Performance of the 2017 European Alliance of Associations for Rheumatology/American College of Rheumatology Classification Criteria for Idiopathic Inflammatory Myopathies in Patients With Myositis-Specific Autoantibodies

Abstract

Objective: We undertook this study to 1) determine the sensitivity of the European Alliance of Associations for Rheumatology (EULAR)/American College of Rheumatology (ACR) classification criteria for idiopathic inflammatory myopathies (IIMs) to properly classify myositis-specific autoantibody (MSA)-positive myositis patients, 2) describe the phenotype and muscle involvement over time in different MSA-positive patients, and 3) compare MSA subgroups to EULAR/ACR criteria-defined myositis subgroups for their capacity to predict clinical phenotypes in patients with IIMs.

Methods: The study included 524 MSA-positive myositis patients from the Johns Hopkins Myositis Center. Each patient was classified using the EULAR/ACR classification criteria. Patient phenotypes were summarized using factor analysis of mixed data (FAMD). We compared the ability of MSAs to that of the EULAR/ACR classification subgroups to predict the phenotype of patients by applying the Akaike information criterion (AIC) and the Bayesian information criteria (BIC) to the linear regression models.

Results: Overall, 91% of MSA-positive patients met the EULAR/ACR criteria to be classified as having myositis. However, 20% of patients with anti-hydroxymethylglutaryl-coenzyme A reductase (anti-HMGCR) and 50% of patients with anti-PL-7 were incorrectly classified as not having myositis. Furthermore, ~10% of patients with anti-signal recognition particle (anti-SRP) and patients with anti-HMGCR were misclassified as having inclusion body myositis. FAMD demonstrated that patients within each MSA-defined subgroup had similar phenotypes. Application of both the AIC and BIC to the linear regression models revealed that MSAs were better predictors of myositis phenotypes than the subgroups defined by the EULAR/ACR criteria.

Conclusion: Although the EULAR/ACR criteria successfully classified 91% of MSA-positive myositis patients, certain MSA-defined subgroups, including those with autoantibodies against HMGCR, SRP, and PL-7, are frequently misclassified. In myositis patients with MSAs, autoantibodies outperform the EULAR/ACR-defined myositis subgroups in predicting the clinical phenotypes of patients. These findings underscore the need to include MSAs in a revised myositis classification scheme.

Figure 1.

Evolution of the strength (blue line) and creatine kinase (orange line) using locally weighted polynomial regression (LOWESS) of patients with different myositis-specific autoantibodies over time.

Figure 2.

Factor analysis of mixed data summarizing the clinical phenotype of the autoantibody-positive adult myositis patients. The factor analysis of mixed data included the gender, race, age at onset, median and maximum creatine kinase, presence of anti-Ro52 autoantibodies, and presence or absence during the course of the disease of: muscle weakness, interstitial lung disease, arthritis, heliotrope or Gottron’s rash, calcinosis, Raynaud’s phenomenon, mechanic’s hands, dysphagia, and fevers. The only two factors explaining more than 10% of the variance (factor 1 nd factor 2) were retained for further analysis.