Validation of the International Criteria for Behçet's disease (ICBD) in Iran

Aim: To evaluate the performance of the new International Criteria for Behçet's Disease (ICBD) in Iran. The ICBD was created in 2006 to replace the International Study Group (ISG) criteria (created in 1990). In this study, 14 existing diagnosis/classification criteria (Curth, Hewitt, Mason and Barnes, Japan original, Hubault and Hamza, O'Duffy, Cheng and Zhang, Dilsen, Japan revised, International Study Group, Iran traditional, Iran Classification Tree, revised Dilsen, Korea), were evaluated and compared to ICBD by calculating their sensitivity, specificity and accuracy. Methods: All patients from the Behçet's Disease registry (6128) and controls (3400), up to January 2008, entered the study. The diagnosis was clinical and not by any of the above-mentioned criteria. Sensitivity, specificity, accuracy (percent agreement), and 95% confidence interval were calculated. The performance of ICBD and ISG were compared by the McNemar test. Results: The sensitivity was 98.2% for ICBD and 78.1% for ISG (P < 0.0001). The specificity was 95.6% for ICBD and 98.4% for ISG (P < 0.0001). The accuracy was 97.3% for ICBD and 85.5% for ISG (P < 0.0001). Discussion: In Iranian patients, ICBD has 20.1% higher sensitivity and 11.8% higher accuracy than ISG, while the specificity was only 2.8% lower. ICBD was better optimized than ISG (difference between sensitivity and specificity 2.6. % vs 20.3%). Conclusion: ICBD has by far better performance than ISG in Iran. The difference is even more prominent in Iranian patients than in the cohort of the international patients upon whom the criteria were created.