Recruitment of Minority Patients in Metabolic Dysfunction-Associated Steatotic Liver Disease Clinical Trials

Abstract

Purpose

Metabolic dysfunction-associated steatotic liver disease (MASLD) is the second leading indication for liver transplant in the United States, but the prevalence is unevenly distributed across ethnic groups with 22.3% of Hispanic patients affected. This study aimed to assess racial and ethnic diversity among MASLD clinical trials.

Methods

We performed a meta-analysis of 91 randomized controlled trials (RCTs) of MASLD therapies in the US and Canada, and multinational RCTS involving the US from 2005 to 2024.

Results

78 studies (85.7%) reported racial data and only 56 (61.5%) included ethnicity data. Meta-analysis revealed a pooled prevalence of 78.8% (95% CI 72.9 – 84.6, I² = 97.0%) in White patients, 6.2% (95% CI 4.1 – 8.3, I² = 94.1%) in Asian patients, 2.5% (95% CI 1.6 – 3.5, I² = 70.2%) in Black patients, and 31.7% (95% CI 26.8—36.6, I² = 96.2%) in Hispanic patients. This prevalence was higher than the reported national Hispanic prevalence, and Hispanic enrollment also increased over time from 20.3% (2009 – 2014), 30.8% (2015–2019), to 40.0% (2020–2024).

Conclusion

Our results show that compared to previous efforts, clinical trials are increasingly including Hispanic patients, yet they still need to increase inclusion of ethnicity data. Since MASLD disproportionately affects Hispanic patients, it is imperative that clinical trials make a targeted effort to diversify patient recruitment.

Introduction

Metabolic dysfunction-associated steatotic liver disease (MASLD) is increasingly prevalent worldwide, affecting as much as 30% of the global population [1]. However, it is not proportionately distributed across the United States population, as Hispanic patients have the highest prevalence among US ethnic groups (22.3%) [2]. The cause of this high prevalence is likely multifactorial, related to genetics in addition to diet and other social determinants of health.

Treatment options remain limited; currently resmetirom and semaglutide are the only FDA-approved medications for the treatment of metabolic dysfunction-associated steatohepatitis [3, 4]. While numerous clinical trials investigating other therapeutics are underway, it is imperative that recruited patients reflect the populations most affected by MASLD and are most likely to benefit from treatment. A previous meta-analysis of MASLD therapy studies in 2020 showed that only 11.6% of reported participants were of Hispanic heritage [5]. The study also noted that 73% of the analyzed studies reported ethnicity/race data, with only 45% including information regarding inclusion of Hispanic patients. Characterization of the ethnic and racial makeup of study cohorts has historically been limited by underreporting.

The purpose of this study was to characterize the racial and ethnic composition of patient enrollment in MASLD randomized clinical trials (RCTs) from 2005 to 2024.

Methods

A systemic review of RCTs for MASLD therapies spanning from 2005 to 2024 was conducted. The studies were found through a search of the PubMed (United States National Institutes of Health, Bethesda, MD, United States) database using the following terms: “metabolic dysfunction associated liver disease,” “metabolic dysfunction associated steatohepatitis,” “fatty liver disease,” “non-alcoholic steatohepatitis” and “non-alcoholic fatty liver disease.”

Only English language studies assessing therapeutic interventions for patients with MASLD were included. All patients included in the study were adults aged 18 and older. All included studies were conducted in the United States or Canada, or were multinational RCTs that included the United States. The aggregate list of studies collected from PubMed was examined independently by three reviewers, and all included studies were selected by consensus. Demographic data regarding patient enrollment including age, sex, ethnicity, and race was collected when available, as was each study’s funding source.

Meta-analysis was conducted to determine the pooled prevalence of Hispanic patients enrolled in the collected studies, and this prevalence was compared to the reported national Hispanic prevalence in the United States.

We performed an analysis of the studies that provided explicit data regarding Hispanic ethnicity. We performed a random-effects meta-analysis using the meta package in R and the metagen function. We employed a random-effects model to account for both within-study sampling error and between-study heterogeneity. The raw proportions (sm = “PR”) were used as the summary measure, with between-study variance estimated using the restricted maximum likelihood (method.tau = “REML”). To account for uncertainty in the estimation of the raw proportions, we applied the Hartung-Knapp adjustment (method.random.ci = “HK”) for the random-effects model, which generally results in more conservative (wider) confidence intervals, reducing the risk of Type I errors. Study-specific estimates were weighted by the inverse of their variance.

Results

The initial search on the PubMed database produced 1,011 results (Fig. 1). 920 studies were excluded because they did not meet inclusion criteria. A total of 91 studies with 16,111 patients were included in the meta-analysis.

Fig. 1.

PRISMA flow diagram

The majority of trials (n = 57 trials; 62.6%) were conducted in the US, 4 trials (4.4%) were conducted in Canada, 2 (2.2%) were conducted in the US and Canada, and 28 (30.8%) were conducted in multiple countries that included the US and Canada. The trial therapeutics included medications (n = 84), lifestyle interventions (n = 6), and fecal microbiome transplant (n = 1). When reported, the median age of study participants was 52.6 years (range 24.1—69.7) with 58% female subjects (Table 1).

Table 1.

Characteristics of the RCTs (n = 91) included in the meta-analysis

78 studies (85.7%) included data on race and 56 (61.5%) included ethnicity data. The majority of trials (n = 58, 63.7%) were supported by pharmaceutical and industry companies, 20 trials (22.0%) by academic funding, and 13 trials (14.3%) by both industry and academic funding. Of the trials that did not report ethnicity data, 74.3% were either partially or wholly funded by industry. 14 (53.8%) of these studies were conducted in the US and Canada, and 12 (46.2%) were multinational that included the US.

Among those with ethnicity data, the pooled prevalence of White patients in the meta-analysis was 78.8% (95% CI 72.9 – 84.6, I² = 97.0%), 6.2% in Asian patients (95% CI 4.1 – 8.3, I² = 94.1%), and 2.5% in Black patients (95% CI 1.6 – 3.5, I² = 70.2%). Meta-analysis revealed a pooled prevalence of 31.7% (95% CI 26.8—36.6, I² = 96.2%) Hispanic patients (Fig. 2).

Fig. 2.

Pooled prevalence of Hispanic subjects in RCTs conducted in the US, Canada, and in multi-national RCTs

When the studies with ethnicity data were limited to the US and Canada only (n = 40), the pooled prevalence of White patients was 75.7% (95% CI 67.4—84.0, I² = 97.4%), Asian patients 4.3% (95% CI 2.4—6.3, I² = 67.1%), Black patients 5.2% (95% CI 1.9—8.6, I² = 73.5%), and Hispanic patients 33.5 (95% CI 26.9—40.2, I² = 96.8%) (Fig. 3). This prevalence of Hispanics was higher than the reported national Hispanic prevalence of 22.9% in the US.

Fig. 3.

Pooled prevalence of Hispanic subjects in RCTs conducted in the US and Canada

Discussion

It is projected that by 2045, nearly half of the US population will be comprised of ethnicities other than non-Hispanic White [6]. Under-represented populations who have a larger burden of disease, such as Hispanics, should be appropriately represented in MASLD RCTs. In our analysis of the studies that reported ethnicity data, enrollment of Hispanics has notably increased over time. In US and Canada, we found it had increased from 20.3% (2009 – 2014), to 30.8% (2015 – 2019), to most recently 40.0% (2020 – 2024) (Figs. 4). This was an encouraging trend since Hispanic patients in the US have the highest prevalence of MASLD. [7]

Fig. 4.

Hispanic recruitment in MASLD RCTs over time

Our meta-analysis also demonstrates the persistent under-reporting of race and particularly ethnicity data of participants. While 78 studies (85.7%) reported race, only 56 (61.5%) reported ethnicity, usually categorized as Hispanic or non-Hispanic. Our findings are consistent with previous studies showing significant underreporting of race and ethnicity data in major gastroenterology journals [8]. Of the publications that reported race, only White and Black races were most commonly included.

Accurate reporting of demographic data in clinical trials is needed to ensure that the findings are generalizable to diverse populations. This concept is illustrated by the experience with interferon-based treatment for hepatitis C, where early clinical trials predominantly included White patients [9]. However, real-world practice and subsequent studies revealed that Black patients had significantly lower sustained virologic response rates (26—28% vs 39—52% in White patients). [10, 11]

Notably, nearly 75% of trials that did not report ethnicity data were found to be funded by pharmaceutical companies. This is consistent with previous studies that show many industry-funded trials either fail to report race or ethnicity data or significantly underrepresent Black and Hispanic patients [12–14]. Though the underlying root causes behind this association are unclear, it is imperative to implement standardized requirements for reporting ethnicity and racial data in all clinical trials. In 2022, the FDA issued guidance encouraging companies to recruit underrepresented populations, implement systematic plans to enhance diversity in clinical trials, and broaden eligibility criteria [15]. While this guidance was not available for a period of time, it is currently publicly available on the agency’s website. However, its enforcement and long-term durability remain uncertain in the current sociopolitical climate.

The pooled prevalence estimates should be interpreted as descriptive summaries rather than precise population-level estimates. In our meta-analysis, the pooled prevalence of Black patients represented in US and Canadian clinical trials was 5.24% (95% CI 1.92—8.55, I² = 74%). This is much lower than the known prevalence of the disease in this population (approximately 18%) [16]. Additionally, it is worth noting that the representation of Black patients is also much lower than the 14.4% of Black people in the general US population [17]. In a recent study comparing Non-Hispanic White and Non-Hispanic Black patients with MASLD, Black patients had lower rates of biopsy-proven advanced fibrosis than White patients, but they exhibited more severe lobular inflammation and a higher prevalence of MASH. [18] There may be some protective mechanism at play that necessitates a higher threshold of required necroinflammation required for progression to advanced fibrosis in Black patients compared to other ethnic groups, which needs further investigation. Ultimately, the inclusion of Black patients remains crucial for understanding potential differences in therapeutics for MASLD so that they can be appropriately applied to the treatment of this affected population.

Overall, it is important to keep in mind that ethnic representation in clinical trials cannot be compared one-to-one with the prevalence in the general population, per se. The composition of participants in each clinical trial is not only affected by its specific inclusion and exclusion criteria but also the geographic locations and referral patterns of the participating centers. For example, study centers located in areas with a much larger Hispanic population (e.g. San Antonio, Texas) will understandably have a higher representation of this group in their clinical trials compared to those with a smaller Hispanic population in other parts of the country.

It is also crucial to acknowledge the significant upstream role that various social determinants of health and structural inequities play here. For example, most clinical trials are conducted at large tertiary academic centers in urban areas of the U.S. but they may not necessarily be within reach of neighborhoods where the highest-risk populations live. Minority patients may be more likely to receive care in community-based primary care settings in rather than hepatologists at academic centers who are either directly involved in clinical trials or are generally more aware of ongoing studies to which they can refer patients. Trial site selection itself represents a structural barrier that skews towards patients with the means to live near or travel to these centers. Also, clinical trials tend to have strict exclusion criteria which screen out patients with significant comorbidities. Because Black and Hispanic populations have higher rates of certain metabolic conditions such as poorly controlled diabetes [19], which is often excluded in MASLD clinical studies, the eligibility criteria may inadvertently but disproportionately exclude a majority of underrepresented patients affected by the very disease that is being studied.

There are also patient-level factors that impact enrollment in clinical trials. In addition to mistrust of clinical research, logistical constraints such as time, travel, and reimbursement required to participate [20], many minority patients are unaware of their MASLD diagnosis and therefore may not know they are eligible to participate in studies. In a systematic review of NHANES data (2011 – 2016), although 90% of MASLD patients had seen a healthcare provider in the prior year, only 5.1% were aware of their diagnosis, independent of education level [6]. It is unclear if this unawareness reflected a failure of the provider to diagnose or communicate the condition, or patient misunderstanding or recall. Improving awareness of and communication about MASLD may help facilitate more diverse patient enrollment in clinical trials.

While our meta-analysis has several strengths, including a comprehensive search strategy and large number of trials, it was limited by significant heterogeneity, as indicated by the I² statistic ranging from 70 to 94.6% for the pooled prevalence of each racial and ethnic group. There was a large variation in the sizes of the RCTs, with the smallest study having n = 8 and the largest with n = 1769. Additionally, since 38.5% of studies did not report any ethnicity data, it is possible that the true representation of Hispanics in these trials may actually be lower.

High I² values are indicative of substantial heterogeneity among studies. While concerning in the context of predicting an effect size, here it is mainly reflective of the broad observed heterogeneity in inclusive of Hispanic individuals in the reported studies. In comparison with other studies that look at reporting and representation of race and ethnicity, high I² are also observed for Hispanic ethnicity. [21, 22]

In conclusion, while the inclusion of Hispanic patients in MASLD clinicals trials has improved over time, gaps remain in the overall reporting and representation of minorities. Further efforts are needed to not only increase enrollment but also address the complex structural and systemic barriers that contribute to the underrepresentation of minority populations in clinical research. Finally, journals should require the reporting of race and ethnicity data for publication to ensure not only transparency but also accountability. Science must continue to ensure that the study populations accurately reflect those most affected by the disease and those most likely to benefit from emerging therapies.

Author Contributions

SEP and YW conducted the data collection. SEP, YW, and SP wrote the manuscript text. ALH performed the statistical analysis.

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Conflicts of interest

The authors declare no competing interests.