Article Text

Download PDFPDF

Poor predictive value of antinucleosome and antineutrophil cytoplasmic antibodies in a 270 inception cohort of patients with early naked arthritis of less than one year's duration
  1. J-M Berthelot1,
  2. A Saraux2,
  3. M Audrain3,
  4. P Le Goff2,
  5. M Hamidou4,
  6. J-Y Muller3,
  7. P Youinou5
  1. 1Rheumatology Unit, Nantes University Hospital, 44093, CHU, Nantes, France
  2. 2Rheumatology Unit, Brest University Hospital, 29629, CHU, Brest, France
  3. 3Laboratory of Immunology, Nantes University Hospital, 44093, CHU, Nantes, France
  4. 4Internal Medicine Unit, Nantes University Hospital, 44093, CHU, Nantes, France
  5. 5Laboratory of Immunology, Brest University Hospital, 29629, CHU, Brest, France
  1. Correspondence to:
    Dr J-M Berthelot, Rheumatology Unit, CHU Nantes, Hô, 44093, Nantes-Cedex 01, France;

Statistics from

In most studies examining the outcome of early arthritis, lupus and vasculitides are rare events, occurring in only 1–3% of cases.1 However, they should be diagnosed as quickly as possible. Thus, it might be worth testing patients with early arthritis for the presence of anticytoplasmic neutrophil antibodies (ANCA).2 Indeed, a strong positivity for ANCA might suggest that vasculitides, including Wegener's disease and microscopic polyangiitis, were probably present before the onset of the clinical features typical of these disorders.3,4 Similarly, and in addition to looking for antinuclear antibodies (and/or anti-dsDNA antibodies), it has been claimed that testing for antinucleosome antibodies might be a valuable additional test for the early detection of lupus; this last subset of autoantibodies has shown good sensitivity (56%) and excellent specificity (97%) for longlasting systemic lupus erythematosus (SLE).5 However, these assumptions have not yet been supported by data from an inception cohort of patients with early unclassified arthritis tested by routine methods.

We tested for ANCA by indirect immunofluorescence (IIF) (1:20 dilution) and for antinucleosome-IgG by a commercial enzyme linked immunosorbent assay (ELISA) kit (BMD DNA-NUC-LISA) in the baseline sera of 270 patients with early onset arthritis without clinical signs suggestive of visceral disease. We then followed up these patients for a mean (SD) of 28.5 (12.1) months.

For the ANCA testing, although 23/270 (9%) baseline sera were positive, neither of the two patients later diagnosed as having vasculitides were positive by IIF-ANCA. Moreover, for both patients, even testing for anti-proteinase 3 (anti-PR3) and anti-myeloperoxidase (anti-MPO) by ELISA was negative at baseline.

For antinucleosome testing, at the 10 IU threshold suggested by the manufacturer, 109/270 (40%) sera were positive, including 51/114 (45%) rheumatoid arthritis (RA) and 20/34 (59%) spondyloarthropathy (SpA). However at the 20 IU threshold only 4% were still positive, including 2/33 (6%) unclassified arthritis, 5/105 (5%) RA, 2/52 (4%) SpA, 1/5 (20%) sicca syndrome, and only 1/5 (20%) patients later diagnosed as having SLE. Moreover for this single patient, the diagnosis was already firmly established, especially as the antinuclear antibody titre was 1/1000 and anti-dsDNA were positive. Hence the positive predictive value for SLE of antinucleosome testing was 1/11, and the sensitivity 1/5.

These low sensitivities do not support the working hypothesis that systematic screening for IIF-ANCA and antinucleosome antibodies is useful for the diagnosis of vasculitides in early arthritis and of SLE in patients with “naked” arthritis. Such results were somewhat expected, given that systemic vasculitis or SLE with arthritis as precursor are quite uncommon events. However, they confirm that determination of ANCA and antinucleosome antibodies should not be carried out in the absence of clinical extra-articular manifestations, especially as low ANCA titres have already been demonstrated in a large percentage of more benign conditions like early RA and early SpA.6–9 We would therefore quite agree with the conclusion of Merkel et al, that instead of systematic screening for ANCA by IIF, only those patients with features atypical for RA, SpA, or undifferentiated arthritis should be tested for ANCA, using an ELISA for anti-PR3 and anti-MPO together with IIF-ANCA.9

Likewise in our cohort, antinucleosome antibodies were positive at low values at baseline in several conditions other than SLE and in only 1/5 patients with SLE. There are few reports on antinucleosome antibodies and SLE, and thus definite conclusions cannot be reached about the overall additional value of this test to support the diagnosis of SLE.10 However, our results strongly suggest that systematic testing for antinucleosome antibodies should not be a substitute for a careful search for all visceral signs suggestive of lupus in a patient presenting with seemingly “naked” arthritis.


Funding was provided by the Centre Hospitalier de Nantes and the Programme Hospitalier de Recherche Clinique 1995.


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.