Although a number of reports have now described an association between polymorphism of the LMP2 gene and disease phenotype in HLA-B27 positive individuals with ankylosing spondylitis (AS), some describe associations with acute anterior uveitis, others with juvenile onset disease, and one report provides no association. A recent study describes yet a further association with disease severity in patients with juvenile rheumatoid arthritis. We therefore hypothesized that the discrepant findings in adult disease may be a reflection of an underlying association with disease severity. Our study population consisted of 100 HLA-B27 positive Caucasians with AS of ten or more years duration. Clinical assessment of disease severity was based on a metrology index scoring five measurements, the modified health assessment questionnaire for the spondyloarthropathies, and a disease activity index consisting of a visual analog scale to score the amount of pain, stiffness and fatigue. LMP2 genotypes were assigned following polymerase chain reaction amplification from genomic DNA and restriction enzyme digestion with CfoI. Despite confirmation of a significantly higher prevalence of the LMP2 BB genotype in AAU positive (66.0%) versus AAU negative (45.2%) patients (P < 0.05), we observed no association between LMP2 genotypes and any of the indices of disease severity. Furthermore, although a significant association was noted between the presence of peripheral synovitis and the functional index score (P < 0.05), a history of AAU was not associated with more severe disease. Our data is thus internally consistent in demonstrating no association between LMP2 genotypes and either disease severity or peripheral arthritis, and supports the notion that polymorphism of LMP2 primarily influences the development of AAU and not some other phenotype of AS.