Article Text


Classifying childhood arthritis
  1. Department of Rheumatology, University of Birmingham, Birmingham B15 2TT

    Statistics from

    The year 1997 marks the 100th anniversary of the publication of Sir George Frederic Still’s landmark paper “On a form of chronic joint disease in children”.1 In recognising childhood arthritis as distinct from adult arthritis, Still fired the first shots in the war of words about its nomenclature and classification which continues to this day. It is now well recognised that childhood arthritis covers a heterogeneous group of diseases, many of which (but not all) have important differences from adult arthritides. Unfortunately, the names of the diseases, and their classification criteria, remain problematic. Still warned against the pitfalls of such debates in the preface to the fourth edition of his paediatric textbook; “… we must be careful lest we mistake words for things, and think that new nomenclature means new discovery, and that difference of terms is equivalent to distinction of kind”.2

    The aim of this leader is to discuss the classification of childhood arthritis. Over recent times this has been the focus of several papers,3 4 a couple of lively debates,5 and many heated private arguments! It has been difficult to reach agreement on the need for a new classification, let alone arrive at a consensus on the classification itself. However, most paediatric rheumatologists now recognise that the disparities between the European designation “JCA” and the North American designation “JRA” need to be resolved to facilitate international communication and research efforts in the aetiology and treatment of chronic arthritis in children. It is my personal view that the latter remains the prime purpose for reclassifying these diseases—to facilitate research by studying homogeneous disease groupings, rather than to provide a system of nomenclature covering every single form of arthritis in children.

    The arguments concerning the current systems of classification have been well rehearsed,3 and will not be discussed extensively here. Instead, consideration will be given to a proposal for the development of classification criteria for idiopathic arthritides of childhood published just over a year ago by the ILAR/WHO task force for classification criteria in paediatric rheumatology.6The classification refers to seven diseases characterised by idiopathic arthritis beginning before the 16th birthday (table). It is based on the clinical experience of paediatric rheumatologists from each of the four regional leagues of the International League Against Rheumatism. As such, it represents the first attempt to reach an international consensus on this subject.7

    The proposed system has several advantages over current classification systems. Firstly, and perhaps most importantly, it attempts to group clinically distinguishable diseases as distinct entities, rather than put them all under an umbrella term. In fact, each of the seven diseases is supposed to be mutually exclusive. Any patient who has clinical features, apart from arthritis, of more than one of the seven diseases is unclassifiable in this system, an example being the child with psoriatic arthritis who has circulating rheumatoid factor. This reflects the fundamental bias of the system towards research rather than everyday clinical use, as undoubtedly many patients will be unclassifiable. Despite its exclusivity, the system has a degree of flexibility not present in current classifications. For each disease, this takes the form of descriptors which allow further refinement of each disease group, including the age of the child at onset of the arthritis, the pattern of the arthritis, its natural history, and associated laboratory tests. Additionally, the disease “extended oligoarthritis” covers an important group of patients in whom the pattern of arthritis changes from an initial oligoarthritis to a polyarthritis in outcome. The system recognises that many extra-articular clinical features are useful for classification; with particular emphasis on enthesitis and psoriasis. Finally, the proposed system specifies that only a six week period of persistent arthritis is required before the diagnosis can be made, eliminating another JCA/JRA discrepancy.

    It must be emphasised that the proposed classification is not finalised, validated, universally acclaimed, or ready for clinical use! Even to colleagues not trained in paediatric rheumatology, it has many shortcomings.8 One of the most difficult problems is that it is based on inherently subjective criteria. It relies on the examiner’s observation of clinical features, including the number of joints involved and the presence of extra-articular features such as enthesitis, psoriasis, “systemic” rash, “generalised” lymphadenopathy, organ enlargement, sacroiliac joint “tenderness”, and “inflammatory” spinal pain. Considerable interobserver variability in joint counts has been demonstrated, even between very experienced paediatric rheumatologists,9 10 and it is likely that similar variability in other clinical features could be documented. Unfortunately there are no objective laboratory criteria on which to base the diagnosis of arthritis in childhood. The biological significance of using the 16th birthday to discriminate childhood from adulthood is questionable,11 as is the use of a seemingly arbitrary time point (six months after the disease began) at which to apply the classification criteria.12 There are several inconsistencies in the criteria and their definitions which may promote heterogeneity within the disease groups and are potentially confusing. These include the definitions of symmetry, dactylitis, positive family history of psoriasis and spondyloarthropathy, and positivity for antinuclear antibody and rheumatoid factor.8 11 On a more fundamental note, it is clear that any arthritic disease for which an aetiology is elucidated (and is therefore no longer “idiopathic”) will automatically be excluded from this classification, and thus as a long term solution to classification it is going to be self defeating!

    The task force recommended that the proposed classification be validated and compared with existing classifications before being used in a clinical setting. At present, the results of the paediatric rheumatology component of the 12th International Histocompatibility Workshop are being analysed to assess whether the proposed criteria result in patient groups of greater genetic homogeneity than are delineated by the existing classifications. It is only by long term follow up studies, however, that the true validity of the classification will become clear.13 As G F Still wrote “… let us be cautious lest we multiply distinctions; then deem that our puny boundaries are things that we perceive, and not that we have made.”2

    Perhaps an accurate classification cannot be formulated without understanding the aetiology and pathogenesis of the chronic childhood arthritides. It is of great concern, however, that we might never understand the mechanisms of these diseases unless we investigate relatively homogeneous disease groups. We must not allow another 100 years to pass before solving the mysteries of childhood arthritis and it is to be hoped that a flexible, biologically relevant classification will assist that process.


    Table 1

    Proposed classification criteria for the idiopathic arthritides of childhood (abbreviated from reference6)

    View Abstract

    Request permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.