Re-classifying myositis

30 years on from Bohan and Peter Over the past 40 years, almost a dozen attempts have been made to establish diagnostic and classification criteria for patients with idiopathic inflammatory myopathies (IIMs). In 1970, Medsger et al. [1] proposed five criteria similar to the Bohan and Peter criteria. However, the diagnostic criteria included responsiveness to steroids, excluding those patients who failed to respond to corticosteroids alone and yet subsequently responded to immunosuppressive therapy. De Vere and Bradley's [2] proposed criteria consisted of a combination of weakness, pain and tenderness, biopsy, EMG and creatine kinase level. However, the authors did not assess specificity against patients with non-inflammatory myopathies. Over 30 years ago, Bohan and Peter [3, 4] discussed two main topics: the problem of diagnosis and the problem of classification. However, these two terms have since been used interchangeably. They highlighted the need for clearly defined criteria for diagnosis and suggested a scheme in an attempt to establish such guidelines (Table 1)—‘These criteria are empirically derived …’ [3]. No data were provided for the sensitivity or specificity of these criteria, but they were explicit that other conditions needed to be excluded (e.g. motor neurone disease, muscular dystrophy, myasthenia gravis, infections, granulomatous myositis such as sarcoidosis, toxic causes from drugs and alcohol, metabolic disorders and endocrinopathies). Unfortunately, no instruction was given as to how to exclude other forms of myopathy. Most of the criteria are non-specific including the characteristic rashes of DM. Bohan and Peter criteria for the diagnosis of PM and DM Definite DM requires four criteria (including rash) and definite PM requires four criteria (without rash). Probable disease comprises three criteria (including rash) for DM and three criteria (without rash) for PM. Possible disease requires two criteria (including rash) for DM and two criteria (without rash) for PM. Bohan and Peter criteria for the diagnosis of PM and DM Definite DM requires four criteria (including rash) and definite PM requires four criteria (without rash). Probable disease comprises three criteria (including rash) for DM and three criteria (without rash) for PM. Possible disease requires two criteria (including rash) for DM and two criteria (without rash) for PM. In many rheumatic diseases, classification criteria have been developed to ensure meaningful comparison of patient populations both to predict the course of a disease and response to treatment. Several classification schemes had been popular, but Bohan and Peter [3] offered their work as an alternative approach. Their classification system was (i) primary idiopathic PM, (ii) primary idiopathic DM, (iii) PM or DM with malignancy, (iv) juvenile DM (or PM) associated with vasculitis and (v) PM or DM with associated collagen vascular diseases. However, they did not recognize IBM or other inflammatory myopathies, such as granulomatous and eosinophilic myositis. Clinicians widely accept that modern-day histopathology and genetic techniques need to exclude IBM, hereditary and other myopathies when applying the Bohan and Peter criteria. Several classification systems or modifications of the original classification criteria have since been proposed for IIM, but disagreement still continues. The new criteria that have been suggested have either been based on autoantibody subgroups [5–7] or on muscle histopathology [8]. Love et al. [5] suggested a classification based on several specific autoantibodies that define subgroups of patients with their clinical features, response to therapy and prognosis. Of the 212 patients included in their study, the clinical groups classified by Peter and Bohan differed relatively little from one another when examined by demography, signs and symptoms, HLA antigens and prognosis. The only exception was IBM. In contrast, they found that unique serological abnormalities of IIM, the myositis-specific autoantibodies (MSAs), each define more distinctive groups of IIM than the clinical groups for signs and symptoms and immunogenetics, and prognosis. However, few of these autoantibodies are available for routine clinical testing, which severely limits the practicality of their otherwise useful approach. Tanimoto et al. [6] included all five of the Bohan and Peter criteria and added four new criteria. Three of these criteria—arthritis and arthralgias, systemic inflammatory signs (fever, high ESR and CRP) and muscle pain or tenderness—are common features of many rheumatic diseases that are not found in non-systemic myopathies or neuropathies. Thus, they are not helpful in establishing a diagnosis in an individual patient, but may serve to distinguish myositis from the comparison group used in the study [7]. The fourth proposed criterion, the presence of anti-Jo-1 antibodies, represents an important addition, but it is only one of the several currently recognized MSAs now reported in IIM (Table 2). However, the diagnosis of definite PM or DM can be made without a disease-specific criterion by using the three criteria intended to identify inflammation. There were, however, few patients with alternative disorders such as cancer-associated myositis; moreover, patients with juvenile myositis and IBM were excluded. Myositis-specific antibodies in PM/DM (modified from Targoff [12]) NA: no available percentage yet in the literature. Myositis-specific antibodies in PM/DM (modified from Targoff [12]) NA: no available percentage yet in the literature. Targoff et al. [7] suggested that the five Bohan and Peter criteria and the designation as definite, probable and possible disease could be retained with minimal change. One criterion could be added: MSA (an anti-synthetase, anti-Mi-2 or anti-SRP antibodies). In addition, it would be possible to substitute MRI evidence of muscle inflammation for the weakness or enzyme criteria, but this could not be used in addition to these criteria or in place of EMG, biopsy, rash or antibody criteria. They suggest that this limitation is necessary because the specificity of oedema on MRI for IIM is not yet established. However, this proposed restriction does cause difficulty in childhood onset IIM in which MRI scanning is increasingly being used to replace biopsy. With the revised criteria, a patient must still satisfy one of the following disease-specific criteria to qualify as having definite IIM: a compatible biopsy, cutaneous DM features or a myositis-specific autoantibody. This revision would ensure diagnosis of definite PM without a biopsy if an MSA is identified and definite DM without both EMG and muscle biopsy. The committee of American Academy of Dermatology published guidelines of care for DM [9] with recommendations for diagnostic evaluation. They suggested that DM is confirmed either by diagnostic skin lesions or by characteristic muscle features with less specific skin findings. Gottron's sign and papules are considered diagnostic, whereas heliotrope and other rashes are considered to be less specific. A skin biopsy is recommended but not required. Gottron's sign or papules alone serve to establish the diagnosis of amyopathic DM. The Bohan and Peter [3] biopsy criterion (Table 1) is also considered too inclusive and may allow patients with some forms of muscle dystrophy to be classified as PM; as in the early stages of muscular dystrophy inflammatory changes may be observed, which can be very similar to PM. These criteria also misclassify patients with IBM (which was not clearly described when these criteria were proposed) as having PM. Other criteria for IIM have been put forward based solely on histological criteria. Engel and Arahata [10] proposed an immunohistopathology-based criteria. A further classification proposed by Dalakas [8] is based on immunopathological features on muscle biopsy. Classification criteria have been developed for IBM. This phenotype of IIM is largely unresponsive to immunosuppressive treatment and patients should be regarded as a separate entity in clinical trials [11]. These criteria are based on a combination of: (i) clinical features such as duration of >6 months, age of onset >50 years, pattern of weakness (finger flexor weakness, wrist flexor weakness more than extensor weakness and weakness of quadriceps); (ii) laboratory features such as serum CK less than 12 times normal range, muscle biopsy (showing mononuclear inflammatory cell invasion of non-necrotic muscle fibres, vacuolated muscle fibres with either intracellular amyloid deposits or tubulofilaments by electronmicroscopy); and (iii) EMG features of an inflammatory myopathy that may have long-duration action potentials. In definite IBM, patients must exhibit all muscle biopsy features (in which case none of the clinical or laboratory features are required) or possible IBM; patients may show only inflammation and invasion of fibres without vacuoles, and meet all clinical and other laboratory criteria (but may not fulfil the muscle biopsy criteria). The Bohan and Peter criteria have also been used for children. However, many paediatric rheumatologists do not use EMG or muscle biopsy criteria with myositis. Recently, a group of paediatric rheumatologists have started to revise the criteria for this group (PreS network for juvenile DM). Although the Bohan and Peter criteria are still widely used, there is no formal international consensus. However, most physicians believe that there should be a collaborative effort between specialists in neurology, rheumatology, paediatric rheumatology, dermatology, pathology and epidemiology and that it should be international in scope to ensure greatest acceptance. For these reasons, members of International Myositis Assessment and Clinical Studies (IMACS) and other collaborating groups have initiated an international multidisciplinary project to develop classification criteria for adult and juvenile myositis. The two primary goals for this project are: In this project, 50 variables were considered using the nominal group technique in five categories: (i) clinical muscle variables (weakness, muscle pain at rest, muscle tenderness and muscle atrophy); (ii) skin variables (heliotrope, Gottron's, shawl sign, periungal changes, calcification and RP); (iii) other clinical variables (onset, family history, other CTDs and steroid response); (iv) laboratory variables (including EMG, MRI, muscle enzymes and autoantibodies); and (v) muscle biopsy variables. The study is an ongoing retrospective analysis of clinical cases. The disorders to be considered are the subgroups of IIM (such as PM, DM, overlap myositis, etc.) and non-IIM that might mimic IIM (such as non-inflammatory IBM, dystrophies, motor neurone and other neurological diseases, drug or toxin-associated myopathies, metabolic myopathies, mitochondrial myopathies, infectious myopathies, endocrine myopathies, CTDs, PMR and dermatological conditions). to develop classification criteria that distinguish IIM from other major mimicking conditions with high sensitivity and specificity; and to develop classification criteria that separates the major subgroups of IIM from each other with high sensitivity and specificity, and to document the reliability of these new criteria. Bohan and Peter were aware that their early suggestions needed further work, and it has taken many decades to take this work forward. The current international collaboration should hopefully produce criteria that are widely accepted across communities in rheumatology, dermatology and neurology, and should improve the diagnosis and classification of IIMs. Disclosure statement: The authors have declared no conflicts of interest.