Relatively few studies have been undertaken to evaluate the occurrence of RA-associated HLA-DRB1 alleles in southern African people with RA. The results of the current study, on the basis of high-resolution typing procedures in combination with the du Montcel HLA-DRB1 SE classification system , reveal that the incidence of RA risk-associated alleles in a population of predominantly black South African females with RA is comparable to or higher than those reported in European and Japanese populations [19–21] and somewhat higher than those reported in the relatively few studies undertaken in African Americans, which range from 25% to 40% [35–37]. Our data are in agreement with those of an earlier study where low-resolution PCR typing procedures were used, in which HLA-DRB1*04 conferred the most significant risk of RA in a cohort of black South African RA patients . According to the du Montcel classification system , the highest risk of RA is associated with the S2 and S3P alleles. Recently, Barnetche et al.  reported an association between RA susceptibility and HLA-DRB1 alleles (categorized according to the du Montcel classification system ) in a combined analysis of worldwide samples (1,210 cases of RA), which included 23 San people of southern African origin. Although the number of cases was small, 52.2% of the patients were found to be carriers of the S2 RA susceptibility allele compared to 26% in the control group (OR = 3.05) . In addition, in agreement with previous European and Japanese RA cohort studies [19, 40], we observed highly significant associations of SS and SX with seropositivity for aCCP but not RF [21, 40]. We have previously reported that isolated measurement of aCCP is no more accurate than RF in the serodiagnosis of RA . However, when used in combination with SE genotyping, measurement of aCCP appears to distinguish a subset of patients who may differ with respect to response to therapy and outcome.
These observations appear to support the relationship between HLA-DRB1 SE subtypes, presentation of citrullinated epitopes and development of RA . This contention is underscored by the broad associations of circulating cytokines with the du Montcel classification system risk phenotype . SS is most strongly associated with T-helper cell 1 (GM-CSF, TNF, IFN-γ and IL-2), Th2 (IL-4 and IL-6) and macrophage (IL-1β, IL-6, IL-8, IL-12, TNF and VEGF) cytokines, as well as with IL-1Ra. We and others have previously reported that RA is associated with a generalised increase in circulating proinflammatory and anti-inflammatory cytokines and chemokines, which might be compatible with Th1, macrophage and fibroblast activation and a counterregulatory role of Th2 cells [22, 23, 43].
Our findings are somewhat at variance with those recently reported by Singwe-Ngandeu et al.  in a relatively small group of Cameroonian RA patients (n = 56). Thirty percent of the RA patients in their study were either SS or SX compared to 10% in patients with other inflammatory rheumatic diseases and 14% in healthy controls. However, no significant associations between SE positivity (SS or SX) and aCCP or RF were detected. Several possible reasons might explain the differences between their study and our present study. Notwithstanding the relatively small number of patients, these differences include the number of alleles typed (21 in Singwe-Ngandeu et al. vs 43 in our present study) and, most importantly, the effects of chemotherapy. Whereas the patients in our study were corticosteroid- and DMARD-naïve at the time of presentation, the patients in the Singwe-Ngandeu et al. study were receiving prednisone (91%), methotrexate (77%), sulfasalazine (12%), azathioprine (5%), leflunomide (5%) and D-penicillamine (2%). Although this difference may explain the absence of associations of SE alleles with aCCP and RF, it is unlikely to explain the differences in frequencies of SE alleles detected in the two studies. Notwithstanding the larger number of alleles typed in the current study, it is also possible that our patients represent those at the extreme end of the severe disease spectrum. In the health care setting of the developing world, the presentation of patients to specialised RA clinics, of which there are few, is likely to be delayed, and our study cohort may therefore reflect a selection bias of patients seen in a tertiary setting.