Spondyloarthritis in Indigenous Populations: How Classification Shapes Epidemiology

The classification of spondyloarthritis (SpA) has undergone profound changes over the last 2 decades. In 2009, the Assessment of SpondyloArthritis International Society (ASAS) proposed classification criteria for axial spondyloarthritis (axSpA), introducing the concept of nonradiographic axSpA (nr-axSpA),¹ which expanded the spectrum of disease beyond the traditional definition of ankylosing spondylitis (AS; now termed “radiographic axSpA”).² Recently, ASAS-Spondyloarthritis Research and Treatment Network (SPARTAN) revised classification criteria have been proposed, with greater emphasis on imaging specificity and the exclusion of disease mimics.³ Another important aspect of these new criteria is the requirement that they be applied only to patients previously diagnosed with axSpA, avoiding their indiscriminate application to patients without symptoms of axSpA or to those with clinical abnormalities that do not correspond clinically to axSpA. In this context, a prior diagnosis of axSpA by a rheumatologist becomes very important, especially in countries where standardization of imaging assessment is not yet fully established. These refinements have improved our ability to classify patients for research purposes, but they have also introduced important challenges when interpreting epidemiologic studies performed across different time periods.

The prevalence of radiographic axSpA varies widely, being higher in regions where HLA-B27 is more frequent, particularly among native populations in Alaska, Canada, and Siberia.⁴ In this issue of The Journal of Rheumatology, Morrison and colleagues report contemporary prevalence estimates of SpA among Alaska Native and American Indian (AN/AI) peoples of Alaska.⁵ Their study found an age-adjusted prevalence of 355 per 100,000 adults, with psoriatic arthritis (PsA) representing the most common subtype, followed by AS (107 per 100,000 adults), undifferentiated SpA (uSpA; 59 per 100,000 adults), and reactive arthritis (ReA; 23 per 100,000 adults).⁵ Although the overall prevalence remains higher than that reported in many other populations worldwide, these findings differ substantially from studies conducted in the same population during the 1980s and 1990s, when uSpA and ReA represented a much larger proportion of cases.^6-8 These discrepancies are possibly linked to the diagnostic criteria used in these studies^4,6-8 and the nature of the database analyzed.

The most important message from the study by Morrison et al may not be the prevalence estimate itself, but rather what it reveals about the evolution of disease concepts in SpA.⁵ Historically, the category of uSpA encompassed a heterogeneous group of patients that shared characteristic features of the SpA spectrum but did not fulfill criteria for a more specific diagnosis. With the development of the ASAS criteria for axSpA¹ and peripheral SpA⁹ and the widespread adoption of the Classification Criteria for Psoriatic Arthritis (CASPAR) for PsA,¹⁰ many patients previously classified as having uSpA can now be assigned to more specific disease categories. Consequently, a reduction in the prevalence of uSpA over time should not necessarily be interpreted as a true epidemiologic decline but may instead reflect improvements in disease recognition and classification. In centers where axial magnetic resonance imaging (MRI) is not yet available for routine diagnosis, many cases of nr-axSpA may be diagnosed as uSpA.

The analysis of the Morrison et al study highlighted this issue, leading us to question whether contemporary classification criteria should have been retrospectively applied to all patients included in their study.⁵ Such concerns are understandable, particularly given the substantial advances in classification criteria and imaging technology over recent decades. However, epidemiologic studies based on existing medical records face important practical limitations. Classification criteria are designed primarily to ensure homogeneous populations in clinical research and were not developed as diagnostic criteria. In contrast, prevalence studies seek to estimate the burden of disease in real-world populations, where complete clinical, laboratory, and imaging data are not always available. Under these circumstances, physician-confirmed diagnoses supported by detailed chart review may represent an appropriate and feasible methodology.

The distinction between diagnosis and classification has become particularly relevant in axSpA. The introduction of MRI transformed the recognition of early inflammatory disease and allowed identification of patients without definite radiographic sacroiliitis.² At the same time, growing awareness of the limited specificity of some imaging findings has led to renewed attention to disease mimics and to the importance of structural lesions when interpreting MRI studies. The recent ASAS-SPARTAN update reflects this evolution and emphasizes the need for careful integration of clinical and imaging information. Although these advances improve classification performance, they also complicate direct comparisons between historical and ongoing prevalence studies.

Suspected cases of axSpA that do not present radiographic confirmation of sacroiliitis should undergo MRI to evaluate inflammatory and structural changes compatible with nr-axSpA. For this purpose, criteria for image acquisition and analysis must follow an established protocol and be evaluated by musculoskeletal imaging specialists.¹¹ The update of the classification criteria for axSpA³ should be accompanied by a strategy to strengthen the definition of a positive image of the disease in all medical specialties involved in its diagnosis. The diagnostic interpretation of axSpA should also rule out frequent mimicking factors, such as previous pregnancies, overweight, advanced age, and physically demanding jobs.¹² Since the characteristic bone marrow edema has low specificity for axSpA, being responsible for a significant number of misdiagnoses in recent years, the evaluation of associated structural lesions has become important for establishing the diagnosis of axSpA.¹³

The findings reported by Morrison et al⁵ are especially important because they address a population that has historically played a central role in our understanding of SpA epidemiology. Indigenous populations of Alaska, Canada, and other circumpolar regions have long been recognized for their high prevalence of HLA-B27 and their increased burden of SpA. The persistence of prevalence rates that remain above those observed in many other populations supports the continued importance of genetic susceptibility in disease development. At the same time, the shift from predominance of uSpA and ReA toward greater recognition of PsA illustrates how advances in disease classification and clinical practice can reshape the epidemiologic profile of a disease.

The study by Morrison et al also highlights the importance of continuing epidemiologic research in historically underrepresented populations.⁵ Indigenous communities remain insufficiently represented in rheumatology research despite their unique genetic background, geographic challenges, and healthcare disparities. As such, this study⁵ reminds us that epidemiology is inseparable from the concepts used to define disease. As classification schemes continue to evolve, interpreting prevalence data requires consideration not only of populations and genetic backgrounds but also of the diagnostic frameworks through which disease is recognized. In this regard, the work by Morrison et al⁵ provides important contemporary data while highlighting the ongoing challenge of reconciling clinical diagnosis, classification criteria, and epidemiologic research in SpA.

Future studies should incorporate standardized clinical assessments, up-to-date classification criteria, and specific imaging protocols to better understand the relative contributions of genetics, environment, and healthcare systems to disease expression in different populations around the world.

• Copyright © 2026 by the Journal of Rheumatology