Differences are demonstrated in the association of myositis and malignancy between polymyositis and dermatomyositis.

The association between malignancy and dermatomyositis (DM) has been well described. There are, however, other aetiological differences between polymyositis (PM) and DM. Thus, although HLA-DRB1 is associated with myositis, there are substantial differences in the strength of association between HLA-DRB1 between PM and DM. AIMS: 1) To determine if differences exist in the association of malignancy with PM, DM and myositis / overlap disease. 2) To determine whether differences exist in HLA-DRB1 associations between malignancy and non-malignancy associated myositis. METHODS: Data was collected on 267 adult Caucasian patients, with definite myositis, according to the Tanimoto criteria (J Rheumatol 1995, 22:668-674); PM (n=112), DM (n=90), overlap (n=65); who were thoroughly screened for malignancy. DNA was amplified by the polymerase chain reaction (PCR) and HLA-DRB1 genotyping was performed using the Dynal RELITM SSO test. The frequency of malignancy was compared between the myositis groups. HLA-DRB1 alleles were compared between those with and without malignancy, as well as with a randomly selected, ethnically matched control group (n=537). RESULTS: There was 18 malignancies overall in the myositis patients. This was highest in those with DM (11.8%) compared to PM (1.8%) and overlap disease (6.1%) (odds ratio for DM vs PM = 7.3, 95% CI 1.6-33.3). The female preponderance was less obvious in the malignancy group (F: M 1.25 malignancy vs 2.56 remainder, p=0.14). The average age of onset was lower in the PM associated malignancy, compared to the remainder of PM patients (46.5 years vs 49.2 years), but by contrast, was higher in the DM (53.6 years vs. 44.1 years) and myositis/overlap group (54.3 years vs. 45.5 years). None of these differences were significant. Breast carcinoma (n=6) was the most common malignancy, followed by large bowel (n=2) and bladder (n=2). HLA-DRB1, particularly HLA-DRB1*03 was associated with all myositis patients, but no differences were detected in the frequency of the alleles between the malignancy-associated group and the remainder of patients. CONCLUSIONS: The increased association of malignancy with DM confirms previous reports in smaller series of subjects, but the size of the increase is greater than previously thought. There is no evidence from the HLA-DRB1 data that there are immunogenetic differences between those with and without malignancy, and other aetiological explanations should be sought.