Objective: To identify clusters of Takayasu arteritis (TAK) based on clinical features and their prognostic relevance.
Methods: Using agglomerative hierarchical clustering, clusters were defined by phenotype at presentation from 564 patients with TAK from three cohorts from India and Türkiye. Angiographic involvement, disease activity and Vasculitis Damage Index (VDI) at presentation, initiation of glucocorticoids or immunosuppressive agents, requirement for vascular procedures, and new-onset vascular complications were compared between clusters [unpaired Student's t-test or odds ratios (OR) with 95% CI]. Mortality rates [hazard ratios (HR) with 95%CI] were computed using Cox regression.
Results: Three clusters were identified (1 and 2, with sub-clusters 2A and 2B). Cluster 2 was younger and comprised relatively fewer women than cluster 1. Clinical features and angiography in cluster 1 reflected more frequent involvement of the aortic arch and its branches as opposed to a greater involvement of the abdominal aorta and its branches in cluster 2. Cluster 2B more often had pan-aortic disease and left subclavian involvement than 2A. Fewer patients in cluster 2A were initiated on glucocorticoids [OR adjusted for centre (aOR) 0.51 (95% CI: 0.31, 0.83)] or immunosuppressants [aOR 0.35 (95% CI: 0.20, 0.60)] or developed new-onset vascular complications [aOR 0.22 (95% CI: 0.07, 0.74)] than in cluster 1. Cluster 2B had higher VDI (3.98) than clusters 1 (3.50) or 2A (2.46) and required more frequent vascular interventions than cluster 1 or 2A (aOR 1.73, 1.88, P < 0.05). Survival was similar across clusters [HR 2A vs 1: 0.99 (95% CI: 0.43, 2.29), 2B vs 1: 1.20 (95% CI: 0.84, 1.70)].
Conclusion: Phenotypic clusters of TAK reflected distinct angiographic involvement with prognostic implications.
Keywords: Takayasu arteritis; cluster analysis; phenotype; prognosis; unsupervised machine learning.
© The Author(s) 2025. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For commercial re-use, please contact reprints@oup.com for reprints and translation rights for reprints. All other permissions can be obtained through our RightsLink service via the Permissions link on the article page on our site—for further information please contact journals.permissions@oup.com.