Characterization of non-B27 susceptibility genes will be required to know the pathogenesis of AS. The aim of this study was to examine whether ankylosing spondylitis (AS) susceptibility is controlled by B27 alone rather than B27 haplotypes and, whether other closely related class I loci, such as MICA and TNFA genes might play a role in AS. Three hundred eleven B27-positive samples from Caucasoid, Asian, and African populations were selected and genotypes were carried out by PCR/SSOP (HLA-B27 and HLA-C), PCR/SSP (MICA-TM polymorphism in the transmembrane region), PCR/SSCP (MICA alleles), and PCR-RFLP (TNF-alpha). Of these, 161 were AS patients, chosen in order to investigate the contribution of TNFA and MICA loci to AS in HLA-B27 positive individuals. Some findings can be concluded from the study: (a) No significant differences of TNF-alpha promoter alleles at position -308 and -238 (A/G) were found between AS patients and B27 matched alleles from healthy controls; (b) strong linkage disequilibrium was found between the B27 and the MICA alleles. The MICA-A4 was found to be in association with B*2705,02,03 and 08; MICA-A5 with B*2704 and B*2707 and MICA-A.5.1 with B*2706; (c) no significant differences of MICA alleles were found between AS and controls carrying the B27-associated alleles, and therefore no evidence is provided for an additional role of MICA gene in AS susceptibility; (d) there are a striking correlations between the structure of B27 extended haplotypes (from MICA region to HLA-C) and the ethnic distribution of these subtypes. The results of differential linkage disequilibrium with HLA-B27 subtypes suggest that B27 itself remains the primary gene for AS susceptibility, and TNFA and MICA are not involved in the pathogenesis of the disease.