What Do We Actually Know About Autoimmune Disease in Latino Populations?

Part 1 of a series on evidence, measurement, and uncertainty in Latino health.

Do autoimmune diseases affect Latino patients differently?

The interesting question is not whether an ethnic association exists, but what causal information, if any, that association contains.

To answer it, we first have to clarify three things: what counts as an autoimmune disease, which Latino population we are studying, and what outcome we mean by “different.”

The first is less straightforward than the label suggests. Autoimmune diseases are defined through evolving combinations of clinical phenotype, serology, pathology, and classification criteria that do not always map neatly onto molecular endotypes. Patients can sit near diagnostic boundaries, accumulate manifestations over time, or share pathogenic pathways across nominally different diseases. I have written previously about the implications of this diagnostic uncertainty for drug development. [1]

The second problem is equally important: Hispanic/Latino is not a biologically homogeneous category. Mexican Americans are not Puerto Ricans, and self-identified ethnicity does not map cleanly onto genetic ancestry, environmental exposure, migration history, or healthcare context.

And the third is the endpoint. Prevalence is not severity; severity is not organ involvement; and none of these is the same as treatment response, disability, damage accrual, or mortality.

Only after defining all three can we ask the harder question:

What, if anything, explains an observed difference?

The deeper we look into the evidence, the less useful a single answer becomes.

And perhaps that is the first important finding.

“Autoimmune Disease” Is Not a Fixed Biological Category

Shared susceptibility does not necessarily imply shared mechanisms, phenotypes, or population effects.

A 2026 analysis integrating GWAS and functional genomic data across 15 autoimmune diseases illustrates why broad autoimmune categories can be misleading. Although 50.8% of susceptibility loci were associated with more than one disease, only 14.7% of the more finely resolved association signals were shared. Even apparently shared loci could implicate different regulatory mechanisms and immune cell types. [2]

That matters when we interpret population differences. A reproducible association in SLE cannot automatically be extrapolated to rheumatoid arthritis, multiple sclerosis, autoimmune hepatitis, or “autoimmunity” as a whole.

The scientifically useful question is not:

Is autoimmunity mechanistically different in Latinos?

It is:

For which disease, which phenotype, and which population is there a reproducible signal, and what evidence supports a mechanism?

“Latino” Is an Ethnicity Category, Not a Genetic-Ancestry Variable

Self-identified ethnicity captures population history and social context, but it does not map cleanly onto genetic ancestry.

A 2025 analysis of approximately 230,000 unrelated whole genomes from the All of Us Research Program found continuous gradients of genetic variation within self-identified racial and ethnic groups rather than discrete genetic clusters. Hispanic or Latino participants spanned a broad spectrum of African, Native American, and European ancestry. [3]

That has direct implications for biomedical interpretation.

Two patients can both self-identify as Latino while differing substantially in continental and subcontinental ancestry. Country of origin, nativity, migration history, environmental exposures, socioeconomic conditions, and healthcare context add further heterogeneity.

This does not mean ethnicity should not be measured. It means it should not be asked to explain more than it can.

When a study reports that “Hispanic patients had more severe disease,” the relevant question is not simply whether the association is statistically significant.

It is:

What variable, or combination of variables, is ethnicity standing in for?

Lupus Shows Reproducible Population Differences, but Their Magnitude Changes Across Settings

Studies across the U.S., Latin America, and Europe suggest that ancestry, healthcare context, socioeconomic conditions, and disease biology cannot easily be separated.

SLE is a particularly useful example because population differences appear across several clinically relevant outcomes. Pooled U.S. surveillance data estimated SLE incidence among Hispanic females at 6.8 per 100,000 person-years, compared with 5.7 among White females. [4] In the longitudinal California Lupus Epidemiology Study, Hispanic patients had higher rates of renal disease (HR 2.9, 95% CI 1.8–4.7) and multiorgan disease (HR 3.3, 95% CI 1.8–5.9) after SLE diagnosis than White patients. [5] Mortality surveillance from California also found an age-standardized mortality ratio of 3.9 among Hispanic/Latino patients with SLE relative to the corresponding general population, rising to 5.8 among Hispanic/Latina women.[6]

These findings establish that measurable population differences exist. They do not establish why.

In one informative comparison, 114 Hispanic patients with SLE living in Texas were compared with 619 Mestizo patients from Latin America with broadly similar Amerindian and African ancestry proportions. The U.S. cohort accumulated more damage and had more than twice the mortality hazard. This contrast is consistent with contributions beyond broad ancestry estimates, although the observational comparison cannot isolate genetic, healthcare, environmental, and selection-related explanations. [7]

European cohorts add a useful counterpoint. In a small 2024 single-center study, Latin American Hispanic and White European patients showed no major differences in clinical manifestations, autoantibody profiles, or treatment. [8] Yet, a larger 2026 Spanish cohort found similar major clinical manifestations but higher disease activity and lower remission rates among non-Spanish patients, most of whom were Latin American. [9]

Taken together, these studies argue against two simple conclusions: that population differences in SLE are predominantly genetic, or that they disappear when healthcare access is more equal.

A population difference is an observation. The setting in which it appears is part of the evidence.

The more informative question is which differences persist across settings strongly enough to warrant a search for underlying biology, and which attenuate when healthcare, socioeconomic, or environmental contexts change.

Rheumatoid Arthritis Shows a Different Signal

Population differences are also detectable in RA, but their interpretation is no less difficult.

A 2024 analysis of 9,363 patients in the U.S. CorEvitas Rheumatoid Arthritis Registry included 545 Hispanic patients. Adjusted Clinical Disease Activity Index (CDAI) scores (a composite of joint counts and patient/physician global assessments) were higher among Hispanic than White patients at both evaluated time points, and improvement over time was smaller among Hispanic patients. [10]

Differences can reflect inflammatory burden, comorbidities, pain perception and reporting, treatment history, therapeutic access, clinician assessment, or several of these simultaneously. 

And that is precisely why findings in lupus cannot simply be generalized to rheumatoid arthritis, or rheumatoid arthritis to all autoimmune disease.

Ethnicity Can Identify a Signal Without Explaining Its Mechanism

Even adjusted associations can combine ancestry, environment, healthcare, and social context in ways that are difficult to disentangle.

This is the central methodological problem.

An association that remains after statistical adjustment is not proof of an independent biological effect. Unmeasured social, environmental, or clinical factors may still contribute.

The National Academies addressed this directly in its 2025 report on race and ethnicity in biomedical research, recommending that investigators measure the variables relevant to the scientific question rather than treating racial or ethnic categories as biological proxies. [11]

A recent Latin American lupus cohort provides a concrete example. Among 1,341 patients, lower socioeconomic status, renal and serosal involvement predicted greater hospitalization risk, whereas antimalarial use and remission were associated with lower risk. [12]

These factors do not occupy separate biological and social compartments. Disease severity can affect treatment intensity. Access can affect treatment. Treatment affects disease activity. Socioeconomic conditions influence multiple points along the same pathway.

So when ethnicity remains associated with an outcome, we should ask:

Are we observing an ethnicity effect, or are we assigning explanatory meaning to ethnicity because the variables that would provide the explanation were not fully measured?

Ethnicity may tell us where to look.

It rarely tells us what mechanism we have found.

Why This Matters for Clinical Research and Drug Development

Population heterogeneity affects external validity, subgroup interpretation, biomarkers, and our ability to distinguish prognostic from predictive effects.

For drug developers, this distinction is consequential.

Suppose a self-identified Hispanic subgroup in a lupus trial enters with greater baseline renal involvement.

If that difference reflects molecular biology relevant to treatment response, it may represent true effect modification.

If it reflects later diagnosis, previous treatment exposure, access to specialists, or baseline disease severity, it may primarily be prognostic.

Those interpretations lead to different conclusions about stratification, biomarkers, subgroup effects, and generalizability.

FDA’s 2025 guidance on clinical-trial participation similarly emphasizes evaluating demographic and non-demographic baseline characteristics that may influence safety or effectiveness. [13]

The broader implication is simple:

Representation alone is not the scientific endpoint. Characterization is.

The question is not only who is enrolled, but which baseline characteristics matter, whether they are prognostic or predictive, and whether the available data allow us to distinguish between them.

So, What Do We Actually Know?

The evidence supports population differences in some autoimmune diseases, but not one biologically coherent entity called “Latino autoimmunity.”

Do autoimmune diseases affect Latino patients differently?

In some diseases, populations, and outcomes, measurable differences have been reported.

But the signal varies by disease, phenotype, cohort, geography, healthcare setting, and analytic approach.

And self-identified ethnicity rarely tells us why.

That leaves four questions I think are more useful than the original one:

Which disease?

Which Latino population?

Which outcome?

And what evidence allows us to move from association to mechanism?

Those questions will guide this broader series on what we actually know about Latino health.

Next month, for Hispanic Heritage Month, I will turn to a very different example where the same problem appears from another direction known as "The Hispanic Paradox."

There, the uncertainty shifts from which disease are we measuring? to an even more fundamental question:

What do we mean by health in the first place?

Stay tuned!

---

References

1. Mina-Osorio P. Diagnostic Uncertainty in Drug Development. In: Lockshin MD, Crow MK, Barbhaiya M, eds. Diagnoses Without Names: Challenges for Medical Care, Research, and Policy. Cham: Springer; 2022:45–57. doi:10.1007/978-3-031-04935-4_5.

2. Dang X, Wang FQ, Zhang C, et al. Identifying genetic and cellular connections and distinctions among 15 autoimmune diseases using an in-silico approach. Communications Medicine. 2026;6:235.

3. Gouveia MH, Meeks KAC, Borda V, et al. Subcontinental genetic variation in the All of Us Research Program: Implications for biomedical research. American Journal of Human Genetics. 2025;112(6):1286–1301.

4. Izmirly PM, Ferucci ED, Somers EC, Wang L, Lim SS, Drenkard C, Dall’Era M, McCune WJ, Gordon C, Helmick C, Parton H. Incidence rates of systemic lupus erythematosus in the USA: estimates from a meta-analysis of the Centers for Disease Control and Prevention national lupus registries. Lupus Science & Medicine. 2021;8(1). doi:10.1136/lupus-2021-000614.

5. Aguirre A, Izadi Z, Trupin L, Barbour KE, Greenlund KJ, Katz P, Lanata C, Criswell L, Dall’Era M, Yazdany J. Race, Ethnicity, and Disparities in the Risk of End-Organ Lupus Manifestations Following a Systemic Lupus Erythematosus Diagnosis in a Multiethnic Cohort. Arthritis Care & Research. 2023;75(1):34–43. doi:10.1002/acr.24892.

6. Gianfrancesco MA, Dall’Era M, Murphy LB, Helmick CG, Li J, Rush S, Trupin L, Yazdany J. Mortality Among Minority Populations with Systemic Lupus Erythematosus, Including Asian and Hispanic/Latino Persons — California, 2007–2017. MMWR Morb Mortal Wkly Rep. 2021;70:236–239. doi:10.15585/mmwr.mm7007a2.

7. Ugarte-Gil MF, et al. Disease features and outcomes in United States lupus patients of Hispanic origin and their Mestizo counterparts in Latin America: a commentary. Rheumatology. 2016;55(3):436–440.

8. Marri L, et al. Clinical Characteristics of Systemic Lupus Erythematosus in Caucasians and Latin American Hispanics: Data from a Single Tertiary Center. Autoimmune Diseases. 2024;2024:5593302.

9. Heras-Recuero E, García-Fernández A, Quiroga-Colina P, et al. Ethnic differences in disease activity among patients with systemic lupus erythematosus in a universal public healthcare system. Lupus. 2026;35(4):378–386. doi:10.1177/09612033261422634.

10. O’Brien J, Park SH, Blachley T, et al. Disparities in burden of disease in patients with rheumatoid arthritis across racial and ethnic groups. Clinical Rheumatology. 2024;43:921–927.

11. National Academies of Sciences, Engineering, and Medicine. Rethinking Race and Ethnicity in Biomedical Research. Washington, DC: The National Academies Press; 2025.

12. Pons-Estel GJ, Quintana R, Ugarte-Gil MF, et al. Predictors of first hospitalization due to disease activity and infections in systemic lupus erythematosus patients. Lupus. 2024;33(13):1492–1501.

13. U.S. Food and Drug Administration. Enhancing Participation in Clinical Trials: Eligibility Criteria, Enrollment Practices, and Trial Designs. Guidance for Industry. Final guidance. December 15, 2025.