Background An association between Behçet's syndrome (BS) and HLA-B*51 is widely reported among many different ethnic groups. A few reports from Taiwan, Greece, and Japan were also published that indicated HLA-A*26 is associated with BS independently from HLA-B*51. Recently, we reported the strong associations between HLA-A*26 and BS from our cohort led by rheumatologist. But our previous study may have some limitations because the patients were diagnosed by clinical manifestations and expert opinions and therefore it might have contained the bias that HLA typing influenced the clinical diagnosis of BS.
Objectives In this study, we analyze the associations between HLA typing and BS among the patients who fulfilled the International Study Group criteria (ISG).
Methods Our cohort data consist of the clinically diagnosed BS patients seen at tertiary care centers of Kameda Medical Center and St. Luke's International Hospital in Japan between January 2003 and December 2013. Charts were reviewed for HLA typing and disease manifestations at baseline visit and during follow up. For this study, we classified the patients into two groups whether the patients fulfilled ISG or not. As a control group for the general population in Japan, we adopted data from the central bone marrow data center registry of Japanese Red Cross Society (n=263016). The significance of the distribution of alleles was analyzed by one-sample proportions test and the primary p-values were corrected by using the Bonferroni correction (Pc). We used student t test for continuous variables and Fisher's exact probability test for categorical variables.
Results From our overall cohort data, 89 and 96 patients were tested for HLA-A and HLA-B typing respectively. Allele frequency (AF) of HLA-A*26 was significantly higher than the general Japanese population (20.8% vs. 12.0%, Pc=2.9x10-3). The same is true for HLA-B*51 (24.5% vs. 9.0%, Pc=2.1x10-12). Secondly, we also analyzed the ISG positive group. HLA-A*26 and HLA-B*51 were significantly more frequent than in the general Japanese population as well (Table 1). No significant differences were seen among the rest of HLA typing including HLA-B*52. And we also compare ISG positive group and negative group. There is no difference of HLA-A26 and B51 phenotype frequency (40.4% vs 37.8% p=0.83; 43.9% vs 43.6% p=1.00) nor manifestation of BS patients stratified by HLA-A*26 or HLA-B*51.
Conclusions In this study, we found that HLA-A*26 has a higher frequency among not only the clinically diagnosed patients, but also the patients who fulfilled ISG. This result may suggest that HLA-A*26 is an allele for BS susceptibility in Japanese patients without the diagnostic bias.
Disclosure of Interest : T. Kobayashi: None declared, M. Kishimoto Speakers bureau: Santen Pharmaceutical, Mitsubishi Tanabe Pharma, Pfizer, and AbbVie Japan, Y. Ohara: None declared, K. Tokunaga: None declared, N. Takizawa: None declared, H. Nakano: None declared, M. Minoda: None declared, H. Oshikawa: None declared, K. Yoshida: None declared, M. Okada: None declared, K. Matsui: None declared