Major histocompatility complex haplotypic associations in Felty's syndromeand large granular lymphocyte syndrome are secondary to allelic association with HLA-DRB1*0401

Objective. To investigate the role of HLA class I in susceptibility to Felt y's syndrome (FS) and large granular lymphocyte (LGL) syndrome. Methods. Fifty caucasoid FS patients, and 55 patients with LGL syndrome, of whom 26 had arthritis and 29 did not, were studied. Complete HLA class I a nd HLA-DR typing including, where relevant, DRB1*04 subtyping was carried o ut by molecular methods. Comparison was made with 78 unselected healthy cau casoid controls and a further 29 DRB1*0401+ individuals. Results. A significant association was found between HLA-A*02 and FS [odds ratio (OR) 3.9, 95% confidence interval (95% CI) 1.8-8.4, P = 0.0004]. At t he B locus, there was an association between B*44 and LGL with arthritis [O R 3.5 (1.3-9.2), P = 0.01]. For HLA Cw*0501, there was an association with FS [OR 4 (1.7-9.2) P = 0.0008]. For both FS and LGL with arthritis, the ext ended haplotype HLA-A*02;B*44;Cw*0501;DRB1*0401 was significantly associate d [OR 9.5 (2.6-35), P = 0.0001; OR 4.6 (1-22.4), P = 0.05, respectively]. T here was no association between HLA class I or II and LGL without arthritis . All the allelic and haplotypic associations were lost on comparison with HLA-DRB1*0401+ controls. The strongest HLA association was with HLA-DRB1*04 01 for FS [OR 27.9 (10.3-75.5), P = 10(-13)], and LGL with arthritis [OR 35 .4 (9.6-131.3), P = 10(-10)]. Conclusions. The major histocompatibility locus (MHC) associations with FS reported here are due to linkage disequilibrium with HLA-DRB1*0401. LGL syn drome with arthritis shows identical class II associations with FS, althoug h there may be subtle immunogenetic differences between the two in the clas s I region. One of the extended haplotypes reported in a number of studies for FS and rheumatoid arthritis (summarized as HLA-A*02;Cw*0501; B*44;TNFbS ;TNFa6;TNFd4;C4A*3;C4BQ*0;DRB1*0401;DQB1*0301) is likely to be attributable to strong primary association with HLA-DRB1*0401, rather than to epistatic interaction between these loci.