Idiopathic Inflammatory Myopathy Classification Essay
The European League Against Rheumatism/American College of Rheumatology (EULAR/ACR) has developed new classification criteria for idiopathic inflammatory myopathies (IIMs) that generally perform better than existing criteria.
The most common subgroups of IIMs in adults are dermatomyositis (DM), polymyositis (PM), and inclusion body myositis (IBM). In children, juvenile dermatomyositis (JDM) is most common.
Led by Ingrid E. Lundberg, MD, PhD of the Karolinska University Hospital and Karolinska Institutet in Stockholm, Sweden, the International Myositis Assessment and Clinical Studies group (IMACS) has developed consensus on outcome measures and definitions of improvement for myositis.
Classification criteria development
The objective of this international project was to develop classification criteria for adult and juvenile IIM and their major subgroups. The researchers used a probability-score model because of its superior discriminating performance.
Based on statistical models, 16 variables from six categories best distinguished IIM cases from comparators. Each variable was assigned a weight (score) based on its influence to discriminate IIM from non-IIM, and a total score was computed by adding score points corresponding to each criterion being present. As pediatric experts generally use fewer muscle biopsies for classification of JDM in clinical practice than adult rheumatologists, a second model not including biopsy variables was also developed.
With or without muscle biopsy, the classification criteria provide a score and a corresponding probability of having IIM. To facilitate use of the new criteria, a web-based calculator has also been created.
A patient classified as having IIM by the new EULAR/ACR classification criteria can be further subclassified using a classification tree developed by IMACS. Adult IIM is distinguished from juvenile IIM by age at onset of first symptom. Based on this data set, juvenile patients with skin rash can be classified into JDM. For adults, clinical findings and muscle biopsy features can separate PM, IBM, DM, or amyopathic DM (ADM). Because of small sample sizes, these criteria are unable to separate three additional subgroups: juvenile PM, immune-mediated necrotizing myopathy (IMNM), and hypomyopathic DM.
Performance of EULAR/ACR criteria
“The EULAR/ACR classification criteria are the first myositis criteria to be validated and tested for sensitivity in other cohorts and revealed no misclassification,” the researchers write. Compared with most previous criteria, the new criteria are superior in sensitivity, specificity, and classification accuracy. The criteria including muscle biopsy displayed high sensitivity (93%) and specificity (88%). Performance was slightly lower without biopsy variables (sensitivity of 87% and specificity of 82%).
While the previous Targoff criteria showed the highest sensitivity (93%) and specificity (89%), it was not able to capture all subgroups of IIM, as ADM patients were not included. Additionally, the variables of the Targoff criteria are not defined clearly enough and it requires testing more variables, including electromyography, which is not always easily accessible and may be painful for patients. By comparison, the EULAR/ACR criteria use a limited number of accessible, defined clinical and laboratory variables and can be used for both adults and children.
Other criteria had either high sensitivity and low specificity (Bohan and Peter, Tanimoto critiera) or low sensitivity and high specificity (Dalakas and Hohlfeld, and European Neuromuscular Center criteria).
Despite utilizing data from centers in Europe, America, and Asia, a potential limitation of the new EULAR/ACR criteria is that a majority of the patients were Caucasian. Thus, the new criteria must still be validated in Asian and African populations. Revision of the criteria will also be important in the future when additional validated myositis autoantibody tests, imaging, and other tests become available. “Importantly, these criteria are proposed as classification criteria in research and clinical trials, not as diagnostic criteria,” the researchers conclude.
Lundberg IE, Tjärnlund A, Bottai M, et al; International Myositis Classification Criteria Project Consortium, the Euromyositis Register, and the Juvenile Dermatomyositis Cohort Biomarker Study and Repository (UK and Ireland). 2017 European League Against Rheumatism/American College of Rheumatology Classification Criteria for Adult and Juvenile Idiopathic Inflammatory Myopathies and Their Major Subgroups. Arthritis Rheumatol. 2017;69:2271-2282.
Interstitial lung disease
ILD is a frequent manifestation of dermatomyositis/polymyositis with the prevalence of clinically relevant ILD ranging from 17% to 36% [9, 47–51]. In a prospective series of patients with systematic assessment of lung involvement, the prevalence was estimated to be 65% . ILD occurs after the diagnosis of polymyositis or dermatomyositis in up to 40% of patients, and precedes the diagnosis of overt CTD in 20–30% of cases [47, 52]. In large cohorts of JDM/JPM, ILD occurs in 8–13% of patients and thus seems to occur less frequently than in adult dermatomyositis [53, 54]. Risk factors for ILD in a patient with dermatomyositis/polymyositis include genetic predisposition [55, 56], some clinical signs such as arthralgia , and autoimmune features (table 2) [22, 56, 57]. After muscular involvement, dermatomyositis/polymyositis-associated ILD is the second major contributor to morbidity . Up to half of patients with IIM-associated ILD may eventually die from respiratory failure . The mortality rate is especially high in acute forms of ILD despite immediate and aggressive treatment [48, 58].
The signs and symptoms related to ILD in IIM do not differ from idiopathic ILD, and include dyspnoea on exertion, cough, decreased exercise tolerance, digital clubbing and asthenia. Physicians who diagnose ILD have to pay particular attention to the extrapulmonary features of IIMs to distinguish secondary ILD from primary ILD. The assessment of such patients encompasses careful general (weight loss and fever), cutaneous (rash, sclerodactyly, Raynaud's phenomenon, calcinosis, cutaneous ulcers and mechanic's hands), muscular (loss of muscle mass, muscular testing and myalgia), and rheumatologic (arthralgia, arthritis and Jaccoud's deformity) examination. Subcutaneous calcinosis is present in ∼25% of cases of JDM, in contrast to its rarity in adult onset dermatomyositis.
Beyond clinical assessment, the diagnosis of hypomyopathic forms of myositis requires creatine kinase and aldolase testing. In cases of high suspicion of IIM-related ILD (e.g. rash evocative of dermatomyositis or MSA positivity), supplementary investigations adapted for each patient comprise electromyography, muscular resonance magnetic imaging and muscle biopsy.
Testing for the presence of antinuclear autoantibodies completes the clinical evaluation; further testing for MSA or MAA depends on the pattern of nuclear fluorescence in most instances. The clinical spectrum associated with anti-ARS autoantibodies is highly heterogeneous and includes arthritis (62%), mechanic's hand (28%), fever (43%), Raynaud's phenomenon (47%) (fig. 1), myositis (57%) and ILD (70%) in its classical description [8, 15, 23]. Because of the high frequencies of nonspecific signs associated with anti-ARS autoantibodies (i.e. Raynaud's phenomenon, fever, hyperkeratosis and cracking of the skin on the fingers), we favour a revised definition of anti-ARS syndrome that was proposed by Connorset al.  based on major (arthritis, ILD and myositis) and minor criteria (table 3). In a meta-analysis of clinical manifestations, the frequency of ILD and fever was increased by 20–30% in patients with anti-ARS autoantibodies other than anti-Jo1 autoantibodies (especially anti-PL7 and anti-PL12 antibodies) as compared with those with anti-Jo1 autoantibodies . By contrast, the frequency of mechanic's hands and of articular and muscular involvement was increased by 50% in patients with anti-Jo1 autoantibodies as compared with other anti-ARS antibodies . Patients with anti-PL7 or anti-PL12 antibodies frequently present with isolated ILD, and are therefore mostly seen by respiratory specialists, whereas other signs are missing . Patients with double positivity for anti-ARS and anti-U1RNP antibodies frequently experience manifestations of dermatomyositis (55%) and Raynaud's phenomenon (82%) . In patients with anti-ARS antibodies, the presence of mechanic's hands may be associated with a higher risk of developing ILD .
Raynaud's phenomenon in a patient positive for anti-PL7 antibodies.
The clinical features described in patients with anti-PM/Scl antibodies largely overlap with those in patients with anti-ARS autoantibodies [19, 32]. Indeed, anti-PM/Scl positive patients commonly experience arthralgia (74%), Raynaud's phenomenon (63%), ILD (38%), fever (33%) and mechanic's hands (11%) . Cutaneous manifestations of dermatomyositis and sclerodactyly are reported in 49% and 76% of cases, respectively . The rash is usually mild and transient, with isolated heliotrope rash or Gottron's papules .
In cases with the presence of anti-MDA5 autoantibodies, digital ulcers are located mainly over the Gottron's papules and digital pulps or periungueal areas, but can occur in other sites of the dermatomyositis rash, i.e. the dorsal side of the elbows, knees, auricles or buttocks (fig. 2) [57, 58]. One European study reported an increased frequency of panniculitis . Arthritis (80%), Raynaud's phenomenon (45%) and mechanic's hands (80%) (fig. 3) are frequently found . Anti-MDA-5 autoantibodies have now been added to the differential diagnoses to consider, especially when the clinical features would suggest anti-ARS or anti-PM/Scl autoantibodies in dermatomyositis patients . Importantly, screening for antinuclear autoantibodies may be negative in patients with anti-MDA5 autoantibodies, which therefore have to be measured specifically upon clinical suspicion .
Clinical features associated with anti-MDA5 autoantibodies include severe cutaneous ulcers of a, b) the fingers and c) the dorsal side of the elbow.
Mechanic's hands in a patient with anti-MAD5 positive dermatomyositis. Image courtesy of Sébastien Debarbieux (Dept of Dermatology, Centre Hospitalier Lyon Sud, Lyon, France).
Proposed criteria for myositis associated with anti-tRNA synthetase antibody
Rapidly progressive ILD
Rapidly progressive ILD is defined by an acute interstitial pneumonia progressing over several weeks or a few months. Among all CTDs, this aggressive presentation of ILD is suggestive of IIM and especially of dermatomyositis, CADM or hypomyopathic dermatomyositis [60, 61]. Thus, physicians have to search thoroughly for minimal skin modifications or mild muscular involvement in the setting of rapidly progressive ILD.
The presence of anti-ARS autoantibodies seems relatively protective against rapidly progressive ILD, although this presentation has been reported [62, 63]. The detection of anti-MDA5 autoantibodies identifies dermatomyositis patients with a high risk of developing rapidly progressive ILD, with a pooled sensitivity of 77% and a specificity of 86% in a meta-analysis of 631 dermatomyositis or polymyositis patients [64, 65]. Fulminant presentations with ILD, CADM and anti-MDA-5 autoantibodies may be especially frequent in Japanese patients compared with other east Asian or non-Asian patients [35, 64], although a publication bias cannot be excluded. In a Japanese retrospective series of 54 patients with JDM , 10 had rapidly progressive ILD and 19 had chronic ILD, illustrating the particularly frequent pulmonary involvement in the Asian population.
Aside from recent advances in patients with anti-MDA5 autoantibodies, the pathogenesis of dermatomyositis- or polymyositis-related ILD is poorly known, contrary to that of SSc . Immunohistochemistry on muscle biopsy provides some clues as to immune mediated mechanisms. CD8+ T-cells are usually found in polymyositis with a diffuse cytotoxic effect leading to muscle cell necrosis , whereas B-cells and CD4+ T-cells predominate in perivascular areas in dermatomyositis and may be responsible for muscular vasculitis [67, 69]. In lung biopsy specimens of patients suffering from dermatomyositis- or polymyositis-associated ILD, numerous activated CD8+ lymphocytes are found in both affected and relatively preserved tissue. The proportion of these cells in the bronchoalveolar lavage (BAL) fluid might predict the outcome after treatment with corticosteroids, as CD8+ T-cells are significantly increased in corticosteroid-resistant cases of dermatomyositis- or polymyositis-associated ILD . This observation needs confirmation, however, before it can be implemented in the clinic.
In addition, correlations are emerging between some genotypes and the clinical phenotype or prognosis. For example, a polymorphism in the promoter of the tumour necrosis factor-α gene is increased in dermatomyositis/polymyositis patients with ILD (OR 2.1, 95% CI 1.0–4.3) . The HLA-DRB1*03 haplotype, a determinant factor in the development of ILD among patients with CTDs [72, 73], predicts a higher risk of ILD in dermatomyositis/polymyositis , irrespective of the myositis subtype or the presence of autoantibodies . Anti-PM/Scl and anti-ARS antibodies also share the same prevalence of HLA-DRB1*03, DQA1*05 and DQB1*02 haplotypes [75, 76].
Recently, gain of function mutations in the IFIH1 gene encoding the protein MDA5 were reported in association with various inflammatory conditions including a skin rash reminiscent of JDM skin lesions . Anti-MDA5 is a cytosolic double strand RNA receptor, and gain of function mutations of IFIH1 lead to enhanced type I interferon production. Similarly, gain of function mutations in the TMEN173 gene encoding the DNA sensor STING (stimulator of interferon genes), which is responsible for enhanced type I interferon production, were found in patients with vascular involvement and severe ILD . Taken together, these findings suggest that type I interferon may be a key cytokine linking anti-MDA5 autoantibodies to lung involvement and ILD.
High-resolution computed tomography (HRCT) is essential for evaluation of ILD. Pulmonary HRCT scanning is a noninvasive and sensitive test, demonstrating evidence of pulmonary changes in all patients suffering from ILD. These changes are predominantly localised in the lower lobes, as reported by previous studies [47, 79]. Linear opacities, ground-glass attenuation (hyperattenuated areas in which the bronchi and vessels remain visible), reticulation and peribronchovascular thickening are the most common HRCT abnormalities at the initial imaging of ILD. Consolidation corresponding to areas with increased attenuation and decreased visibility of bronchial walls and vessels are commonly found, and correspond pathologically to areas of organising pneumonia. A combination of patchy areas of consolidation and reticulation is frequent, and suggestive of the diagnosis of myositis in subjects with ILD, especially of subacute or acute onset. Honeycombing (areas of small cystic spaces with thickened walls), traction bronchiectasis (bronchial dilation due to traction by fibrous tissue) and subpleural bands are less frequently observed .
Pathologic and imaging correlations in IIM are generally similar to those in the idiopathic setting. Overall, a variety of HRCT patterns may be observed in IIM-associated ILD, with especially those of organising pneumonia, nonspecific interstitial pneumonia (NSIP) (fig. 4) or mixed NSIP–organising pneumonia pattern, whereas the usual interstitial pneumonia (UIP) pattern (fig. 5) is less frequent . However, studies that correlate imaging to pathology in the context of IIM are limited and suffer from selection bias, as clinicians tend to find lung biopsy indicated in the less typical cases. Therefore, caution must be exerted when interpreting the imaging, and HRCT patterns cannot be directly transposed into histological considerations.
High-resolution computed tomography from a patient with anti-Jo1 positive polymyositis showing basilar predominate reticulation and ground-glass opacity without honeycombing, in a pattern suggestive of nonspecific interstitial pneumonia.
High-resolution computed tomography from a patient with polymyositis showing honeycombing in a pattern consistent with usual interstitial pneumonia.
HRCT findings do not differ significantly between the various presentations of polymyositis- or dermatomyositis-ILD ; however, acute and subacute onset of ILD seem to be associated with more frequent and more extensive consolidation, which is the hallmark of this ILD subtype. Other HRCT features include diffuse patchy ground-glass opacity, basal irregular lines and some reticulation . HRCT patterns at diagnosis can help predict the prognosis of patients with polymyositis- or dermatomyositis-ILD , and consolidation (mostly associated with a histological pattern of organising pneumonia) is generally reversible with treatment.
Polymyositis-ILD may be associated with a better prognosis than other forms. The HRCT patterns of dermatomyositis-ILD can be categorised in three groups to help clinical decisions . Dominant ground-glass attenuation and/or reticular opacity without honeycombing, subpleural bands, and traction bronchectasis are often associated with severe ILD and a possibly fatal outcome. Prominent consolidation is typical of acute onset ILD that responds well to corticosteroids and often immunosuppressive therapy . The outcome of the HRCT pattern with ground-glass attenuation, predominant reticulation, bronchiectasis, and accessory findings is less predictable.
In patients with anti-ARS autoantibodies, the imaging patterns suggestive of NSIP and organising pneumonia predominate, and consolidation is present in half of the cases that resolve most often . ILD related to anti-PL12 autoantibodies is more frequently associated with reticulation and traction bronchiectasis than anti-Jo1 autoantibodies. As expected, patients with anti-Jo1 autoantibodies who experience severe ILD more frequently have a UIP pattern on imaging than those with less progressive disease (54% versus 18%, OR 5.02, 95% CI 1.05–243) , suggesting that the imaging pattern of ILD may have prognostic significance in this setting, similar to what is observed in idiopathic interstitial pneumonia.
Pulmonary function tests
ILD is typically associated with restrictive lung function impairment (total lung capacity (TLC) and diffusing capacity of the lung for carbon monoxide (DLCO) <80% predicted). DLCO is commonly reduced in patients with ILD, often at an earlier stage of the disease than TLC and forced vital capacity (FVC). Pulmonary function tests add important information to HRCT to evaluate severity, as they only partially correlate with HRCT scores and pattern. In addition, pulmonary function is more sensitive to changes upon therapy than HRCT abnormalities.
Interpretation of pulmonary function tests may be difficult in IIMs, because of the potential coexistence of ILD, respiratory muscle weakness and/or PAH, with all three impairing DLCO. Moreover, a restrictive pattern may be driven by ILD but also by respiratory muscle weakness, the evaluation of which is often difficult to interpret. Conversely, pseudo-stabilisation of a restrictive pattern may be spuriously attributed to improved diaphragmatic strength . In one study, pulmonary function parameters improved early in the course of disease in some patients treated with a combination of high-dose corticosteroids and immunosuppressive agents, and even normalised in some patients despite persisting radiographic changes . The clinical significance of persisting abnormalities on HRCT despite normal lung function is currently unclear. Clinicians must be aware of the possible discrepancy between imaging and lung function in patients with dermatomyositis- or polymyositis-ILD to interpret response to therapy and outcome.
BAL cellular profile
When combined with the results of a careful history, physical examination and thoracic imaging, analysis of the BAL differential cell count can contribute to the diagnosis of ILD. BAL differential cell count per se does not predict the pathology pattern, however, nor does it predict the response to therapy. Thus, the main role of BAL is to rule out infection. BAL is crucially important in patients with acute or subacute onset ILD, which may correspond to bacterial infection or to ILD related to IIM. It is even more relevant in subjects with progressive disease despite corticosteroid treatment, in order to diagnose rare occult infections including atypical mycobacteria, Pneumocystis jirovecii, Nocardia, intracellular bacteria, cytomegalovirus and fungi, especially in immunocompromised patients.
Attempts have been made to correlate BAL cellular patterns to imaging or histological patterns, or to the clinical presentation. For example, increased lymphocytes and eosinophils in the BAL were higher in dermatomyositis-ILD than in polymyositis-ILD in one study . The CD4+/CD8+ ratio was greater in CADM than in those with classic dermatomyositis and rapidly progressive ILD [62, 70, 87]. Increased neutrophil counts in BAL were associated with a worse outcome in dermatomyositis- or polymyositis-ILD in two studies [47, 63]. However, BAL cellular patterns lack reproducibility and have no clear clinical relevance or treatment implications. In addition, the BAL cellular profile is subnormal or cannot be categorised in a large proportion of cases despite the presence of ILD .
ILD can present with a variety of histological patterns in the setting of IIM. Although the available studies suffer from selection bias, histopathological patterns include in order of decreasing frequency NSIP (65%), UIP (15%), organising pneumonia (9%) and diffuse alveolar damage (8%) according to pooled data from six independent series (table 4) [13, 47, 79, 86, 89, 90], whereas lymphocytic interstitial pneumonia was present in only 1% of patients. The pattern of organising pneumonia may be underrepresented in these series, as clinicians may identify consolidation as a feature associated with a relatively good response to treatment and prognosis, and refer less frequently for surgical biopsy. Two patterns may coexist in various amounts in a patient, with for example mixed patterns of organising pneumonia and NSIP. The patterns found are very similar to those observed in the idiopathic setting, with subtle differences noted including fewer fibroblastic foci, smaller honeycombing spaces, higher germinal centres and total inflammation scores in UIP-associated with CTD than in idiopathic UIP (idiopathic pulmonary fibrosis) . A correlation was reported in one study between pathology and response to treatment, with a good prognosis in cases of organising pneumonia, contrary to those with patterns of diffuse alveolar damage or UIP as expected .
Overall, the pathology found in IIM-ILD is rather similar to that of idiopathic ILD. However, lung biopsy is now rarely performed in patients with CTD and ILD, as it generally does not alter management when compared with decisions based on less invasive procedures, e.g. clinical evaluation, pulmonary function tests, HRCT, and possibly BAL. A lung biopsy is discussed on an individual basis in rare cases with atypical imaging features, or when malignancy is considered.
Results of lung histopathology in patients with dermatomyositis- or polymyositis- associated interstitial lung disease
No randomised controlled trial has been performed to assess the efficacy and tolerance of corticosteroids and immunosuppressive drugs in patients with ILD associated with IIM. The efficacy of immunosuppressive drugs has mainly been reported in uncontrolled small retrospective series, which suggest a benefit of treatment but do not permit firm conclusions to be drawn. Thus, the indication for an immunosuppressive agent and the choice of the molecule are mostly based upon expert opinion, familiarity and experience of the clinician with a particular drug, and the tolerance profile. Treatments have to be initiated in cases of symptomatic or progressive ILD. Asymptomatic ILD must be carefully followed at diagnosis to determinate the disease evolution. Rapidly progressive ILD is poorly sensitive to corticosteroids alone and invariably requires combination with immunosuppressive therapies. Given a mortality rate close to 50%, the aim of therapy is to prevent progression to acute respiratory failure and death [51, 52].
Corticosteroids remain the cornerstone and first-line treatment of dermatomyositis- or polymyositis-associated ILD . Oral regimens include 0.5–1 mg·kg−1 of prednisone daily for 3–8 weeks, followed by a gradual taper. In severe cases, one to three daily pulses of 1000 mg intravenous methylprednisolone are administrated [13