The recent article by Ravelli et al, which describes more homogeneous subsets of juvenile idiopathic arthritis (JIA), is very important, and the association between a positive result for antinuclear antibodies (ANAs) and certain clinical characteristics in JIA is indisputable (1). However, there are several arguments against the notion that ANA-positive patients should be grouped as a separate category within the classification of JIA.
The method used to detect ANAs is of critical importance. In their study, Ravelli et al detected ANA positivity using indirect immunofluorescence (IIF) testing, with rat liver or HEp-2 cells as substrate. Testing for ANA using an enzyme-linked immunosorbent assay (ELISA) is a standardized operator-independent method used increasingly in laboratories, because it is cost effective and has better antigenic specificity in systemic lupus erythematosus and other rheumatic diseases in adults. Detection of ANAs through use of ELISA has replaced the use of IIF in many laboratories, even though ELISA-detected ANA positivity has no clinical value in JIA (2). In a previous study, we showed that ANAs were detected by ELISA in only 4 of 100 children with JIA and were not associated with uveitis or with positive IIF-detected ANAs (2).
Rates of ANA positivity detected by IIF testing vary considerably between different cohorts (3, 4). The traditional IIF test is operator-dependent and has been criticized for the lack of both reproducibility and a uniformly accepted cutoff titer (5, 6). In the study by Ravelli and colleagues, 11% of patients did not have a conclusive ANA status as defined by the criteria that were set by the authors (1). In our study of 100 Norwegian children with JIA, the overall proportion of patients tested by IIF with a positive ANA result was 68% (2). Of these children, 61 had previously undergone IIF testing for ANA in an epidemiologic JIA study in the Nordic countries (7), and only 36% of the 61 patients were shown to be positive for ANA, as tested by IIF (unpublished observations). In the clinical setting, different laboratories yield different rates of positivity on ANA testing, and after more than 60 years of clinical use, the test is still difficult to standardize. In an accompanying editorial in the same issue of Arthritis and Rheumatism as the report by Ravelli and colleagues, Fritzler discusses the utility and future of the IIF test for ANA (8).
The antigenic specificity and target of ANAs in JIA have not been elucidated, and there is no evidence of a role for ANAs in the pathophysiology of JIA. The International League of Associations for Rheumatology (ILAR) criteria for JIA aim to be useful both for research purposes and in everyday clinical practice.
In conclusion, we find it doubtful that categorizing patients by ANA status can provide more homogeneous subsets than those provided by the present ILAR classification. Even if the ANA-positive patients in the Ravelli cohort share common clinical characteristics, a more well-defined and reproducible descriptor than ANA is probably needed for the purpose of research.