Article Text


Diagnostic performance of three assays for anticitrullinated protein antibodies in the very early arthritis ‘ESPOIR’ cohort
  1. P Nicaise-Roland1,
  2. L Nogueira2,3,
  3. C Demattei4,
  4. S Grootenboer-Mignot1,
  5. L De Chaisemartin1,5,
  6. N Rincheval6,
  7. M Cornillet3,
  8. P Dieudé7,
  9. M Dougados8,
  10. A Cantagrel9,
  11. O Meyer7,
  12. G Serre2,3,
  13. Sylvie Chollet-Martin1,5
  1. 1Immunology Department, APHP, Bichat, Paris
  2. 2Laboratory of Cell Biology and Cytology, Institut Fédératif de Biologie, CHU de Toulouse, Toulouse, France
  3. 3Laboratory of Epidermis Differentiation and Rheumatoid Autoimmunity, CNRS-Université de Toulouse, France
  4. 4BESPIM, Nimes, France
  5. 5INSERM UMR996, Paris-Sud University, Paris, France
  6. 6Department of Biostatistics and of Clinical Research, IURC, Montpellier, France
  7. 7Rheumatology Department, APHP, Bichat, Paris, France
  8. 8Rheumatology Department, APHP, Cochin, Paris, France
  9. 9Rheumatology Department, Purpan Hospital, Toulouse, France


Background and objectives Anticitrullinated protein antibodies (ACPA) are recognised as the most specific markers for rheumatoid arthritis (RA). Their detection can be performed with various ELISAs using either the autoantigenic targets (human mutated vimentin: MCV and human fibrinogen: AhFibA) or cyclic peptides (CCP2). We investigated the diagnostic value of these ELISAs in a French cohort of undifferentiated arthritis of less than 6-month duration (ESPOIR cohort).

Material and methods 685 patients were followed until 2 years after inclusion. Antibodies were determined at baseline. The cut-off of each assay was previously determined in order to obtain a similar 98% diagnostic specificity. Diagnostic sensitivity was determined for patients classified as RA at 2 years, according to the 1987 American College of Rheumatology (ACR) criteria.

Results 592 (86.4 %) patients were classified as RA at 2 years. 47–48.5 % of RA patients were positive for one single test at baseline without any significant differences between the 3 assays, and 317 (53.5 %) were positive for at least one. Discrepant results were observed for eight patients positive only for anti-CCP2, 17 for anti-MCV antibodies, and 16 for AhFibA. 95 patients classified as RA at 2 years were not classified as RA at baseline: 30 of them (31.5%) were positive for at least one assay at baseline. If the positivity of at least one assay is included in the diagnostic criteria, 543 patients are classified as RA (79.3%) at inclusion, instead of 497 (72.5%) without it.

Diagnostic sensitivities of ACPA assays

Conclusions The performance of the three ACPA assays is similar but the tests do not totally overlap. In very early undifferentiated arthritis, the detection of the three ACPA subfamilies increased the sensitivity of 6% as compared with the assays of only 1 marker with the risk of a slight decrease in specificity. Moreover, the positivity of one ACPA assay 2 years before the diagnosis was predictive for RA in 30% of the 95 patients without a diagnosis at inclusion. When ACPA were included among the diagnostic criteria, 7% more patients were classified as RA at inclusion. These results are in perfect agreement with the new ACR/EULAR classification criteria for RA which now include ACPA.

Statistics from

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.