fetching data ... 
Background: Behçet’s syndrome (BS) is a complex multisystem inflammatory disease of unknown etiology. Given the extreme heterogeneity in patient manifestations and disease course, the 2018 EULAR recommendations emphasized the need for precise patient classification to enable individualized management strategies. While traditional cluster analysis introduced new insights in identifying patient subsets, its deterministic approach may not fully capture the broad complexity of BS.
Objectives: The primary objective of this study was to define the clinical phenotypes of BS using latent class analysis (LCA), a probabilistic model-based clustering method that identifies hidden classes based on unobserved patterns. The secondary aim was to explore sex-related differences in clinical manifestations and assess therapeutic needs across the identified classes.
Methods: We conducted a cross-sectional, observational, single-center study, including all adult patients followed in our department for BS (ISBD and/or ISG criteria) between 2012 and 2022, with a target sample size of 500 patients. LCA was performed using clinically relevant indicators (sex, oral ulcers, genital ulcers, skin lesions, articular involvement, and major organ involvements). Models were compared based on class number, fit indices, class separation, individual probabilities for class membership, and class size. The clustering algorithm was run with 100 random initializations, testing solutions with 2 to 6 classes. The final model was selected based on both clinical relevance and statistical performance.
Results: A total of 553 patients (409 males, 144 females) were enrolled, with a male-to-female (M/F) ratio of 2.8:1, and a mean age at BS onset of 30 ± 11 years. The main clinical manifestations comprised oral (95%) and genital (85%) ulcers, papulopustular lesions (40%), and arthralgias (30%). Uveitis (48%) and vascular lesions (41%) represented the most frequent major organ involvements, followed by the neurological (13%), cardiac (7%), and gastrointestinal (4%) systems. Patients were mostly treated with corticosteroids (85%), azathioprine (70%), cyclophosphamide (48%), and TNF-a inhibitors (8%). Anticoagulants and antiplatelet agents were prescribed in respectively 33% and 12% of cases. Mortality was recorded in 12 patients (2%).
Using LCA, five latent classes (C1-C5) were identified:
C1 (n=214; 39%) – “Vascular type”: This was the largest class, featuring the highest male predominance with a M/F ratio of 4.3:1 and a mean age at BS onset of 29 ± 10 years. All patients presented vascular involvement (80% venous, 36% arterial), with the highest prevalence of cardiac lesions (12%).
C2 (n=171; 31%) - “Ocular type”: this class exhibited the second highest male predominance (4:1), with a mean age at BS onset of 30 ± 11 years. It was characterized by 100% uveitis and frequent mucocutaneous lesions.
C3 (n=40; 7%) - “Neurological type”: the M/F ratio was 2.3:1 and the mean age at BS onset was 30 ± 10 years. Neurological parenchymal involvement was observed in all patients with 40% of concomitant uveitis, mostly of the posterior segment (83%).
C4 (n=98; 18%) - “Skin-mucosa and articular type”: the highest female distribution was observed in this class with a M/F ratio of 1.1:1. The mean age at BS onset was 31 ± 12 years. All patients reported oral and genital ulcers, with the highest prevalence of papulopustular lesions (54%) and articular involvement (48%).
C5 (n=30; 5%) - “Uncertain BS”: This was the smallest group, with a less pronounced male predominance (M/F ratio=1.5:1), and a later BS onset (38 ± 14 years). This class was marked by frequent uveitis (60) and vascular involvement (48%), with the fewest mucocutaneous lesions (12% oral ulcers, no genital ulcers, and 4% papulopustular lesions).
Treatment requirements varied across classes, with the highest doses of corticosteroids prescribed in the neurological (median=30mg/d) and ocular (median=20mg/d) classes. Conventional immunosuppressants were administered in the neurological (80%), vascular (76%), and ocular (75%) classes, followed by the “uncertain BS” (62%). TNF alpha inhibitors were mainly used in the ocular (17%) and uncertain BS (10%) (p<0.001).
Conclusion: This study is the first to apply LCA for the identification of clinical phenotypes in BS, offering a probabilistic patient classification better suited to address the complexity and heterogeneity of BS. Five latent classes were identified, namely the “vascular type”, “ocular type”, “neurological type”, “skin-mucosa and articular type”, and “uncertain BS”. Significant sex disparities and varying therapeutic needs were detected across classes. These findings provide critical guidance for advancing precision medicine in BS and, ultimately, improving patient outcomes.
REFERENCES: [1] Ilgen U. Cluster analysis as a clinical and research tool in Behçet’s syndrome. Curr Opin Rheumatol. 2024;36:3–8.
[2] Zou J, Luo J-F, Shen Y, Cai J-F, Guan J-L. Cluster analysis of phenotypes of patients with Behçet’s syndrome: a large cohort study from a referral center in China. Arthritis Res Ther. 2021;23:45.
[3] Soejima Y, Kirino Y, Takeno M, Kurosawa M, Takeuchi M, Yoshimi R, et al. Changes in the proportion of clinical clusters contribute to the phenotypic evolution of Behçet’s disease in Japan. Arthritis Res Ther. 2021;23:49.
Acknowledgements: NIL.
Disclosure of Interests: None declared.
© The Authors 2025. This abstract is an open access article published in Annals of Rheumatic Diseases under the CC BY-NC-ND license (