Background/Purpose: Behçet’s syndrome (BS) is formally diagnosed using the International Study Group (ISG) criteria, where recurrent oral ulceration and any two other symptoms (recurrent genital ulceration, uveitis, skin lesions and pathergy positivity) are required. The allowance of various symptomology in the ISG criteria has led to the reporting of varied manifestations, and differences in clinical presentation can complicate BS diagnosis, especially in areas where the disease prevalence is low. The purpose of this study was to explore clinical BS symptoms present at initial patient visit that are predictive of ISG criteria diagnosis at follow-up.

Methods: Data from consecutive patients monitored in outpatient clinics in New York and Amsterdam were abstracted. Patients were included if diagnosis at initial visit was “suspected” or “probable BS”; patients given a formal diagnosis by ISG criteria at initial visit or a non-BS diagnosis at initial visit were excluded. Demographic data, including ancestry/ethnicity, clinical symptoms, duration of symptoms and RAPID3 were abstracted from initial visit, with follow-up ISG status (defined as meeting criteria ISG+ vs not meeting criteria ISG-) abstracted from last visit.

Univariable logistic regression was used to screen initial visit clinical features and symptoms with follow-up ISG status. All variables that passed screening at P ≤ 0.10 were included in the final multivariable model². 

Results: 189 patients were included: 169 from NY and 20 from Amsterdam. 71 (37.6%) patients were classified as ISG+ with an average of 9.4 years (± 8.3 years) of symptoms. Age, gender, ethnicity, duration of symptoms at enrollment, duration of follow up as well as RAPID3 and almost all clinical manifestations at baseline were comparable between ISG+ And ISG- patients.

Presence of morning stiffness, family history of BS, genital ulceration, labial ulceration, skin lesions, eye disease and retinitis were each identified in the univariable model as being possibly associated with prevalence of ISG+. The final multivariable model did not include correlated symptoms (i.e. genital and labial ulceration as well as eye disease and retinitis).  In the final model, presence of morning stiffness, genital ulcers, skin lesions, and eye disease were associated with increased odds of ISG+, adjusting for age, symptom duration and family history (Figure). Area under the curve was 0.718, indicating acceptable predictive capability of the final model.

Conclusion: Based on our data, over a third of patients with suspected or probable Behçet’s developed new manifestations over time that led to classification as ISG+ Behçet’s. Despite development of these new manifestations, the presence of morning stiffness, genital ulcers, skin lesions, and eye disease at initial visit were independently associated with significantly higher odds in developing ISG+ Behçet’s during follow up.