• autoantibodies
• polymyositis
• dermatomyositis

It was with much interest that we read the recent European League Against Rheumatism/American College of Rheumatology classification criteria for idiopathic inflammatory myopathies.1 These criteria include Jo-1 autoantibodies, and the authors discussed that future updates of the criteria should also include the more recently identified myositis-specific autoantibodies.1 2 The interest in autoantibodies for classification is also illustrated by a recent proposal for a new clinicoserological classification of adult autoimmune myositis, which is based on the association of autoantibodies with distinct clinical phenotypes.3 4 For example, antibodies to synthetases (eg, Jo-1, PL-7 and PL-12) define the antisynthetase syndrome, anti-MDA-5 antibodies are associated with myositis with overlap features such as interstitial lung disease, anti-TIF-1γ and anti-NXP-2 define a subgroup of dermatomyositis and anti-SRP and anti-HMGCR are associated with necrotising autoimmune myositis.2

As autoantibodies play a role in the newly proposed classifications,1 2 it is expected that measurement of myositis-specific autoantibodies will be increasingly introduced in clinical practice. Most of the myositis-specific autoantibodies have been identified by immunoprecipitation. Alternative, easy-to-use commercial line/dot immunoassays are available. However, such assays are not standardised and may suffer from low specificity.2 Therefore, these assays need to be further validated.

We evaluated a cohort of 144 patients with inflammatory myopathy (IIM) and 240 controls (blood donors, chronic inflammatory demyelinating polyneuropathy, rheumatoid arthritis, systemic sclerosis, Sjögren’s syndrome and systemic lupus erythematosus; 40 of each) for myositis-specific autoantibodies using assays from Alphadia (myositis 12 IgG dot for Bluediver) (Mons, Belgium), Euroimmun (Euroline Autoimmune Inflammatory Myopathies) (Lübeck, Germany) and Trinity Biotech (ImmcoStripe Myositis Advanced LIA) (Buffalo, New York, USA).

The results are shown in table 1. We observed differences in specificity (reactivity in controls) between the manufacturers and between individual antibodies. For example, 2.9% and 2.4% of controls tested positive for anti-Jo-1 by Euroimmun and Trinity, respectively, compared with 0.4% by Alphadia. Overall, Euroimmun and Trinity showed more reactivity in controls than Alphadia, except for anti-SAE for which Euroimmun showed less reactivity in controls. Differences in reactivities between manufacturers were also observed in myositis patients, with the most pronounced difference for anti-TIF-1γ (2.1% with Alphadia versus 12.4% with Euroimmun and 11% with Trinity). It should be noted that even for an established marker such as anti-Jo-1 antibodies, differences between manufacturers were observed in patients with IIM. The likelihood ratio (LR) (prevalence of antibodies in patients divided by prevalence of antibodies in controls) gives a good estimate of how the test result affects the post-test probability (an LR >10 indicates a clinical significant difference in pretest to post-test probability). The LRs are shown in table 1 and further illustrate differences between individual antibodies and between manufacturers.

Table 2 shows the corresponding phenotype of myositis-specific antibodies in patients with IIM. The association between antisynthetase antibodies and interstitial lung disease, arthritis and Raynaud’s phenomenon was highly significant for all assays. The association of TIF-1γ antibodies and dermatomyositis was high for two of the three assays tested. The association of other antibodies with certain phenotypes (eg, association of TIF-1γ and NXP-2 with malignancy, of Mi-2 with dermatomyositis, NXP-2 with calcinosis and MDA-5 with amyopathic IIM) were weaker and differed between the assays (table 2), indicating that the assays did not perform similarly.

Taken together, as myositis-specific autoantibodies are included in classification criteria, it is important that clinicians and laboratory professionals are aware of the performance characteristics of the assays used to detect such antibodies. Initiatives to harmonise assays across manufacturers are needed.