In the present study we have analysed the effect of HLA-DRB1 and -DQB1 alleles on disease progression and genetic predisposition among 201 RA patients. We clearly confirm the association of RA with HLA class II alleles sharing the (Q)R/KRAA amino acid (AA) cassette in the third hypervariable region (HVR3) of the DRβ-chain. The HVR3 (Q)R/KRAA motif was significantly overrepresented among RA patients (79% vs. 40%, P < 0.001), with one third of the patients homozygous (28% vs. 6.7%, P < 10–9) and the number of rheumatoid factor positive (RF +) patients was significantly increased among HVR3 (Q)R/KRAA homozygous in comparison to HVR3 (Q)R/KRAA negative individuals. Erosive disease defined by the Larsen Score and personal disability determined using the Health Assessment Questionnaire (HAQ) was significantly increased among patients positive for the HVR3 motif with the worst outcome among HVR3 (Q)R/KRAA homozygous patients. In contrast, there was no association of the shared HVR3 AA cassette and disease severity in the majority of patients presenting systemic (extraarticular) disease. Homozygosity for the shared HVR3 motif was only marginally increased among patients presenting ‘severe’ extraarticular disease in comparison to patients with articular disease (33% vs. 43%, P = ns). Similarly, patients with nodular disease were not more often homozygous for the HVR3 (Q)R/KRAA motif. Furthermore, we observed no HLA–DR independent association of DQB1 alleles among HVR3 (Q)R/KRAA positive patients and controls. Our analysis supports the predominant role of HLA-DR for genetic susceptibility to RA. In the clinical setting, however, HLA-DR typing may be limited to assess the individual risk of patients for disease progression.