29 Apr 2022

Historically, Alzheimer’s disease trials in the U.S. and Europe have enrolled mostly people of Caucasian heritage. The field is now increasing efforts to enroll more racially and ethnically diverse populations. To aid this process, the Food and Drug Administration on April 14 issued a new draft guidance for industry. It lays out criteria for improving enrollment of underrepresented groups. The guidance applies to trials in all indications, not just AD.

At the same time, two recent studies underscored both the need for, and complexity of, diverse populations in trials and biomarker studies. In the April 19 JAMA, researchers led by Kristine Yaffe at the San Francisco Veterans Affairs Health Care System strengthened prior findings of higher dementia risk in Hispanic and black populations with data from nearly 2 million veterans, the largest nationwide study to date. In the April 21 Neurology, Suzanne Schindler and colleagues at Washington University School of Medicine in St. Louis reported that plasma p-tau was a less consistent marker of brain amyloidosis in groups of black than white people. These data hint that biomarker findings from predominantly white populations may not translate well to trials and broad clinical practice.

“It is critical that treatments be evaluated in cohorts that reflect the patient population,” Schindler told Alzforum. “The FDA’s guidance to have a plan to include a diverse cohort in clinical trials is very reasonable, although it may be practically very difficult to achieve a representative cohort.”

FDA Guidance. The agency outlines five major categories to consider when devising a plan to include diversity in clinical trials.

Previous studies have suggested differences in dementia risk by race. Most have reported higher risk for African Americans and Latinos than for whites and Asians, with inconsistent findings on risk among other minorities (Mayeda et al., 2016; Dec 2020 news; Alzheimer’s Association 2021 report). Many of these studies were small or geographically limited, with a particular paucity of data on Asian and Native American populations (Mehta and Yeo, 2017). 

Yaffe sought to get around this limitation by examining retrospective data from the Veterans Affairs database. First author Erica Kornblith randomly selected 5 percent of patients seen nationwide each year over the 20-year period from 2000 to 2019, for a total of 1,869,090 veterans included in this analysis. Nearly all were men, their average age was 69, and they were assessed over an average of 10 years, with at least three healthcare visits each. About 10 percent, or 176,795 participants, were African American. Other minorities were less well-represented, with 1 percent (20,663) Hispanics, 0.5 percent (9,391) Asians, and 0.4 percent (6,865) Native Americans, reflecting the makeup of the armed forces in those years. In the 2020 census, the U.S. population was 18.7 percent Hispanic, 12.4 percent black, 6 percent Asian, and 1.1 percent Native American.

Overall, 13 percent of the cohort developed dementia. The age-adjusted rates by race were 11.5 percent for whites, 12.4 percent for Asians, 14.2 percent for American Indian or Alaska Natives, 19.4 percent for blacks, and 20.7 percent for Hispanics. In terms of hazard ratios, Native Americans were not significantly different from whites. Asians had 1.2 times the risk of whites, while blacks and Hispanics had 1.5 and 1.9 times the risk, respectively.

It is unclear what drove this higher risk. The numbers were adjusted for age, sex, education, and co-morbidities, and were similar in different parts of the country. Because all participants were veterans, they all had access to roughly the same quality medical care. Other socioeconomic factors might be responsible for worse health outcomes for minorities, the authors suggested.

In an accompanying editorial, Gwen Yeo at Stanford University School of Medicine, Palo Alto, California, noted that because education was assessed indirectly in this study, by the percentage of people in the participant’s ZIP code who had a college degree, the authors may have been unable to fully account for its effects on dementia risk.

How can researchers assess whether trials are capturing disease demographics? A recent paper from the cancer field describes a “participation to prevalence” ratio (PPR), where perfect representation would be 1. In CAR-T cancer trials, the actual ratios are far below that, at least for black people, the only minority group studied here. PPRs varied from 0.2 to 0.6, showing that the problem of underrepresented minorities is not unique to AD (Al Hadidi et al., 2022). 

Some AD trials have fallen even shorter. For example, in the aducanumab Phase 3 trials, six out of 1,102 people, or 0.5 percent, who took the approved dosage were black, for a PPR of around 0.04 based on U.S. population and prevalence data (Matthews et al., 2019).

Alas, like most problems in AD research, finding a solution will not be simple. Enrolling representative numbers of each racial/ethnic group may be insufficient because, as the FDA guidance notes, drug response or diagnostic markers may also vary by race.

Race and Biomarkers
Some studies suggest a difference in AD biomarkers such as p-tau and total tau by race, although the findings are typically from small cohorts and as yet inconclusive (Jan 2019 news). For example, Schindler had previously reported that African Americans are more likely than white people to carry TREM2 variants linked to low sTREM2 in cerebrospinal fluid (Apr 2021 news).

In a new study, Schindler matched 76 pairs of African American and Caucasian participants by age, APOE genotype, and cognitive status. All were enrolled in aging studies at WashU. Their average age was 68 and nearly all were cognitively healthy. The researchers found that participants had similar levels of plasma p-tau181, p-tau231, and neurofilament light regardless of race, but black participants had a higher plasma Aβ42/Aβ40 ratio. This higher ratio correlated with CNS markers of amyloidosis, with black people having higher CSF Aβ42/Aβ40 and lower amyloid PET signals than whites.

In other words, black participants had less plaque in their brain. Plasma Aβ42/Aβ40 faithfully reflected this, but other plasma markers did not. P-tau181 and 231 and NfL were higher than expected for the amount of plaque present. This suggests standard p-tau cutoffs might produce false positives in black patients, the authors noted.

Why might black participants have less plaque in their brains than whites with similar cognition? Because black participants were typically recruited by community outreach, while whites were more likely to seek out the study due to a family history of dementia, the black cohort may have less genetic risk for AD, Schindler noted. Polygenic risk scores for this study population to confirm this hunch were unavailable. Hence, the cognitive problems in black participants might be due to issues other than amyloidosis, such as cerebrovascular disease. The black participants in this study had much higher rates of hypertension and diabetes than did whites.

“Diagnostic tests should be studied in diverse cohorts, or we will risk disproportionate misdiagnosis of underrepresented groups,” Schindler told Alzforum.

As blood-based biomarkers are poised to enter clinical use, and researchers are studying them more broadly in aging populations, they are finding that factors other than race/ethnicity can sway them, as well. For example, a recent study correlated chronic kidney disease with elevated AD biomarkers (Syrjanen et al., 2021). 

The FDA guidance seeks to facilitate more diversity in clinical studies of drugs, biologics, or medical devices. The guidance suggests sponsors develop a “Race and Ethnicity Diversity Plan” that they discuss with the FDA no later than the end of Phase 2. The plan should note any evidence that the safety or efficacy of the drug at hand may vary by race. This evidence could include reports from the literature or pharmacokinetic and pharmacodynamic studies done by the sponsor, the FDA noted.

The FDA suggests sponsors set enrollment goals for underrepresented populations that reflect the distribution of the disease. The diversity plan should describe how these participants will be enrolled and retained, and how the data analysis will characterize safety, efficacy, and dosage in these subgroups (see table above).

The guidance is non-binding. It is also vague on specifics. The agency will take comments on this draft guidance through June 13. Meanwhile, the Alzheimer’s Association is hosting a webinar May 23 on how to improve diversity in AD trials.

In the April 27 New England Journal of Medicine, David Blumenthal at the Commonwealth Fund, New York, and Cara James at Grantmakers in Health, Washington, D.C., note that one barrier to enrolling diverse trials is the lack of reliable data on race. Many databases use old definitions of racial and ethnic groups that poorly capture the current composition of the U.S. population. The federal government could standardize definitions, fund database updates, and educate providers on why and how to collect accurate data, the authors suggest.—Madolyn Bowman Rogers

Comments

1. • Joyce Balls-Berry
     Washington University
   • Posted: 02 May 2022

   Another consideration in improving trial diversity is social and structural determinants to health that impact early screening and diagnosis of patients who are historically minoritized. A consideration is the patient's access to appropriate and quality care and ability to receive referrals for ADRD screening and treatment. A key aspect that may impact ADRD diagnosis in patients of color may also relate to discrimination, socioeconomic status, insurance status, educational level, and living environment. I often wonder if we should consider measuring health disparities in the same way that we measure other biomarkers for ADRD.

   View all comments by Joyce Balls-Berry

Make a Comment

To make a comment you must login or register.

References

News Citations

1. In the United States, Racial Disparities in Dementia Risk Persist 3 Dec 2020
2. Do Alzheimer’s Biomarkers Vary by Race? 11 Jan 2019
3. TREM2 Variants and CSF sTREM2 Levels Differ by Race 7 Apr 2021

Conference Citations

1. Increasing Sustained Diversity in Alzheimer's Disease and Related Dementia Clinical Trials - ISTAART Webinar 15 Apr 2022

Paper Citations

1. Mayeda ER, Glymour MM, Quesenberry CP, Whitmer RA. Inequalities in dementia incidence between six racial and ethnic groups over 14 years. Alzheimers Dement. 2016 Mar;12(3):216-24. Epub 2016 Feb 11 PubMed.
2. 2021 Alzheimer's disease facts and figures. Alzheimers Dement. 2021 Mar;17(3):327-406. Epub 2021 Mar 23 PubMed.
3. Mehta KM, Yeo GW. Systematic review of dementia prevalence and incidence in United States race/ethnic populations. Alzheimers Dement. 2017 Jan;13(1):72-83. Epub 2016 Sep 4 PubMed.
4. Al Hadidi S, Schinke C, Thanendrarajan S, Zangari M, van Rhee F. Enrollment of Black Participants in Pivotal Clinical Trials Supporting US Food and Drug Administration Approval of Chimeric Antigen Receptor-T Cell Therapy for Hematological Malignant Neoplasms. JAMA Netw Open. 2022 Apr 1;5(4):e228161. PubMed.
5. Matthews KA, Xu W, Gaglioti AH, Holt JB, Croft JB, Mack D, McGuire LC. Racial and ethnic estimates of Alzheimer's disease and related dementias in the United States (2015-2060) in adults aged ≥65 years. Alzheimers Dement. 2019 Jan;15(1):17-24. Epub 2018 Sep 19 PubMed.
6. Syrjanen JA, Campbell MR, Algeciras-Schimnich A, Vemuri P, Graff-Radford J, Machulda MM, Bu G, Knopman DS, Jack CR Jr, Petersen RC, Mielke MM. Associations of amyloid and neurodegeneration plasma biomarkers with comorbidities. Alzheimers Dement. 2021 Sep 27; PubMed.

External Citations

1. draft guidance

Further Reading

News

1. Do African-Americans Have More, or Different, Alzheimer’s Disease? Too Little Data to Tell 26 Oct 2018
2. Largest Alzheimer GWAS in African Americans Finds New Variants 22 Oct 2020