Response to: ‘Correspondence on ‘EULAR/ACR classification criteria for adult and juvenile idiopathic inflammatory myopathies and their major subgroups’’ by Irfan et al

Abstract

We have with interest read the ‘Correspondence on “EULAR/ ACR classification criteria for adult and juvenile idiopathic inflammatory myopathies and their major subgroups” by Irfan et al’ with subtitle: improving the subtyping of dermatomyositis in the 2017 EULAR/ACR classification criteria for juvenile and adult idiopathic inflammatory myositis: value of nonclassic cutaneous findings, to be published in Annals of the Rheumatic Diseases. Here, the authors present performance of the EULAR/ ACR classification criteria in a cohort from India of 26 consecutive adult patients with idiopathic inflammatory myopathies (IIM). Patients with characteristic skin rashes (those included in the EULAR/ACR criteria as well as in the older Bohan and Peter criteria) were classified as having dermatomyositis (DM). The remainder of the patients were classified as polymyositis (PM) (n=12). All patients had documented muscle weakness. Of the 12 patients classified as PM, seven had nonclassical skin rash such as generalised erythematous rash or malar rash and had been given the diagnosis DM by their physician. Moreover, five of these seven patients had myositisassociated autoantibodies that largely supported the clinical diagnosis of DM. The authors conclude that the sensitivities of Gottron’s papules (60%–80%) and heliotrope rash (30%) are low. The authors suggest that ethnic or geographical differences in the prevalence of different skin rashes may explain this discrepancy in performance of the EULAR/ACR criteria in their cohort of Indian patients compared with the original cohort. They suggest that adding nonclassic cutaneous findings for subclassification would substantially improve the sensitivity of the criteria for subtyping patients with DM across different populations. We agree with the observed limitation of new EULAR/ACR classification criteria for DM. Other reports on the performance of the new criteria have described misclassification of DM with varying results. In one cohort with mainly Caucasian patients 3/151 patients who were clinically subclassified as DM did not have any of the EULAR/ACR criteria skin variables according to the EULAR/ACR criteria; these patients had typical shawl or Vsign. In a cohort from China, 7 of 40 patients with a clinical diagnosis of DM were classified as PM using the EULAR/ACR criteria. These patients had shawl sign, holster sign, erythroderma or other rashes frequently seen in DM patients. In a recent study of 194 adult Japanese patients diagnosed as DM by the physician, 14 were classified as PM by applying the EULAR/ ACR criteria. In a cohort of 99 patients with wellcharacterised amyopathic DM (ADM) from a dermatology centre, 6% did not have any of the ‘classic skin rashes’ and 26% could not be classified as ADM/DM using the EULAR/ACR criteria. In this cohort, Gottron’s sign was the most common skin rash (present in 90%) followed by malar rash (82%), periungual changes (77%), Varea of the neck (76%) and rest of the face rash (76%). Sixtytwo percent had a heliotrope, similar to what was found in the international population used to derive the EULAR/ACR criteria. In the development of the EULAR/ACR classification criteria, 93 variables were tested including 14 skin variables. In addition to the three classical skin rashes, data were collected on shawl and Vsign rashes, periorbital oedema, linear extensor erythema, calcification, periungual erythema, mechanic’s hands, photodistributed violaceous erythema, Raynaud’s phenomenon, cuticular overgrowth and poikiloderma. In the statistical analyses based on 976 IIM patients (DM, n=239; ADM, n=44; hypomyopathic DM, n=12; juvenile (DM) (JDM), n=248), and 624 nonIIM patients, only the three classical skin rashes were found to be significant in the classification of IIM from nonIIM patients, and were included in the final version of the criteria. A difference in the presence of DMassociated skin variables in various ethnic populations is suggested by data from the classification criteria cohort. In the DM subgroup, seven skin manifestations were more prevalent among Caucasian compared with Asian patients, whereas shawlsign was more common among Asian patients. In the subgroup ADM, Caucasian patients had a higher prevalence of seven skin manifestations compared with Asian patients. Still the classical skin rashes were among the most common in both Asian and Caucasian patients (figure 1). Among DM patients, 9.6% were negative for all three EULAR/ ACR criteria skin rashes and among the ADM patients no patient was negative for all three. As discussed in previous letters to Annals of the Rheumatic Diseases, the EULAR/ACR criteria have a limitation as most myositisspecific autoantibodies, except antiJo1 autoantibodies, were not available to be included in the criteria due to historical reasons. Thus, an update of the criteria is needed. To test the performance of the whole panel of myositisspecific autoantibodies, and potentially other variables, we welcome new collaborators with large cohorts of patients with IIM and comparator cases with validated results of autoantibody profiles, as well as thorough clinical data on muscle and extramuscular variables, including skin rashes. An ongoing collaborative effort between dermatology and rheumatology to examine prospectively collected skin data in DM will further examine the skin variables.