New Classification Criteria for Systemic Sclerosis (Scleroderma)




The American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) criteria for the classification of systemic sclerosis (SSc) were developed to classify more patients with SSc for studies and improve on previous criteria. The new classification criteria have the following criteria: skin thickening of the fingers extending proximal to the metacarpophalangeal joints. Seven additive items are each given a numerical weight: skin thickening, fingertip lesions, telangiectasia, abnormal nail fold capillaries, lung involvement, Raynaud’s phenomenon, and SSc-related autoantibodies. The ACR/EULAR classification criteria for SSc have good sensitivity and specificity but do not substitute for diagnostic criteria.


Key points








  • The American College of Rheumatology/European League Against Rheumatism classification criteria for systemic sclerosis (SSc) are more sensitive and specific than the previous 1980 American Rheumatism Association criteria for SSc.



  • Classification criteria are not diagnostic criteria and were developed for clinical studies, but most patients diagnosed with SSc should meet classification criteria.



  • There is no gold standard for SSc diagnosis and the criteria should only be applied if SSc diagnosis is suspected.






Introduction


Systemic sclerosis (SSc; also called scleroderma) refers to an autoimmune connective tissue disease with autoantibodies, vasculopathy, and fibrosis. Most patients have sclerodactyly (thickened skin of the fingers) and may or may not have more extensive skin fibrosis. There are large differences in prevalence of internal organ involvement and other features, although nearly all have Raynaud phenomenon (RP), many have esophageal dysmotility, and 8% to 50% have pulmonary arterial hypertension (PAH), cardiac involvement, interstitial lung disease, inflammatory arthritis, and digital ulcers. There is no laboratory test to diagnose SSc. With heterogeneous features and no diagnostic gold standard, it is important to classify patients with SSc along the disease spectrum as accurately as possible. Because SSc is rare, if a significant proportion of patients are not included in the classification then they will be missed when studying the disease.


Classification criteria are not diagnostic criteria. However, classification criteria should be similar to diagnostic criteria (with a large overlap of patients included in both criteria). This article discusses the evolution of SSc classification criteria, strengths and weaknesses, and how to use the new American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) SSc classification criteria.




Introduction


Systemic sclerosis (SSc; also called scleroderma) refers to an autoimmune connective tissue disease with autoantibodies, vasculopathy, and fibrosis. Most patients have sclerodactyly (thickened skin of the fingers) and may or may not have more extensive skin fibrosis. There are large differences in prevalence of internal organ involvement and other features, although nearly all have Raynaud phenomenon (RP), many have esophageal dysmotility, and 8% to 50% have pulmonary arterial hypertension (PAH), cardiac involvement, interstitial lung disease, inflammatory arthritis, and digital ulcers. There is no laboratory test to diagnose SSc. With heterogeneous features and no diagnostic gold standard, it is important to classify patients with SSc along the disease spectrum as accurately as possible. Because SSc is rare, if a significant proportion of patients are not included in the classification then they will be missed when studying the disease.


Classification criteria are not diagnostic criteria. However, classification criteria should be similar to diagnostic criteria (with a large overlap of patients included in both criteria). This article discusses the evolution of SSc classification criteria, strengths and weaknesses, and how to use the new American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) SSc classification criteria.




Previous systemic sclerosis classification criteria


Classification criteria have been published in SSc. Classification criteria are used to include patients with a similar clinical entity for research studies, but are used in a clinical setting to help to classify patients. The American Rheumatism Association (ARA) 1980 preliminary scleroderma (SSc) criteria were the most commonly cited criteria for SSc. The 1980 preliminary scleroderma criteria were an important step in SSc classification. There was 1 sufficient (major) criterion and 3 minor criteria. The major (sufficient) criterion was proximal cutaneous sclerosis/skin thickening (nonpitting) of the fingers that also extended proximal to the metacarpophalangeal joints (MCPs). If this criterion was not met, then at least 2 of 3 items had to be present: (1) sclerodactyly, (2) digital pitting scars of fingertips or loss of substance of the distal finger pad (digital tuft resorption), and (3) bibasilar pulmonary fibrosis.




Limitations of previous criteria


There are currently more patients with SSc who are at the mild end of the spectrum and in the limited cutaneous SSc subset (lcSSc) compared with patients in whom the 1980 criteria were tested, in whom a larger proportion of patients with diffuse cutaneous SSc (dcSSc) were identified; possibly because of the evolution of SSc over time, more recognition caused by available commercial autoantibodies, and/or earlier diagnosis. Many patients whom experts would now classify as SSc were not classified in the 1980 criteria. For example, a patient with sclerodactyly, telangiectasia, calcinosis, RP, proven PAH, and a positive anticentromere antibody test would not be classified as having SSc in the 1980 scheme. Patients with sclerodactyly, RP, anticentromere antibodies, dysphagia, dilated nail fold capillaries, and calcinosis would also not meet the previous criteria for SSc classification.


Often 20% of patients with lcSSc did not meet the criteria and were excluded from clinical studies. This percentage means that 1 in 5 patients with lcSSc would not be classified with SSc even though they had a diagnosis of SSc.




Other previously proposed systemic sclerosis classification criteria


LeRoy and colleagues proposed criteria that included clinical features, autoantibodies, and capillaroscopy. In 2001, LeRoy and Medsger suggested revising the classification criteria to include early cases of SSc, using nail fold capillary changes and SSc-specific autoantibodies. The addition of nail fold capillary abnormalities and telangiectasia to the 1980 SSc criteria improves their sensitivity. Adding RP to the 1980 criteria for SSc increased the number of patients classified with SSc.




The American College of Rheumatology/European League Against Rheumatism systemic sclerosis classification criteria


The ACR and the EULAR jointly funded the development of new SSc criteria, which was conducted in multiple phases. Candidate items for the classification criteria were identified through surveys of all potential items, followed by a Delphi process that reduced the 168 original features to 23 items. Patients from international SSc cohorts were compared with control patients with diseases similar to SSc (mimickers) to test the discriminative ability of the items. Next, SSc cases were purposively sampled from 2 European and 3 North American SSc centers to reflect the spectrum of probability that the subject could have SSc (ie, cases that had high probability, low probability, and in between). Data on the 23 candidate criteria for each of these cases were collected using a standardized form and provided to experts in a randomized order who ranked the cases from highest to lowest probability of having SSc. A consensus meeting helped to define to whom the criteria should be applied (inclusion criteria: patients suspected of having SSc) and should not be applied (exclusion criteria: patients in whom a different diagnosis is far more likely and patients who have had skin fibrosis but without finger involvement), and which manifestations/items are sufficient to allow a patient to be classified as having SSc (sufficient criteria: sclerodactyly with continuous skin fibrosis proximal to the MCPs). Multicriteria decision analysis helped to determine the relative weighting of each of the items. Some items were deleted and some items were clustered (ie, the individual autoantibodies were grouped into an item about presence of SSc-related antibodies). Weights of items were rounded to whole numbers. Cases were reranked with the scores of each item to determine a numerical threshold to exclude SSc cases that were less probable. There were cases for which all SSc experts thought the patients had SSc and others for which all agreed they did not have SSc, but there were also cases for which some experts thought the patient had SSc and others did not. Further cases were obtained and tested in this gray zone to identify a threshold that maintained specificity and sensitivity. A numerical range of the total score in this gray zone was tested in cases (patients with incident and prevalent SSc) and controls (SSc mimickers) that were prospectively collected at multiple sites and the data were divided randomly into derivation and validation cohorts. Low-frequency and statistically redundant items were removed. The resulting criteria were tested in the derivation sample of cases and controls and validated in the other sample. Table 1 shows the ACR/EULAR SSc criteria. The sensitivity and specificity of the criteria were higher than for the 1980 criteria. In the validation sample, the 1980 criteria had a sensitivity of 75% and specificity of 73%, whereas the new criteria had greater than 90% sensitivity and specificity.



Table 1

The ACR/EULAR criteria for the classification of SSc
















































Category Subitems Weight
Skin a Skin thickening of the fingers of both hands extending proximal to the MCPs b 9
Puffy fingers 2
Whole finger, distal to MCP 4
Fingertip lesions a Digital tip ulcers 2
Pitting scars 3
Telangiectasia 2
Abnormal nail fold capillaries 2
PAH and/or interstitial lung disease 2
Raynaud’s Phenomenon (RP) 3
Scleroderma-related antibodies (any of anticentromere, anti–topoisomerase-I [anti–Scl-70], anti–RNA polymerase-3) 3
Total score:

These criteria are applicable to any patient considered for inclusion in an SSc study. Features can have been present at any time (ever).

Exclusion criteria: (1) these criteria are not applicable to patients having an SSc-like disorder better explaining their manifestations, such as nephrogenic sclerosing fibrosis, scleredema diabeticorum, scleromyxedema, erythromyalgia, porphyria, lichen sclerosis, graft-versus-host disease, and diabetic cheiroarthropathy; (2) patients with skin thickening that has never occurred on the fingers are also not classified as having SSc.

Adapted from van den Hoogen F, Khanna D, Fransen J, et al. 2013 classification criteria for systemic sclerosis: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Ann Rheum Dis 2013;72:1750; and van den Hoogen F, Khanna D, Fransen J, et al. 2013 classification criteria for systemic sclerosis: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Arthritis Rheum 2013;65:2741; with permission.

a Only the highest score from each category is used and the sum is totaled; a score of 9 or more classifies a patient as having SSc.


b Sufficient criterion: skin thickening of all fingers of both hands and extending proximal to the MCPs is sufficient for SSc classification (ie, it is scored as 9 and is sufficient for classification).



Using the American College of Rheumatology/European League Against Rheumatism Systemic Sclerosis Classification Criteria


The new ACR/EULAR SSc classification criteria are not to be applied if the patient has never had sclerodactyly but has fibrotic skin involvement elsewhere, thereby excluding eosinophilic fasciitis, generalized morphea, and other diseases that could be in the differential diagnosis of SSc. SSc mimickers were collected prospectively using cases and controls at multiple SSc sites, and from databases. Validation for the final ACR/EULAR SSc classification criteria was from the prospective cases and controls, including patients in whom the criteria would not be applied, so common sense and clinical judgment are needed to use the criteria. Expert opinion was used to reduce the number of potential items to be tested.


Methodological Advances


State-of-the-art methodology was used throughout the development of the 2013 SSc criteria establishing a comprehensive platform for classification criteria development in the modern era. Historical classification criteria in the rheumatic diseases have been threatened by a bias of circularity of reasoning in which investigators who developed classification criteria also contributed patient data in which the criteria were tested. Through each phase of SSc criteria development, caution was taken to ensure that separate groups of investigators and centers were involved in data-driven and expert-based methods. This approach also resulted in broad geographic representation of experts and multiracial representation of patients throughout North America and Europe. Candidate criteria were generated and reduced using advanced expert-based methods, including separate Delphi exercises in Europe and North America and nominal group technique. Multicriteria decision analysis facilitated further item reduction, weighting, and threshold identification. Through these methods, years of expertise, knowledge, and experience were used to inform the criteria development in a manner that was least susceptible to bias. Bayesian methods were used to evaluate the ability of each proposed criterion to distinguish cases from controls, and the psychometric properties of face and construct validity of each criterion was evaluated.


This methodological approach reflects increased collaboration between clinical experts and clinical epidemiologists; a balanced use of expert-based and data-driven methods, and implementation of bias reduction strategies and measurement science. This approach is being used to inform the development of classification criteria in other rheumatic diseases, including gout, myositis, and systemic lupus erythematosus (SLE).


Sensitivity and specificity were tested for the 2013 ACR/EULAR SSc classification criteria on serially collected patients from several outpatient clinics in Europe and North America. Approximately half the cases were early SSc so the performance of the new criteria in both established and early disease could be studied. The patients and controls (patients diagnosed with a disease that could mimic SSc) were from clinics that have expertise in SSc.


It was thought by experts in the 2013 criteria (using Delphi and other exercises) that some patients with mixed connective tissue disease (MCTD) could have SSc (overlaps were allowed) and some patients with current undifferentiated connective tissue could have several criteria for SSc (and perhaps be classified with SSc depending on the manifestations). A patient with MCTD with many features of SSc who scores 9 or more points in the 2013 ACR/EULAR SSc classification criteria would be classified as SSc and would presumably have a similar prognosis (when adjusting for the activity and severity of each SSc item) to other similar patients with SSc who do not have the other features of MCTD such as SLE, inflammatory myositis, or Sjögren syndrome features.


SSc disease may be a continuum (preclinical, mild, moderate, severe) with an arbitrary cutoff (above which the disease may be classified and below which the disease is not classified) and patients also may take time to meet criteria (undifferentiated connective tissue disease evolving into SSc). There were trade-offs between sensitivity and specificity at the 9 points needed for SSc classification and some experts classify patients with fewer criteria and occasionally do not classify a patient with SSc despite having the required 9 or more points. There are trade-offs between having a simple useful classification scheme versus a more comprehensive one. Important SSc features were removed, such as scleroderma renal crisis (SRC) and calcinosis, because they were too rare (former) or did not add greater statistical value than the other included items (former and latter) or did not differentiate from other diseases in the differential diagnosis of SSc. Both the 1980 and the 2013 SSc classification systems allowed SSc sine (without) skin involvement (in the former meeting 2 of 3 minor criteria) and by scoring items other than skin involvement in the recent SSc classification. The absolute criterion of skin involvement of the fingers and continuing proximal to the MCPs occurs in both schema. Examples of how to score cases are given later in this article.

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Sep 28, 2017 | Posted by in RHEUMATOLOGY | Comments Off on New Classification Criteria for Systemic Sclerosis (Scleroderma)

Full access? Get Clinical Tree

Get Clinical Tree app for offline access