To analyze the genetic contribution of HLA in development of psoriatic arthritis (PsA) and to study whether MICA is primarily associated with PsA or whether its association is secondary to linkage disequilibrium with centromeric genes, such as MICB, TNFA, or HLA-DRB1.DNA samples from 81 Spanish patients with PsA and 110 healthy controls were examined by polymerase chain reaction (PCR) sequence-specific primers to type HLA-Cw and HLA-DRB1, PCR sequence-specific oligonucleotides to determine HLA-B, and PCR restriction fragment length polymorphism for tumor necrosis factor-alpha promoter polymorphisms at positions -238 and -308. Analysis of microsatellite polymorphisms in the transmembrane region of MICA and in intron 1 of MICB was also carried out.HLA-Cw*0602 was significantly increased in PsA [60% vs 17%; p(c) < 0.00002, OR 7.33, etiological fraction (EF) 0.52]. MICA-A9 (60% vs 30%; p(c) = 0.0002, OR 3.57, EF 0.43) and the microsatellite MICB-CA-22 allele (23% vs 7%; p(c) = 0.028, OR 3.9, EF 0.17) were also significantly increased in PsA. MICA-A9 was in linkage disequilibrium with MICB-CA-22 (delta = 0.6). The association of MICA-A9 was independent of MICB-CA-22 and Cw*0602, since it was also associated in MICB-CA-22 negative (p(c) = 0.0015, OR 2.96, EF 0.34) and in Cw*0602 negative patients (p(c) = 0.034, OR 2.83, EF 0.34). TNFA and DRB I alleles were not significantly associated with PsA.Cw*0602 and MICA-A9 appear to be the strongest genetic susceptibility factors for PsA. However, MICA-A9 was associated independently of Cw6. HLA-B alleles and MICB-CA22 are associated secondarily to linkage with MICA. TNFA and HLA-DRB1 were not associated with PsA susceptibility, and our data suggest that their reported association may only reflect the linkage disequilibrium with MICA-A9 among the different populations studied.