Prediction models for rheumatoid arthritis during diagnostic workup: evaluation of combinations of rheumatoid factor, anti-citrullinated protein/peptide antibodies and the HLA shared epitope

Bert Vander Cruyssen ^1*, Ilse EA Hoffman ², Isabelle Peene ², Ann Union ², Herman Mielants ², Lydie Meheus ² and Filip De Keyser ²

¹ UZ Gent, Belgium
² Ghent University Hospital, Belgium

^* To whom correspondence should be addressed. E-mail: bert.vandercruyssen{at}ugent.be.

Accepted 26 June 2006

Abstract

Objectives: A prediction model for a disease may depend on the studied population. In the present analysis, our objective was to calculate the probabilities for rheumatoid arthritis (RA) in a consecutive cohort of patients during diagnostic work- up. Therefore, we fitted different logistic regression models evaluating the value of HLA-shared epitope (SE) determination and testing for rheumatoid factor (RF) and anti-citrullinated protein/peptide antibodies (ACPA).

Methods: The study included 1003 consecutive patients, presenting a new diagnostic problem for which RA was included in the differential diagnosis. All patients were tested for ACPA, RF and HLA-SE.

Results: After 1 year, diagnoses were established: 153 patients had definite RA and 629 patients had RA excluded. RF, used as a continuous marker is useful to evaluate the probability for RA. Combined RF and SE testing, or combined ACPA and RF testing, provided additional predictive information. The redundancy of SE testing in a model that includes ACPA testing can be explained by the high association between ACPA and SE both in RA and non-RA patients. The value of RF testing increased if patients presented with at least one swollen joint at baseline.

Conclusion: In the present study, we calculated valid probabilities for RA during routine diagnostic work-up and showed that the potential additional value of SE testing disappears when ACPA testing is available. Combined RF and ACPA testing is useful, especially when RF is considered as a continuous parameter reflecting an increasing probability for RA at higher RF titers. The value of (continuous) RF testing increases when the a priori chance is higher.

Keywords: HLA shared epitope, anti-citrullinated protein/peptide antibodies, prediction model, rheumatoid arthritis, rheumatoid factor

CiteULike    Complore    Connotea    Del.icio.us    Digg    Reddit    Technorati    What's this?

This article has been cited by other articles:

• EMAD, Y., RAGAB, Y., SHAARAWY, A., ABOU-ZEID, A., SAAD, A., FAWZY, M., JOKHDAR, H., RASKER, J. J. (2009). Can Magnetic Resonance Imaging Differentiate Undifferentiated Arthritis Based on Knee Imaging?. The Journal of Rheumatology 36: 1963-1970 [Abstract] [Full Text]  
• Vander Cruyssen, B., Nuytinck, L., Boullart, L., Elewaut, D., Waegeman, W., Van Thielen, M., De Meester, E., Lebeer, K., Rossau, R., De Keyser, F. (2007). Polymorphisms in the ficolin 1 gene (FCN1) are associated with susceptibility to the development of rheumatoid arthritis. Rheumatology (Oxford) 46: 1792-1795 [Abstract] [Full Text]