Inclusion of ethnoracial populations and diversity remains a key challenge in Alzheimer’s disease biofluid-based biomarker studies

Along with a rapidly growing population of older adults, Alzheimer’s disease and related dementia (ADRD) are prevalent worldwide (50 million in 2020) and will continue to increase (152 million in 2050) in subsequent decades [1]. Low-and-middle-income countries will see more than two-thirds of dementia cases and added burden by 2050 [2]. These growths will also be accompanied by more racial and ethnic diversity among adults worldwide. Alzheimer’s disease (AD) alone accounts for 60–80% of dementia cases—a significantly higher risk for ADRD has been reported for Blacks/African Americans and Hispanics/ Latinos when compared to non-Hispanic Whites [3,4].

Based on the biological contributions of amyloid and tau proteins in AD pathology, biomarkers provide a valuable window into the pathophysiological process that undergird the disease. Collection of biological data can help to potentially inform and improve the efficacy and effectiveness of clinical trials and translation of clinical knowledge. However, conducting biofluid and neuroimaging biomarker studies can be prohibitive due to the high costs of equipment, tracers, and reagents, the time to establish strict protocols and collect data, the burden for participants, and the requirement of specialized, quality control infrastructures. This challenge is further exacerbated by the notable un-derrepresentation of Blacks/African Americans and Hispanics/Latinos populations in biomarker research studies. With the increased risk for some underrepresented groups, recruitment, enrollment, and retention into prospective studies is a precedent for understanding ADRD etiology, progression, and protective and risk factors [5,6].

In this issue of the Journal of the Neurological Sciences, Drs Chaudhry and Rizig conducted a meticulous systematic review and meta-analysis of the literature to determine differences in established AD biofluid biomarkers between African Americans and Black Africans compared to White adults. The authors introduced the study, providing a detailed narrative by defining dementia, frameworks for classification of AD, biomarker types, systemic barriers leading to White, homogenous samples, and the need to study disparities and ethnoracial factors beyond biology. The review aimed to compare levels of Aβ_1–42, phos-phorylated tau, and total tau measured between African Americans/ Black Africans and Whites. Studies were required to 1) use at least one AD biomarker in fluid (cerebrospinal fluid [CSF], urine, saliva or plasma) in both African American/Black African and White groups, 2) establish the cognitive status of the sample (cognitively normal, mild cognitive impairment [MCI], AD), and 3) be an observational study with cross-sectional measurement of biomarkers at one timepoint. Following the search, quality assessment, and extraction, only five studies met the inclusion criteria in the past 20 years across established databases.

The handful of cross-sectional studies were conducted in the United States with no published articles capturing Black Africans or Whites in Africa. Four studies used CSF and one used plasma biomarkers. Race was self-reported and a total of 359 African Americans and 1171 Whites were pooled across the five studies. In the quality appraisal using BIO-CROSS [7], Chaudhry and Rizig reported concerns with population representation where participation rate was not reported in three studies, power calculations were not reported in four studies, reproducibility was questioned since the laboratory/place of measurement was not listed in four studies, and modeling to detect outliers in biomarker samples was only reported in one study. Further examination of confounding variables in the five studies was also reported in comprehensive detail. Overall, only two studies were deemed to be high-quality and the other three were classified as mid-quality.

Chaudhry and Rizig compared differences in AD biomarkers between African Americans and Whites separately in participants with normal cognition and in those how had cognitive impairment. The four CSF studies were then meta-analyzed to find the standardized mean difference in biomarker levels. Interestingly, in participants with normal cognition, there were no differences in CSF Aβ, yet CSF ptau₁₈₁ and t-tau were significantly higher in White participants compared to African Americans. The authors found that there was substantial heterogeneity in both CSF ptau₁₈₁ (I² = 74%) and t-tau (I² = 78%) meta-analysis. In the fifth study using plasma samples, African Americans had higher ptau and t-tau levels, but lower levels of Aβ_1–42 when compared to whites. In participants with MCI or mild dementia, three CSF studies examined biomarker differences. Differences in results were consistent with the cognitively normal group where there was no difference in CSF Aβ, but there were statistically significant higher values of CSF ptau₁₈₁ and tau in Whites compared to African American participants. However, no heterogeneity was found in either CSF ptau₁₈₁ (I² = 0%) and t-tau (I² = 0%) meta-analysis. In the single plasma study of MCI, Aβ_1–42 and t-tau were lower, but ptau was higher in African Americans compared to White participants.

In the interpretation of these results, Chaudhry and Rizig interrogated discrepancies in the extant literature on group differences in racebased clinical evaluation, genetic encoding and expression, and neuropathological diagnosis data, but concluded that a biological explanation for the differences found in this study to be unclear. The authors further discussed disparities in representation within AD research by exploring barriers to the recruitment of non-white participants. Most importantly, explanations delved beyond the onus of the individual-level around passivity and disinterest toward research. Instead, actionable drivers shifted to researchers and institutions creating and investing resources in inclusive access and strategies for sustainable representation in underrepresented populations. Limitations of this systematic review and meta-analysis were noted to include human error whilst reviewing/ coding studies, language restriction to English, cross-sectional study designs, and the study-specific limitations—mainly with small sample sizes of African American participants. Drs. Chaudhry and Rizig advanced their study by providing practical strategies for improvement of participant engagement at the level of the researcher, community, and institution. They also stressed the need for greater ethnoracial representation, effortful outreach practices, transparency of methodologies in publications, and replication of results in longitudinal studies. Overall, this systematic review and meta-analysis presented a thorough assessment of available cross-sectional studies that examine group differences in biofluid biomarker levels between African Americans and Whites.

Molecular biomarkers provide insight into the pathobiological process of dementia and can help to inform clinical assessments and staging of the disease, along with elucidation of other biological processes. Their pragmatic value is weighed with their practical limitations and the onus it places on the participant for safe and uniform data collection. Gold standard procedures like CSF via a lumbar puncture and imaging with various radiotracers are invasive and time and effort intense for participants, leading to disparities in available data for most ethnoracial groups. The advent of blood-based biomarkers (plasma/serum) offers significant promise with the acquisition of various analytes via a common blood draw procedure. Further testing and validation are required for blood-based biomarkers. However, this may help to reduce the burden on participants and patients and, more importantly, increase the representation of diverse, ethnoracial populations in biomarker studies.

Biomarkers by no means are a silver bullet for understanding the multidimensional factors that may influence differences found in the biological process of AD. It is imperative for researchers to work collaboratively in multilingual, multicultural, and multidisciplinary teams to collect data on the life course and social determinants of health since these may be key drivers of observed differences. Additionally, the use of Whites as the gold standard and a control group assumes the same process occurs in both groups to impact the outcome [8]. Control groups are valuable for the comparative effectiveness of interventions or pharmacological treatments; however, within-group designs may serve to better understand how social forces impact the life course and onset of diseases like AD. Furthermore, the active use of inclusive frameworks like the National Institute on Aging’s Health Disparities Research Framework can help to orient and guide multiple levels of collection and analysis [9].

Finally, treatment of race as a biological construct is prevalent in the aging and ADRD literature. Scientists imbued the concept as biology-centric based on overt physical characteristics. This perspective is still commonly used today, a testament to the pervasive institutional ideologies that endure. However, recent discourses eschew this traditional notion and critically revise race as a unidimensional social construct used to delineate groups. Race as a covariate or stratification variable offers very little by way of understanding integral differences in disease manifestation and progression—as an identifier, it serves as a vital proxy for marginalization, structural and institutional discriminatory practices, and historically vulnerable groups. Scientists need to redefine and shift from the use of homogenous binning (e.g. Asians, Latinos), blanket umbrellas (Black, White), hierarchical grouping (majority vs. minority), and geopolitical labels (immigrant, undocumented aliens, citizens). Importantly, race and ethnicity should not be conflated, nor should researchers ignore mixed, or more than one race, category. In the next three decades, increasing globalization, transnationalism, and social connectivity will bring new challenges with the diversification of an aging population and a re-description of identity based on race and ethnicity. Established census classifications will need to be thoughtfully redefined with respect to intergenerational families, language, gender, sex, geographic boundaries, and lived experiences. Researchers and scientists in the fields of aging and ADRD should start to measure these constructs and their multidimensional relationships to better understand if and how they influence differences in biomarkers.