Volume 7 Supplement 1

25th European Workshop for Rheumatology Research

Open Access

Quantifying the contribution of the second HLA-DRB1 susceptibility allele to the risk of rheumatoid arthritis: a meta-analysis

  • IA van Rossum1,
  • AH Zwinderman2,
  • PP Tak1 and
  • N de Vries1
Arthritis Research & Therapy20057(Suppl 1):P93

DOI: 10.1186/ar1614

Received: 11 January 2005

Published: 17 February 2005


HLA-DRB1 alleles encoding the shared epitope (SE) susceptibility sequence predispose to rheumatoid arthritis (RA). Having two copies results in even higher risk. This study analyzes whether the contribution of the second SE allele to RA risk is identical to that of the first SE allele.


In a meta-analysis of 44 datasets of case–control studies meeting pre-defined criteria, the odds ratios OR1/0 and OR2/1 are calculated, as an estimate of the relative risk for RA for individuals having one compared with zero, or two compared with one SE allele; from these, the ratio OR2/1:OR1/0 is calculated.


Overall the weighted mean ratio OR2/1:OR1/0 (95% confidence interval) was 0.90 (0.80–1.01). In the 39 studies in which all patients met the ACR criteria for RA, the overall ratio OR2/1:OR1/0 is 0.89 (0.78–1.01; see Fig. 1). In a multiple regression analysis on a group level the ratio OR2/1:OR1/0 was not influenced by racial origin, the frequency of compound heterozygosity (0401/0404 or 0401/0408), the percentage of female or rheumatoid factor-positive patients, disease duration or age at disease onset.
Figure 1

Ratio OR2/1:OR1/0 (see Methods) with 95% confidence interval of the 44 study groups and the overall weighted mean.


The contribution of the second SE allele to RA risk does not differ significantly from that of the first SE allele. This result will help to pool immunogenetic data across populations, resulting in a detailed description of the contribution of immunogenetic factors to RA risk.

In addition this finding implies that pathogenetic models for RA that incorporate the role of HLA should explain the fact that the first and the second SE alleles increase RA risk equally. Such effects might, for example, be described for models incorporating presentation of pathogenic antigenic peptides.

Authors’ Affiliations

Department of Clinical Immunology & Rheumatology, AMC/University of Amsterdam
Division of Medical Statistics, AMC/University of Amsterdam


© BioMed Central Ltd 2005