Diagnosis and classification of Idiopathic Inflammatory Myopathies

Abstract

The idiopathic inflammatory myopathies (IIMs) are a heterogeneous group of diseases, collectively named myositis, sharing symptoms of muscle weakness and muscle fatigue and inflammation in muscle tissue. Other organs are frequently involved supporting that these are systemic inflammatory diseases. The IIMs can be sub-grouped into dermatomyositis, polymyositis and inclusion body myositis. The myositis-specific autoantibodies (MSAs) identify other and often more distinct clinical phenotypes, such as the anti-synthetase syndrome with antisynthetase autoantibodies and frequent interstitial lung disease (ILD) and anti-SRP and anti-HMGCR autoantibodies that identify necrotizing myopathy. The MSAs are important both to support myositis diagnosis and to identify subgroups with different patterns of extramuscular organ involvement such as ILD. Another cornerstone in the diagnostic procedure is muscle biopsy to identify inflammation and to exclude non-inflammatory myopathies. Treatment effect and prognosis varies by subgroup. To develop new and better therapies, validated classification criteria that identify distinct subgroups of myositis are critical.. The lack of such criteria was the main rationale for the development of new classification criteria for inflammatory myopathies, which are summarized in this review, along with an historical background on previous diagnostic and classification criteria. As these are rare diseases with a prevalence of 10 in 100 000 individuals an international collaboration was essential, as was the interdisciplinary effort including adult and paediatric experts in rheumatology, neurology, dermatology and epidemiology. The new criteria have been developed based on data from more than 1 500 patients from 47 centers world-wide and are based on clinically easily available variables.

“The Master said,… If names are not correct, language is not in accordance with the truth of things. If language is not in accordance with the truth of things, affairs cannot be carried out to success…. Therefore a superior man considers it necessary that the names he uses be spoken appropriately…. What the superior man requires, is just that in his words there may be nothing incorrect.”

Confucius, Chinese sage and philosopher, 551-479 BC, from Analects, Book XIII

Introduction

The idiopathic inflammatory myopathies (IIM), collectively named myositis, is a heterogeneous group of disorders that are characterized clinically by chronic muscle weakness and low muscle endurance and by inflammatory cell infiltrates in muscle tissue. Based on differences in clinical and histopathological features they have for many years been subclassified into polymyositis (PM), dermatomyositis and inclusion body myositis (IBM) (1). The frequent presence of inflammatory cell infiltrates with a high frequency of T lymphocytes in muscle tissue together with the commonly occurring autoantibodies suggests that these are auto-immune disorders. These are not only muscle disorders as other organs are often affected, such as the skin in DM, and the lungs, heart, joints and gastrointestinal tract, in both PM and DM suggesting that these are systemic autoimmune disease. Interestingly, the newly identified so called myositis-specific autoantibodies are associated with distinct clinical phenotypes as reviewed in detail in the review by Dr Betteridge in this issue of JIM (2). More recently other subsets of IIM have been identified, the clinically amyopathic dermatomyositis (CADM) and the so called immune mediated necrotizing myopathy (IMNM), which is characterized predominantly by muscle fibre necrosis and a specific autoantibodies (anti-SRP or anti-HMGCR) (3,4). With the recent observations it is becoming more convincing that different subsets of IIMs may have different molecular pathways that are leading to different clinical phenotypes and different serotypes and where different therapies may be effective. To get a better understanding of the underlying disease processes the classification of IIM and its different subsets is essential. The progress in the development of new classification criteria will be discussed in this review paper on the historical background of currently used diagnostic and classification criteria used in IIM.

Diagnostic and Classification criteria - what is the difference?

Classification criteria are developed to be used in clinical research in order to make it possible to compare different studies with the same disease. Diagnostic criteria are meant for clinical use to aid the clinician in the diagnostic workup. Sometimes classification criteria are in-correctly used as a diagnostic tool leading to a lack of a diagnosis in some patients. Diagnostic criteria should be based on clinical and laboratory manifestations that are present in the patients early in the disease, and preferentially at presentation to the clinician. Diagnostic criteria should in the best case be developed from a prospectively followed cohort, with registration of early symptoms and signs, and where time will tell if they have the diagnosis in question or not. Classification criteria on the other hand are based on accumulated clinical manifestations and laboratory tests that are or have been present in patients who have developed a disease with high certainty of diagnosis. This could be investigated in a case- control study with retrospectively collected data from cases with IIM and from comparator cases that mimic IIM or have been considered to have IIM. As there is no gold standard for IIM, the treating physician´s diagnosis of IIM could be used as the correct diagnosis. The case-control design has been applied in the development of new classification criteria and the process and preliminary results will be discussed further below.

Diagnosis of adult IIM

The diagnosis of IIM is based on clinical symptoms such as subacute development of symmetrical muscle weakness and muscle fatigue, most prominent in proximal muscles, and signs such as laboratory investigations supporting skeletal muscle inflammation and muscle fibre degeneration and repair (regeneration) (Table 1). The most easily available test to demonstrate skeletal muscle involvement is elevated serum levels of muscle enzymes, the most often tested is creatine phosphokinase (CK), others are lactate dehydrogenase (LD), aspartate transaminase (AST) and alanine transaminase (ALT), and less commonly used is aldolase. Importantly, elevated serum levels of muscle enzyme are not specific for myositis as elevated levels could be seen many other myopathies and normal muscle enzymes do not exclude myositis. Muscle biopsy with histopathological evaluation of frozen muscle tissue by an experienced muscle pathologist is a very useful tool and constitutes the core in the diagnostic work-up of adult IIMs both to confirm skeletal muscle inflammation and to exclude other myopathies. A muscle biopsy is recommended by experts to be mandatory to classify IIM patients without a typical dermatomyositis skin rash. A muscle biopsy is also important to subclassify patients into e.g. IBM, IMNM. Immune-histochemistry staining such as for major histocompatibility complex (MHC) class I, for T cells and macrophages may be helpful to confirm signs of inflammation. Notably, a normal muscle biopsy does not exclude IIM as the inflammatory infiltrates can be patchy. Magnetic resonance imaging (MRI) with T1 and T2 (STIR images) of skeletal muscle may be helpful to identify areas of muscle inflammation by detection of skeletal muscle oedema (5). The sites of inflammation can be used to target muscle biopsies (6). Although the oedema in skeletal muscles on MRI is not specific for myositis it is more commonly seen in myositis in comparison to non-inflammatory myopathies. MRI can also detect muscle damage in the case of fat replacement of muscle tissue or fibrosis. Electromyography (EMG) is another tool to detect myopathies and some changes may also distinguish between necrosis and denervation, such as the size, shape, and recruitment pattern of the motor units potential, although there are no specific EMG findings for myositis.

Table 1.

Diagnostic tools for inflammatory myopathies

Clinical history of muscle weakness or muscle fatigue

Clinical exam: muscle atrophy, weakness, skin rash, joints, lungs, heart

History of previous medication

Family history of rheumatic or muscle disease

Muscle enzymes in serum (CK, LD, AST, ALT and aldolase)

Muscle biopsy from affected skeletal muscle

Magnetic resonance imaging of affected muscle with T1 and T2 (STIR) images

Myositis-specific and myositis-associated autoantibodies (listed in Table 2)

Electromyogram (EMG)

When myositis diagnosis has been confirmed other organ involvement should be searched for:

High resolution computed tomography of lungs

Pulmonary function tests and diffusion capacity

Electrocardiography (ECG)

Echocardiography

Autoantibody testing has become an important tool for diagnosis of IIM and also to identify subgroups of IIM with different clinical phenotypes and prognosis summarized in Table 2. There are so called myositis-associated antibodies, that can be present in other systemic autoimmune diseases, e.g. in systemic lupus erythematosus (SLE), Sjögren´s syndrome or systemic sclerosis, such as anti-Ro52, anti-Ro60, anti-La, anti-U1RNP, PM-Scl and anti-Ku. Although these autoantibodies are not specific for myositis, their presence may be helpful to distinguish an inflammatory myopathy from a non-auto-immune myopathy. There are also so called myositis-specific autoantibodies (MSA) as described in more detail by Betteridge et al (2). These are almost exclusively present in IIM or in subgroups of IIM. The most common of these are anti-Jo-1 antibodies, which are present in approximately 20–25% of patients with PM or DM. The anti-Jo-1 antibodies, targeting histidyl-tRNA synthetase, are strongly associated with distinct clinical manifestations: myositis, interstitial lung disease (ILD), arthritis, fever, Raynaud´s phenomenon and skin rash on the hands called mechanic´s hands. These features, together called the anti-synthetase syndrome (ASS), can also be present without clinical muscle involvement. There are seven more anti-synthetase autoantibodies, listed in Table 2, which can be seen with ASS, some of these autoantibodies are more frequently associated with ILD and some more with myositis. These autoantibodies all target antigens that are ubiquitously expressed in the cytoplasm of all nucleated cells with some variation in expression between different organs that may explain the link e.g. between muscle and lung in the antisynthetase syndrome (7) A new myositis-specific autoantibody has been discovered that targets a muscle specific protein, four and half limb domain 1 (FHL1) (8). This antibody was present in approximately 25% of patients with PM, DM or IBM with the predominate clinical features of severe muscle atrophy and dysphagia but an absence of lung or joint involvement, supporting the hypothesis that different autoantibodies are associated with distinct clinical phenotypes and that different molecular pathways may predominate in different subsets of myositis.

Table 2.

Autoantibodies in idiopathic inflammatory myopathies

Myositis-specific autoantibodies	Myositis-associated autoantibodies
▪ Anti-aminoacyl-tRNA synthetases – Anti-Jo-1 – Anti-PL-7 – Anti-PL-12 – Anti-EJ – Anti-OJ – Anti-KS – Anti-Zo – Anti-YRS ▪Anti-Mi-2 ▪Anti-SRP ▪Anti-TIF1-γ ▪Anti-NXP-2 ▪Anti-MDA5 ▪Anti-SAE ▪Anti-HMGCR ▪Anti-FHL1	▪Anti-SSA/Ro ▪Anti-Ro52 ▪Anti-Ro60 ▪Anti-La ▪Anti-PM-Scl 75 ▪Anti-PM-Scl 100 ▪Anti-Ku ▪Anti-U1RNP ▪Anti-cN-1A

FHL1= four and half limb1; cN-1A= cytosolic 5'-nucleotidase 1A

The golden standard assessment for myositis-specific autoantibodies has been immunoprecipitation, but this technique is not feasible to use in clinical practice in all parts of the world. Several different commercial tests have been developed with varying specificity and sensitivity and in the clinical context it is important to check the validity of a commercial test to avoid over-interpretation of such tests. A line –blot assay with most of the myositis-specific antibodies has been partly validated, and can be used in clinical practice (9) Recently an antibody has been detected in patients with IBM and rarely in the other myositis subgroups called anti-Cytosolic 5'-nucleotidase (CN)-1A (10, 11). As this autoantibody could be detected in other autoimmune disorders such as Sjögren´s syndrome and SLE it should be classified into the MAA. The sensitivity and specificity of this antibody is currently undergoing investigation.

Once a diagnosis of IIM has been made, further investigations are recommended to clarify extra-muscular organ involvement as extra-muscular organ involvement may affect management and prognosis. Particularly the lungs and the gastro-intestinal tract are frequently involved. Therefore investigations of lung involvement with high resolution computerized tomography (HRCT) and pulmonary function tests are recommended even in patients without clinical symptoms from the lungs. Likewise radiography or video radiography of the pharynx and esophagus is recommended as part of the diagnostic workup as swallowing problems may lead to aspiration pneumonia and early death. Heart involvement may also occur and ECG and ECHO cardiogram is recommended to screen for heart muscle involvement in cases with IIM as heart involvement can also be subclinical (12). Adult patients with myositis, and especially those with dermatomyositis have an increased risk of having an associated malignant disease, this is less certain for patients with PM and IBM. Therefore there is a general recommendation to screen for malignancies in patients with dermatomyositis, particularly if the patients have anti-TIF-1gamma antibodies and/or if they have a poor response to conventional immunosuppressive treatment. No specific malignancy is linked to dermatomyositis, thus a general screening is recommended based on sex and age, and with more careful surveillance of elderly individuals.

History of the Classification Criteria for PM/DM

Over the past forty-five years investigators and clinicians have struggled with how to best define and separate the IIM from other conditions, which have similar clinical, laboratory and pathologic features, to result in a more homogeneous group of patients for clinical and laboratory studies. Given the rarity and heterogeneity of these disorders, however, the different myositis classification criteria that have been proposed by the specialists who evaluate and study such patients have many limitations. Because most criteria have been based on expert opinion alone and because different studies have used varying criteria, different populations with possibly dissimilar risk factors and pathogeneses have been studied. This has resulted in the inability to compare epidemiologic studies or trials of different therapies or even of the same therapy. An unfortunate consequence of the lack of consistency among classification criteria for IIM may have decreased interest by the pharmaceutical industry and funding agencies in supporting myositis clinical trials.

The major publications focusing on PM/DM classification criteria over the past 45 years are summarized in Table 3. Some studies included IBM criteria to allow for an overall classification of IIM (for a comprehensive review of the IBM criteria see the chapter by David Hilton-Jones and Stefen Brady (13)). One of the first major attempts to classify PM/DM was accomplished by Medsger et al.(14). Their criteria were based on personal experience and emphasized muscle weakness, muscle biopsy showing inflammation, electromyographic (EMG) abnormalities consistent with myopathy, elevated blood activity of the enzymes associated with myositis and clinical responses to corticosteroids. DeVere and Bradley (15) published their classification criteria in a similar series of cases defined by weakness, muscle pain and tenderness, characteristic myopathic EMG changes, an inflammatory muscle biopsy and elevated serum CK activity.

Table 3.

Summary of features of major proposed criteria for idiopathic inflammatory myopathy^*

Criteria proponents	MW	MPT	MBx	EMG	ENZ	CUT	DEF	MSA	SCF	Reference
Medsger et al.	X		X	X	X		X			(14)
DeVere & Bradley	X	X	X	X	X					(15)
Bohan & Peter	X		X	X	X	X	X			(16,17)
Dalakas	X		X	X	X	X	X	X		(19)
Tanimoto et al.	X	X	X	X	X	X	X	X	X	(20)
Targoff et al.	X		X	X	X	X		X	X	(21)
Dalakas & Hohlfeld	Xa		X	X	X	X			Xb	(24)
VanderMeulen et al.	X		X		X	X			X	(23)
Hoogendijk et al.c	X		Xd	X	X	X			X	(25)
Oddis et al.	X		X	X	X	X	X		X	(26)

^*Modified from (30)

CUT, cutaneous features (V sign, shawl sign, Gottron's papules, heliotrope rash); DEF, disease definition (i.e., definite, probable, possible); EMG, electromyography; ENZ, elevated muscle enzymes (i.e., creatine kinase, aldolase); MBx, inflammation observed in muscle biopsy; MPT, muscle pain and tenderness; MSA, myositis-specific autoantibody (e.g., Jo-1 for (19) and (20) but all available for (21)); MW, proximal muscle weakness; Ref, references; SCF, subclassification (i.e., dermatomyositis, inclusion body myositis, polymyositis, cancer-associated, unspecified, isolated, overlap, amyopathic).

^aMyopathic muscle weakness based on a number of exclusions and inclusions: affecting proximal muscles more than distal ones; sparing eye and facial muscles; characterized by a subacute onset (weeks to months); rapid progression in patients who have no family history of neuromuscular disease; no exposure to myotoxic drugs or toxins; no signs of biochemical muscle disease; pattern distinct from inclusion-body myositis.

^bAmyopathic dermatomyositis is a separate category.

^cDetailed inclusion and exclusion criteria.

^dDetailed pathological features.

The seminal publications by Bohan and Peter (16,17) advanced the field by providing the most comprehensive evaluation criteria for PM and DM, which set the stage for the classification criteria that have been used most extensively during the last four decades (Table 4). Key aspects of their criteria that distinguished them from prior attempts were: the requirement first to exclude all other forms of myopathy; an approach to estimate the certainty of diagnosis of both PM and DM by defining possible, probable and definite disease criteria; the first inclusion of “characteristic rashes” to distinguish DM; more detailed descriptions of each of the other criteria; and definitions for five subgroups of PM/DM, including juvenile, overlap and cancer-associated forms of myositis. Despite the major advances provided by these criteria and their continued use to the present day, several limitations have created difficulties in interpreting the criteria, which has likely resulted in different populations being studied in the various studies that used these criteria. First, the Bohan and Peter criteria resulted from their case series of 153 patients, and data were developed at a single institution based on clinical observations. Second, they did not provide clear instructions for how to rule out all other forms of myopathy, and IBM and many other myopathies had not yet been identified. Third, the number of features needed to fulfil certain criteria was not fully specified, and many of the features proposed are observer-dependent and nonspecific. Importantly, the “characteristic rashes of dermatomyositis” were not specified, which resulted in extensive controversy regarding which of the many rashes that occurs in DM could be used for classification. Finally, the sensitivity and specificity of these criteria for differentiating DM/PM from a wide variety of confounding dermatologic or neuromuscular conditions were not studied. However, a subsequent analysis suggested that they had a sensitivity of 93% and a specificity of 93% for distinguishing PM/DM from systemic sclerosis (Ssc) and systemic lupus erythematosus (SLE) (18).

Table 4.

Bohan and Peter Criteria for Polymyositis and Dermatomyositis^*

First rule out all other forms of myopathy! Symmetrical weakness, usually progressive, of the limb-girdle muscles Muscle biopsy evidence of myositis – Necrosis of type I and type II muscle fibers, phagocytosis, degeneration and regeneration of myofibers with variation in myofiber size, endomysial, perimysial, perivascular or interstitial mononuclear cells Elevation of serum levels of muscle-associated enzymes – CK, Aldolase, LD, Transaminases (ALT/SGPT and AST/SGOT) Electromyographic triad of myopathy – Short, small, low-amplitude polyphasic motor unit potentials – Fibrillation potentials, even at rest – Bizarre high-frequency repetitive discharges Characteristic rashes of dermatomyositis

Definite PM = all first 4, probable PM = 3 of first 4, possible PM = 2 of 4; Definite DM = rash + 3 others; probable DM = rash + 2 others; possible DM = rash + 1 other

^*Modified from (16)

ALT/SGPT, alanine transaminase/serum glutamic pyruvate transaminase; AST/SGOT, aspartate transaminase/serum glutamic oxaloacetic transaminase; CK, creatine kinase; LD, lactate dehydrogenase

Based on personal observations, Dalakas proposed a modification of the Bohan and Peter criteria for PM/DM but also added criteria for IBM and identified important muscle biopsy features that could distinguish these disorders (19). He also suggested that definite disease required all the criteria to be present and that probable disease included all of the criteria except the muscle biopsy.

The next important publication relating to IIM classification criteria was a multispecialty effort in Japan (20). A retrospective study by questionnaire was conducted in dermatology, neurology and rheumatology departments throughout the country. It assessed many disease features that had been proposed in prior criteria sets, but also included new ones that had not been assessed yet but were thought by the respondents to be helpful. Consequently, the following nine features were proposed as criteria for PM/DM: 1) heliotrope rash or Gottron's sign or linear extensor erythema; 2) proximal weakness of the upper or lower extremity and trunk; 3) elevation of CK or aldolase; 4) muscle pain on grasping or spontaneous muscle pain; 5) the electromyographic triad of myopathy; 6) presence of anti-Jo-1 autoantibodies; 7) nondestructive arthritis or arthralgias; 8) signs of systemic inflammation (fever >37°C at the axilla, elevated C-reactive protein or erythrocyte sedimentation rate) and 9) muscle biopsy evidence of myositis (inflammatory infiltrate with degeneration or necrosis of muscle, active phagocytosis, central nuclei or active regeneration). With these criteria, definite PM was defined as any 4 of the 9 features without rash, with a sensitivity of 99% and a specificity against all other diseases of 95%. Definite DM was defined as rash plus 4 other features, with a sensitivity of 94% and specificity against SLE and SSc skin lesions of 90%. Although these criteria had the advantage of having been developed by a national multidisciplinary team, some of the limitations were that the evaluators had varying degrees of experience and training in evaluating myositis, and so the consistency of the evaluation and completeness of the questionnaires were unclear. Additionally, there were limited numbers and types of alternative disorders for comparison, and IBM, cancer-associated myositis and juvenile myositis were not part of the evaluation process. Other limitations were that this study was based on questionnaires and retrospective chart reviews, PM was diagnosed only by rheumatologists and neurologists, DM was diagnosed only by dermatologists and rheumatologists, and apparently Gottron’s papules were not assessed at all.

Targoff and colleagues (21) took a different approach and first asked whether a patient met defined criteria for IIM and, if so, then asked whether the patient met subclassification criteria for the IIM phenotypes of PM, DM or IBM. Their intentions were to modify the Bohan and Peter criteria by adding specificity where needed and to include, for the first time, a role for all the known myositis-specific autoantibodies. They restated that all non-idiopathic inflammatory myopathies should be carefully excluded state-of-the art methods, realizing that these will change over time as technologies and our understanding of these disorders evolves. The six primary classification criteria for IIM that they proposed included the previous five identified by Bohan and Peter—but with DM for the first time specifically defined by heliotrope rash or Gottron’s papules or Gottron’s sign (see Figure 1)—plus any of the myositis-specific autoantibodies then commercially available and performed by a validated assay (e.g., autoantibodies against aminoacyl-tRNA synthetases, chromodomain helicase DNA-binding protein 3 [anti-Mi-2 autoantibodies] or the signal recognition particle [anti-SRP autoantibodies]). Using this approach they defined definite IIM as any 4 of the 6 criteria, probable IIM as any 3 of the 6 criteria, and possible IIM as any 2 of the 6 criteria. Regarding the subclassification criteria, DM was defined by the presence of any one of the rashes named above and PM by the absence of those rashes. IBM was defined by also meeting the criteria of Griggs et al.(22), and juvenile myositis was defined by age at onset <18 years.

Figure 1.

Pathognomonic rashes of DM: Gottron’s papules, red or violet papules occurring over the knuckles, the interphalangeal joints and other extensor surfaces (A); Gottron’s sign, red or violet macules occurring over the knuckles and sometimes the interphalangeal joints in patients with dermatomyositis (B); and heliotrope rash, red or violate eruption over the upper eyelids and sometimes extending around the eye (C).Reproduced from (31) with permission.

In a retrospective evaluation, Van der Meulen et al. (23) combined prior criteria with a critical pathology focus and concluded that only a small number of patients suspected of having PM actually met their criteria, which included: subacute onset (<1 year); symmetric, proximal more than distal muscle weakness, or muscle soreness; CK > 2 times the upper limit of normal; and mononuclear cells (MNC) surrounding and ideally invading non-necrotic endomysial myofibers. Definite DM required the “typical skin rash” or perifascicular atrophy on muscle biopsy. Of interest, they found that 23% of the patients studied did not meet these criteria, and so they created a new category called “unspecified myositis”. These subjects had a perimysial or perivascular MNC infiltrate but without endomysial MNC surrounding and invading non-necrotic fibers or perifascicular atrophy or rash. Possible myositis, defined as CK > 2 times the upper limit of normal and necrotizing myopathy with few or no MNC on biopsy, was identified in 18% of their patients. They concluded that PM is over diagnosed and that other categories of myositis should be established. A limitation of their criteria is that they established a new, large category of myositis patients with “unspecified myositis” for which there was no information on how to assess or treat them.

In 2003, Dalakas and Hohlfeld (24) updated the Bohan and Peter criteria by adding more pathologic focus and details, as well as including criteria for amyopathic DM. Their muscle biopsy criteria were as follows: for definite PM, primary inflammation with the CD8/MHC-1 complex and no vacuoles; for probable PM, ubiquitous MHC-I expression but no CD8+ cell infiltrates or vacuoles; for definite DM, perifascicular, perimysial or perivascular infiltrates, perifascicular atrophy and rash present; for probable DM, perifascicular, perimysial or perivascular infiltrates, perifascicular atrophy but no rash present; and for amyopathic DM, a rash is present but biopsy findings are nonspecific or are diagnostic for DM, and no weakness is present.

As part of the European Neuromuscular Centre (ENMC) and Muscle Study Group (MSG) workshops, a group of myologists proposed criteria for IIM, excluding IBM, based on expert opinion (25). Table 5 summarizes these criteria elements and Table 6 describes how to apply these to each category of myositis (25). These criteria are unique in that they include a detailed list of inclusion and exclusion elements to be applied to clinical, laboratory (including magnetic resonance imaging and myositis-specific autoantibodies) and pathologic features.

Table 5.

Elements of the classification criteria for the idiopathic inflammatory myopathies (except IBM) approved by the Myositis Study Group and the 119^th European Neuromuscular Centre workshop^*

Clinical criteria Inclusion criteria Onset usually over 18 years (post-puberty), onset may be in childhood in DM and non-specific myositis Subacute or insidious onset Pattern of weakness: symmetric Proximal > distal, neck flexor > neck extensor Rash typical of DM: heliotrope (purple) periorbital edema; violaceous papules (Gottron’s papules) or macules (Gottron’s sign), scaly if chronic, at metacarpophalangeal and interphalangeal joints and other bony prominences; erythema of chest and neck (V-sign) and upper back (shawl sign) Exclusion criteria Clinical features of IBM (see Griggs et al. [9]: asymmetric weakness, wrist/finger flexors same or worse that deltoids; knee extensors and/or ankle dorsiflexors same or worse than hip flexors) Ocular weakness, isolated dysarthria, neck extensor > neck flexor weakness Toxic myopathy (e.g., recent exposure to myotoxic drugs), active endocrinopathy (hyper- or hypothyroid, hyperparathyroid), amyloidosis, family history of muscular dystrophy or proximal motor neuropathies (e.g., SMA) Elevated serum creatine kinase level Other laboratory criteria Electromyography Inclusion criteria Increased insertional and spontaneous activity in the form of fibrillation potentials, positive sharp waves, or complex repetitive discharges Morphometric analysis reveals the presence of short duration, small amplitude, polyphasic MUAPs Exclusion criteria Myotonic discharges that would suggest proximal myotonic dystrophy or other channelopathy Morphometic analysis reveals predominantly long duration, large amplitude MUAPs Decreased recruitment pattern of MUAPs MRI: diffuse or patchy increased signal (oedema) within muscle tissue on STIR images Myositis-specific antibodies detected in serum Muscle biopsy inclusion and exclusion criteria Endomysial inflammatory cell infiltrate (T-cells) surrounding and invading non-necrotic muscle fibers Endomysial CD8+ T-cells surrounding but not definitely invading non-necrotic muscle fibers, or ubiquitous MHC-1 expression Perifascicular atrophy MAC depositions on small blood vessels, or reduced capillary density, or tubuloreticular inclusions in endothelial cells on EM, or MHC-1 expression of perifascicular fibers Perivascular, perimysial inflammatory cell infiltrate Scattered endomysial CD8+ T-cell infiltrate that does not clearly surround or invade muscle fibers Many necrotic muscle fibers as the predominant abnormal histological feature. Inflammatory cells are sparse or only slightly perivascular; perimysial infiltrate is not evident. MAC deposition on small blood vessels or pipestem capillaries on EM may be seen, but tubuloreticular inclusions in endothelial cells are uncommon or not evident. Rimmed vacuoles, ragged red fibers, cytochrome C oxidase-negative fibers that would suggest IBM MAC deposition on the sarcolemma of non-necrotic fibers and other indications of muscular dystrophies with immunopathology

^*Modified from (25)

DM, dermatomyositis; EM, electromyography; IBM, inclusion body myositis; MAC, membrane attack complex; MHC, major histocompatibility complex; MRI, magnetic resonance imaging, MUAP, motor unit action potential; SMA, spinal muscular atrophy; STIR, short-tau inversion recovery.

Table 6.

Scoring of the classification criteria elements for the idiopathic inflammatory myopathies (except IBM) to define different myositis phenotypes as approved by the Myositis Study Group and the European Neuromuscular Centre workshop^*

Definite polymyositis All clinical criteria with the exception of rash Elevated serum CK Muscle biopsy criteria include a and exclude c, d, h, i† Probable polymyositis All clinical criteria with the exception of rash Elevated serum CK Other laboratory criteria (1 of 3) Muscle biopsy criteria include b and exclude c, d, g, h, i Definite dermatomyositis All clinical criteria Muscle biopsy criteria include c Probable dermatomyositis All clinical criteria Muscle biopsy criteria include d or e, or elevated serum CK, or other laboratory criteria (1 of 3) Amyopathic dermatomyositis Rash typical of DM: heliotrope, periorbital oedema, Gottron’s papules or sign, V-sign, shawl sign, holster sign Skin biopsy demonstrates a reduced capillary density, deposition of MAC on small blood-vessels along the dermal–epidermal junction, and variable keratinocyte decoration for MAC No objective weakness Normal serum CK Normal EMG Muscle biopsy, if done, does not reveal features compatible with definite or probable DM Possible dermatomyositis sine dermatitis All clinical criteria with the exception of rash Elevated serum CK Other laboratory criteria (1 of 3) Muscle biopsy criteria include c or d Non-specific myositis All clinical criteria with the exception of rash Elevated serum CK Other laboratory criteria (1 of 3) Muscle biopsy criteria include e or f and exclude all others Immune-mediated necrotizing myopathy All clinical criteria with the exception of rash Elevated serum CK Other laboratory criteria (1 of 3) Muscle biopsy criteria include g and exclude all others

^*Modified from (25)

^†Letters refer to those elements in Table 3.

CK, creatine kinase; DM, dermatomyositis; EMG, electromyography; IBM, inclusion body myositis; MAC, membrane attack complex.

A comparison of performance of some of the main criteria used to classify IIM including the ENMC criteria was performed in a small single centre cohort (27). The study confirmed the limitations of existing criteria and emphasized the need to develop new and improved classification criteria for IIM.

Oddis et al.(26), as part of the efforts of the International Myositis Assessment and Clinical Study Group (IMACS) to develop multidisciplinary consensus for conducting and reporting myositis trials, also suggested modifications to the Bohan and Peter criteria that defined Gottron’s papules or heliotrope rash as the required rashes for DM and suggested that a muscle biopsy consistent with PM was also a necessary criterion for PM.

The IMCCP criteria - how it started and a summary of the methodology and results

The international interest group, International Myositis Assessment and Clinical Studies group (IMACS) was established in 1999 to develop outcome measures for use in clinical trials and to facilitate collaborative myositis research (28). Through this collaborative network it was also apparent that there was a strong need to develop new classification criteria where novel technologies such as immunohistochemistry staining of muscle biopsies, MRI of skeletal muscles and new myositis-specific autoantibodies were taken into account and to be used for clinical and molecular studies on myositis. Thus a group of experts in myositis and epidemiology met in 2004 to initiate a project to define new classification criteria for myositis that have high sensitivity and specificity and started what resulted in a long project that is now coming to the end.

It was clear that the myositis criteria project needed to be multidisciplinary and international and there was also an aim, if possible, to develop common criteria for paediatric and adult myositis patients. This project was defined as the International Myositis Classification Criteria Project (IMCCP), with an international steering group representing adult and paediatric rheumatology, neurology, dermatology and epidemiology, furthermore, with representatives from the US and Europe. We decided to seek endorsement from major scientific organisations including American college of Rheumatology (AC), European League Against Rheumatism (EULAR), the Muscle Study Group, The International Assessment and Clinical Studies (IMACS), childhood arthritis and rheumatology research alliance (CARRA) paediatric rheumatology European society (PReS) network for JDM and the paediatric rheumatology international trials organisation (PRINTO). The major aims of the project were to develop new classification criteria that should distinguish patients with idiopathic inflammatory myopathies (IIM) from non-IIM and, furthermore, that the criteria should identify subgroups of IIM. It was decided that the process should be data driven and include both paediatric and adult cases and comparator cases with non-IIM. The process started by identifying variables to be collected based on previously published criteria for IIM including IBM and additional variables were added based on expert opinion with contributions from myologists within the large IMACS network. The decision on which variables to be included was based on discussions with project participants using nominal group technique and was finally made by the IMCCP steering group. All variables to be included in the project were then operationally defined. Inclusion criteria for cases and comparators were defined as: i) diagnosis for at least 6 months prior to study inclusion; ii) physician certainty of diagnosis – either known IIM or, as comparators, known non-IIM cases where myositis was considered in the initial differential diagnosis; and iii) patients with the most recent and complete data were prioritized to acquire the most complete data in a consistent manner. A maximum of 40 cases and an equal number of comparators was collected from each center, for pediatric cases a minimum of 5 cases and comparators was required per center. As there is no gold standard for diagnosis of IIM, it was decided that the diagnosis of the expert clinician who enrolled patients should be used as the diagnosis for enrolled patients.

An electronic data base was constructed that included the basic demographics of patients, and totally 93 variables. As a result of the contribution of myositis experts from 47 centres world-wide (23 European, 17 North American, one South American, and six Asian sites) we collected data from 976 IIM patients and 624 comparators (adults and children). Importantly, 20% of the cases and comparators are of Asian origin. The patients represent the major current clinical subgroups of IIM: JDM, PM, DM, IBM, amyopathic DM (ADM), hypomyopathic DM, immune-mediated necrotizing myopathy (IMNM) and juvenile PM. The comparator group consisted of a broad spectrum of conditions that can mimic IIM from rheumatology, pediatric rheumatology, neurology and dermatology clinics.

After a long period to collect data from the cases and comparators the analytical phase started. With the large amount of data a biostatistician with experience from similar work was recruited to the IMCCP project, Professor M Bottai, Karolinska Institutet. As a first step we wanted to identify variables that distinguish IIM from non-IIM with a high sensitivity and specificity. Three classification techniques were explored first a traditional a sum-of-items model, in which a patient was classified as case if the patient had a specified number of items from a set of items similar to e.g. the Bohan and Peter criteria; secondly a probability-score model; and thirdly, a classification tree. The ensuing candidate criteria were examined with respect to statistical performance and clinical relevance. Sixteen variables, which had the best capacity to discriminate IIM from non-IIM using a probability model, were identified (Table 7) (29). Furthermore, the 16 variables were found to have different importance and were therefore given different weights or score. In addition, there was a request among the investigators to test if we could develop one set of criteria without requiring muscle biopsies. This was, in particular, in the context of children with a typical dermatomyositis skin rash. Therefore two versions of the criteria were tested one with and one without muscle biopsies. The best statistical performance with a balance between sensitivity and specificity was found with a 55% probability cut-off for both models. The probability model with the different weights of the variables offers flexibility in the number of variables that needs to be tested. The 55% probability is recommended for research studies such as register based studies or a natural history study, where the sensitivity is important, whereas in a clinical trials a high specificity is recommended, which is a high probability score. Thus the experts in the steering committee recommend a high probability, 90%, to be used in clinical trials. This is comparable to defining different levels of sensitivity and specificity using probable for 55% probability or definite 90% probability for diagnosis. Importantly, the steering committee recommends that for patients without skin rash typical for dermatomyositis the criteria version with muscle biopsy variables should be used. The new criteria were compared to previously developed diagnostic or classification criteria, and performed superior to all criteria concerning sensitivity and specificity with one exception that is the Targoff criteria (21) that performed equally well. However, the Targoff criteria require more variables to be tested e.g. both muscle biopsy and EMG is required, so they are less flexible. Moreover, in the Targoff criteria the variables were not well-defined and the Targoff criteria did not include all subsets of IIM.

Table 7.

Variables used in the development of new classification criteria for idiopathic inflammatory myopathies (29)

Variable
Age of onset of first symptom assumed to be related to the disease ≥ 18 years and < 40 years
Age of onset of first symptom assumed to be related to the disease ≥ 40 years
Objective symmetric weakness, usually progressive, of the proximal upper extremities
Objective symmetric weakness, usually progressive, of the proximal lower extremities
Neck flexors are relatively weaker than neck extensors
In the legs proximal muscles are relatively weaker than distal muscles
Heliotrope rash
Gottron´s papules
Gottron’s sign
Dysphagia or esophageal dysmotility
Anti-Jo-1 (anti-histidyl-tRNA synthetase) autoantibody present
Elevated serum levels of creatine kinase (CK)* or lactate dehydrogenase (LDH)* or aspartate aminotransferase (ASAT/AST/SGOT)* or alanine aminotransferase (ALAT/ALT/SGPT)*
Muscle biopsy features- presence of:
Endomysial infiltration of mononuclear cells surrounding, but not invading, myofibres
Perimysial and/or perivascular infiltration of mononuclear cells
Perifascicular atrophy
Rimmed vacuoles

^*Serum levels above the upper limit of normal

After having classified a patient as IIM, a classification tree was developed to identify subclasses of IIM. Through the subclassification tree juvenile DM, adult DM, clinically amyopathic DM, (CADM) IBM and PM could be identified. However, there were too few patients in our cohort to identify the newly defined subgroup immune mediated necrotizing myopathy, thus these patients are to be found within the PM subgroup. Likewise there were too few cases to identify hyopmyopathic DM. and there were not enough non-JDM IIM cases to identify variables for other subgroups of juvenile myositis. Using extrapolation from the adult IIMs in the classification tree a sub-group of juvenile myositis other than JDM was suggested. The new criteria with the probability score and the subclassification tree have been validated using internal validation using boot strap method and tested for sensitivity in external cohorts with myositis cases with very good results. Further validation will be needed in external cohorts.

The new classification criteria are currently under review by the ACR and EULAR criteria subcommittees, as we are seeking endorsement from these organizations, whereafter they will be submitted for publication.

To support the investigators in making calculations of the probability score a web- calculator has been designed. The web-calculator is available at “http://www.imm.ki.se/biostatistics/calculators/iim” and maintained by the Unit of Biostatistics, Karolinska Institutet, Stockholm, Sweden. With the web-calculator one can enter the variables and the calculator will give you a probability score for the classification of myositis that will increase and narrow down when you add information on your patient. The web-calculator will also give you a subgroup of myositis if sufficient information is available. The web-calculator will be available as an app to upload on smartphones. The uploaded data can be stored in excel sheets.

Discussion with recommendations

The new myositis classification criteria are data driven, based on patients and comparators from many centres world-wide representing different ethnicities. Thus the patients should be representative of various clinical phenotypes of myositis. The variables in the final version of the criteria are easily available, clinically relevant and collected in the routine work-up in most centres. Importantly these are classification criteria that were developed on information from patients with established disease, and thus they are not recommended for diagnosis. The probability model that has been developed is flexible, meaning that you do not need to test all variables to reach a high probability. However, in patients without typical dermatomyositis skin rash a muscle biopsy is considered necessary both to confirm signs of inflammation and exclude other differential diagnoses. Furthermore, these criteria apply to cases with a suspicion of myositis when no better explanation for the symptoms exists. Surprisingly only one of the myositis-specific antibodies was included in the new classification criteria. This can be explained by the long process from when we started to collect the information on the patients, as most of the new myositis-specific autoantibodies were not available as routine tests when data were collected. Thus a revision of the criteria is recommended to take place within a few years when we should have more consistent information on serology. EMG and magnetic resonance imaging (MRI) of muscles were not included in the new criteria, reflecting that they are not often used in clinical practice among the 47 centres that contributed with cases for this study, and interestingly we could classify patients correctly with a high sensitivity and specificity based on the variables in the new proposed classification criteria for myositis.

In summary, many sets of diagnostic and classification criteria for different forms of IIM have been proposed over five decades, each with specific advantages and disadvantages. Most have been based on clinical impressions rather than data analyses, and none has been fully tested for sensitivity and specificity against all the appropriate disease confounders in adequately powered studies (29). Because of these limitations and recent disagreements among rheumatologists, neurologists, dermatologists and others in the appropriate use of varying criteria for IIM that would create difficulties in comparing studies and clinical trials in the future, many myologists concluded that large, multicenter and multispecialty studies, that include comparisons of IIM with many confounding conditions, are needed to define and validate new IIM classification criteria. The result of this multidisciplinary effort is a new set of classification criteria that have been partly validated with good results. More validation in external cohorts with comparator cases is needed.