Background/Purpose: A rigorously ascertained psoriatic arthritis (PsA) cohort demonstrated considerable genetic heterogeneity and provided preliminary evidence that MHC genes determine quantitative traits within the PsA phenotype with different patterns of MHC effect.

Methods: We now extend these findings by performing detailed clinical phenotyping of PsA cases to better characterize the clinical features associated with particular HLA class I alleles and their haplotypes. 

Results: A total of 150 PsA patients [mean age 52±12 years; 46% male; mean PsA duration=25±12 years; 45% with axial involvement; 25% with sacroiliitis; 41% with radiographic erosions; median PASI=1.2] were studied.  In univariate analysis, the inheritance of B*27:05 was apparently associated with a more severe joint disease phenotype including: joint erosions, the requirement for TNF therapy and axial disease manifestations such as spine involvement and sacroiliitis that were temporally preceded by or coincident with skin disease. In striking contrast, the presence of HLA-C*06, whether on a B*57-C*06 or B*37-C*06 haplotype, was associated with a considerable delay in the development of arthritis, and conferred a reciprocal phenotype of significant negative associations with arthritis severity, including the presence of erosions, the requirement for TNF inhibitor therapy, and axial disease. B*08 and B*08-C*07 (EH8.1) were correlated with joint deformities, erosions, TNF inhibitor requirement, osteolysis, and dactylitis, developing after the appearance of psoriasis, suggesting this haplotype denotes a more severe but delayed arthritic phenotype.

Surprisingly, in contrast to the uniform association of disease susceptibility with B*27, the association with phenotypic features was not uniform across all B*27 alleles, and was mainly accounted for by the B*27-C*01 (EH27.1) haplotype and not B*27-C*02 (EH27.2), suggesting the influence of additional genetic effects on EH27.1. The predictive value of these haplotypes was confirmed by logistic regression, which after adjustment for confounders showed, for example, the probability of developing sacroiliitis was almost completely determined by the inheritance of EH27.1, EH8.1 or C*05, figure A. Similarly, the probability of developing peripheral joint erosions was strongly associated with the presence of EH27.1, EH8.1 or C*03, figure B.

Conclusion: Certain HLA alleles, and, most strikingly particular haplotypes, contribute importantly to the magnitude of traits comprising the diverse phenotypes of PsA, but this contribution does not completely parallel the role of these alleles or haplotypes in determining susceptibility.