Objective. To examine the association between the IL1 gene cluster and susceptibility to ankylosing spondylitis (AS) in 3 independent case-control cohorts.Methods. We analyzed 394 patients and 446 controls from Alberta, Newfoundland, and Toronto, Canada. Samples were genotyped using a panel of 38 single-nucleotide polymorphism (SNP) markers within the IL1 gene cluster. Data from 20 informative and nonredundant SNP markers were analyzed using several association test strategies. First, we used the program WHAP to identify single-marker associations. Second, we used WHAP to analyze "sliding windows" of 3 contiguous markers along the entire extent of the IL1 gene cluster in order to identify haplotypic associations. Third, we used the linkage disequilibrium mapping program DMLE to estimate the posterior probability distribution of a disease locus.Results. A total of 14 SNP markers showed significant single-locus disease associations, the most significant being rs3783526 (IL1A) (P ؍ 0.0009 in the Alberta cohort, P ؍ 0.04 in the Newfoundland cohort) and rs1143627 (IL1B) (P ؍ 0.0005 in the Alberta cohort, P ؍ 0.02 in the Newfoundland cohort). Analysis of 3-marker sliding windows revealed significant and consistent associations with all of the haplotypes in the IL1A and IL1B loci in the Alberta cohort and with IL1B in the Newfoundland cohort, especially haplotypes rs1143634/rs1143630/rs3917356 and rs1143630/ rs3917356/rs3917354 (P ؍ 0.006-0.0001). With DMLE, a strong peak in the probability distribution was estimated near IL1A in both the Alberta and the Newfoundland populations.Conclusion. These results indicate that the IL1 locus, or a locus close to IL1, is associated with susceptibility to AS.