The association of HLA-B27 with ankylosing spondylitis accounts for nearly 40% of the total disease risk. However, fewer than 5% of B27-positive individuals in the general population become affected. Genomewide scans suggest that other major histocompatibility complex genes further heighten this risk, although linkage disequilibrium with HLA-B27 has confounded their precise identification. Over 31 variants of HLA-B27 have been identified to date, which have evolved from the original B27 allele (B*2705) along three geographic lines. HLA-B*2705 and B*2702 are the primary subtypes in Caucasians with spondylitis, and B*2704 and B*2707 are the primary subtypes in Asians. HLA-B*2706 and B*2709 are not disease associated. There are four theories of how HLA-27 causes spondyloarthritis: (1) HLA-B27 presents a bacterially derived 'arthritogenic peptide' (not yet identified); (2) misfolding or homodimerization of HLA-B27 heavy chains results in a pro-inflammatory response; (3) HLA-B27-positive individuals have deficient intracellular killing of arthritogenic organisms; and (4) HLA-B27 itself, due to sequence homology with bacterial proteins, becomes autoantigenic.