Background/Purpose: HLA-B27 has been linked to sacroiliitis (SII) in psoriatic arthritis (PsA); however, the contribution to SII of other human leukocyte antigen (HLA) Class I loci encoding the major histocompatibility complex (MHC) has received little study.  In particular, whether symmetric and asymmetric SII are differentially genetically determined.  

Methods: Patients from two prospective PsA cohorts were genotyped for HLA-B and C alleles.  Radiographs were evaluated for SII, which was defined as symmetric (symSII) or asymmetric (asymSII) without knowledge of HLA-B and C status.   Associations of alleles with SII were explored in multivariable logistic regression models adjusting for potentially confounding characteristics.  From these models, allele scores were constructed as the sum of an individual’s positively associated alleles minus the sum of their inversely associated alleles  

Results: A total of 490 PsA patients were studied [mean age=56 years, 54% female, median PsA duration=18 years].  SymSII and asymSII was observed in 44 (9%) and 75 (15%), respectively.  In multivariable analysis, symSII was associated with 11 alleles (listed in Fig1).  SymSII was observed in 3% of patients with a symSII allele score of -3 or -2 compared with 12% of those with a score of 0 (OR=4.05; p=0.009) and in 69% of those with a score of +2 (OR=68.4; p<0.001) (Fig1A).  The area under the receiver operator curve (AUC) for detecting symSII with the allele score was 0.758 (95% CI 0.679, 0.836).  Adjusting for body mass index and PsA duration did not substantially alter the associations.                 A different set of five alleles was associated with asymSII (listed in Fig 1).  The prevalence of asymSII was 7% for those with an asymSII allele score of 0 compared with 18% for those with a score of +1 (OR=3.17; p<0.001) and 42% for those with a score of +2 or +3 (OR=10.18; p<0.001) (Fig1B).  The area under the receiver operator curve (AUC) for detecting asymSII with the allele score was 0.684 (95% CI 0.624, 0.742).  Adjusting for family history of psoriasis, current PASI score, and the presence of radiographic osteolysis did not substantially alter the associations.

Conclusion: Symmetric and asymmetric SII are differentially genetically determined, such that the determination of HLA Class I alleles may be useful in predicting risk for SII within a given PsA patient.