MATERIALS AND METHODS

Reference groups

The prevalence of DIP OA and the distribution HLA-DR alleles in the study group were compared with two reference groups representative of the general Dutch population. As a reference for DIP OA, data were used from the Zoetermeer population⁹, a population survey consisting of 3109 men and 3476 women conducted between 1975 and 1978 in two districts of Zoetermeer, a suburban metropolitan area near Leiden in the Netherlands. Subjects in this study were documented for the presence of OA by radiographs of the hands, forefeet, and lateral cervical spine. As a reference group for the distribution of the HLA-DR antigens, data were used from blood donors in the Leiden area who had been typed for HLA by serological methods. This is a group of about 2400 people.

Patients

The study group consists of subjects from the outpatient clinic of the department of rheumatology aged between 40 and 70 with and without radiological DIP OA. This group obtained between 1993 and 2000 is part of a continuing project, the Early Arthritis Clinic (EAC). Consecutive patients with arthritis in at least one joint with a short history of complaints are admitted to this clinic by general practitioners. Each patient subsequently undergoes full clinical, biochemical, and radiographic assessment. Diagnoses in the EAC are made according to international classification and if necessary revised during up to one year of follow up.¹⁰ For the purpose of this study, patients with a definitive EAC diagnosis after one year of follow up are included. Patients with rheumatoid arthritis, systemic lupus erythematosus, and spondyloarthropathies were excluded because of known associations of these diseases with certain HLA-DR alleles. Patients with psoriatic arthritis were also excluded because DIP involvement is common in psoriatic arthritis and might thus interfere with our readings.

Radiographs and radiographic scoring

Plain dorsovolar hand radiographs were taken during the period of the first visit to the EAC. Only radiographs obtained at a maximum of three months before until three months after the first visit were included.

For the purpose of this study each of the radiographs was independently graded for DIP OA by two of three observers (NR, JS, MK) using the Kellgren/Lawrence scale. Furthermore, the observers scoring the radiographs for the presence of DIP OA were unaware of the underlying EAC diagnosis. This overall score distinguishes five degrees of severity of OA according to the presence of the radiological features: osteophytes, joint space narrowing, subchondral sclerosis, cysts, and deformity. A patient was diagnosed with DIP OA if a Kellgren score of two or more was found in at least one DIP joint. The interrater agreement for the presence or absence of DIP OA was 0.7; Cohen’s κ. In cases of disagreement radiographs were re-evaluated until consensus was reached.

HLA typing

DNA was isolated from blood lymphocytes. Generic DRB typing was performed with a polymerase chain reaction and biotin labelled sequence specific oligonucleotide method as described previously.¹¹

Statistics

The means were compared using an independent sample Student t test. The observed (O) DIP OA frequencies in the study group, depicted for sex and age, were related to the expected number (E) in the reference group by calculating the O/E (a relative risk). Odds ratios (OR) with 95% confidence intervals (95% CI) were calculated to determine the comparability of the distribution of HLA alleles in the study group with the reference group and, furthermore, to assess the association between DIP OA and HLA-DR alleles.

RESULTS

Six hundred and one consecutive patients aged 40–70 visited the EAC. Of these patients, 166 met the study criteria and were included in this study. The following patients were excluded: 64 patients because typing of HLA-DR had not been performed and 21 patients because no appropriate hand radiographs were taken. Furthermore, 240 patients were excluded with RA, 32 with psoriatic arthritis, nine with a variety of systemic diseases, and 69 because there was no definitive diagnosis made within one year.

Table 1 summarises the clinical characteristics of these patients. The patients in this study were included in the EAC with a broad variety of diagnoses. Fifty five (33%) patients of our study group had DIP OA. Twenty nine (53%) of the patients with DIP OA were women. The average age of patients with DIP OA was significantly higher than that of patients with no DIP OA respectively 57.7 (SD 6.3) years compared with 49.2 (6.9) years (p<0.001). Table 2 compares the study group with the Zoetermeer group depicted for age and sex. The prevalence of DIP OA was comparable with that of the reference group: O/E 1.0 (95% CI 0.7 to 1.3).

Table 3 summarises the distribution of HLA-DR in the study group in comparison with the Leiden blood donors. No significant difference was found in the distribution of the HLA-DR between the two groups as is shown by the odds ratios of around one for each separate DR antigen.

Table 4 shows the association between DIP OA and HLA-DR in the study group. An association was found between DIP OA and HLA-DR2 and HLA-DR4 with odds ratios of respectively 2.4 (95% CI 1.1 to 5.0) and 0.3 (95% CI 0.1 to 0.7). No association with the other HLA-DR antigens and DIP OA was seen.

DISCUSSION

This study aimed at investigating the association of HLA-DR with DIP OA. We found that HLA-DR2 was more frequent in patients with DIP OA than in patients with no radiological evidence of DIP OA with an OR of 2.4 (95% CI 1.1 to 5.0). We also found a lower prevalence of DIP OA in HLA-DR4 positive patients with an OR of 0.3 (95% CI 0.1 to 0.7). This is to our knowledge the first time that an association with HLA-DR4 has been established. Other investigators have suggested a positive association with HLA-DR2.

We had the unique opportunity to study the relationship of these two variables in a relatively large group of well documented patients representative of the general population. This consecutive patient group was collected prospectively and HLA typing and radiological examinations were performed irrespective of the EAC diagnosis. Because the study group consisted of patients included in an EAC we took into consideration the existing correlation between certain rheumatic diseases and HLA alleles when selecting the patients for the present study. Patients with diagnoses with known HLA associations and diseases associated with radiological damage of the DIP joints, such as psoriatic arthritis, were excluded from the study group, although radiological damage would not have been very likely because patients are included in the EAC at a very early stage. Furthermore, it was shown that the study group is a representative sample of the general population by finding in this group similar frequencies of DIP OA and HLA antigens.

Only two studies, up until now, have investigated the correlation between HLA class II antigens and hand OA, with conflicting results.^6,7 These studies were performed in ethnically separate groups, a group of Ashkenazi Jews and a group of Mexican mestizos. The study of the Ashkenazi Jews was performed with a small group of patients with clinical characteristics of hand OA and controls consisting of “healthy subjects”. They found a positive association with HLA-DR2, although this association was not significant. Possibly, the limited sample size might have contributed to the absence of significant results in this study. The second publication studied the relationship between Heberden’s nodes and HLA in a group of Mexican mestizos. The authors of this study found no association between these two variables. This might have been caused by using different diagnostic criteria of hand OA in patients compared with controls. Patients were diagnosed with hand OA on the basis of the combination of clinical and radiographic characteristics whereas the controls were screened for the absence of Heberden’s nodes without radiological assessment. Because there is a discrepancy between signs and symptoms and the radiological manifestations of OA, as is shown in patients with radiological OA who have symptomatic disease in only 20–40% of cases, the absence of radiological assessment of the control group might have led to an underdiagnosis of OA in the reference group used in this study. Recently, further evidence of the role of HLA-DR2 as a risk factor for DIP OA was found in two European populations, one in Germany¹² and the other in Iceland (Jonsson HH, personal communication).

No earlier study has found an association between HLA-DR4 and hand OA. Therefore, further studies in other populations are needed to confirm these results and to rule out the possibility of a false positive association (type 1 error) because of multiple testing.

One may suggest a direct role of HLA-DR antigens in OA. HLA class II molecules are involved in the communication between T cells and antigen presenting cells. Although OA is not generally considered to be an autoimmune disorder with an inflammatory nature, several studies suggest a role for T cells and HLA class II molecules in OA. One study on the synovial membranes of patients with OA indicates T cell mediated immune responses in this disorder.¹³ Another recent study demonstrated higher levels of human cartilage glycoprotein 39 (HC gp-39) specific T cells in patients with OA than in healthy controls.¹⁴ Furthermore, activated T cells have been suggested to have a role in bone metabolism and osteoclast regulation through the RANKL molecule. This interaction may directly lead to bone formation.¹⁵

Another possibility is that the role of HLA class II alleles implicated in this study is due to linkage disequilibrium with neighbouring genes of HLA-DR class II alleles which may be involved in the pathogenesis of OA. Whether our findings are a matter of haplotype or whether they suggest a direct role for HLA-DR alleles needs to be further investigated.