 2.1, 95% CI 1.6, 2.7). Observed differences were consistent across APOE ε4 carriers and non‐carriers.ConclusionDifferential rates of amyloid eligibility were observed despite the lack of systematic sample bias due to clinical or cognitive requirements observed in previous studies. Potential explanations for these observations include differences in clinical trial access, incidences of elevated amyloid, needed cutoffs for biomarker assays, and confounding due to comorbidities or other unmeasured covariates.This work is supported by a public‐private partnership between Alzheimer’s Clinical Trial Consortium (U24 AG057437) and Eisai.