Background Psoriasis is known to be associated with HLA component (HLA-B*13, B*17, Cw*06 and DRB1*07). However few data exist on HLA association with PsA especially using highly sensitive HLA-Cw, DR and DQ typing methods.

Objectives The aim of this work was to study the distribution of HLA alleles in PsA patients in comparison with unrelated controls from the same geographic area.

Methods 69 patients (35 males, 34 females) were included according to modified Avila criteria. A first control group was composed of 328 healthy individuals (random controls: RC). Because alleles at the B and Cw loci are in strong linkage disequilibrium, 60 control individuals matched with patients for HLA-B alleles were selected (matched controls: MC). All the individuals were typed for HLA-A and B by microlymphocytoxicity and for HLA-Cw, DRB1 and DQB1 by PCR-SSP after geno-mic DNA extraction of peripheral blood nucleated cells. The significance of differences in phenotype frequencies was determined by Fischer’s exact test.

Results No significant difference in the HLA-B allele distribution exists between patients and MC. Disease expression of the PsA patients was as follows: skin psoriasis in 100%, axial PsA in 37.7%, asymmetric oligoarthritis in 15.9%, symmetric polyarthritis in 34.8% and asymmetric polyarthritis in 46.4% of the patients. HLA-A*01 (p = 0.0011), B*27 (p = 0.03), B*38 (p = 0.003) and B*57 (p = 0.03) were significantly increased, whereas HLA-A*11 (p = 0.017) and B*18 (p = 0.0019) were significantly decreased in PsA patients compared to RC. 34.4% of the PsA patients carried Cw*0602 compared to 11% in RC (p = 0.003). No difference of DRB1 and DQB1 allele frequencies could be evidenced. When PsA patients and MC were compared no difference of HLA class I and class II allele frequencies was shown except for a significant increase of Cw*0102 in patients (p = 0.0025). Early onset of PsA (as well as psoriasis) was significantly associated with Cw*0602 (p = 0.01) and DRB1*07 (p = 0.03). No association between spondilytis (axial PsA) with and B*27 was observed. Finally, although the sample size is small (10 patients), severe forms of disease were significantly associated with HLA-B*38 (p = 0.0035).

Conclusion Taken together, these data show that alleles at class I loci are associated with the disease, although it is difficult to assess their specificity to PsA genetic susceptibility. No particular role of the HLA class II region could be demonstrated.