Clinical sensitivity of antibodies against cyclic citrulinated peptide in patients with rheumatoid arthritis
© BioMed Central Ltd 2001
Received: 15 January 2001
Published: 26 January 2001
The synthetic cyclic citrulinated peptide (CCP) is recognised by rheumatoid arthritis (RA) associated antifilaggrin antibodies, previously determined as antikeratin antibodies or perinuclear factor. Antibodies detected by ELISA using CCP as an antigen (anti-CCP) seem to be of prognostic value in patients with RA.
The objective of our study was to determine the clinical sensitivity of anti-CCP in patients with definite RA (according to ARA diagnostic criteria). RF results were considered when measured in the same serum as anti-CCP.
Sera from 97 RA patients (15 M, 82 F) were tested for anti-CCP in duplicates by Imunoscan RA ELISA (Euro-diagnostica). RF was tested in the same serum in 69/97 patients.
In all 4 assay runs, OD of all calibrators vs. their calculated values completely corresponded to the figure in the manufacturer's analysis certificate. When the units of the lowest calibrator D were calculated by the equation of log calibration curve according to the manufacturer's protocol, they were always about 20% (7-17U) above the defined value (50U). This inconsistency of the curve fitting led to a wrong validation in 26/97 (27%) of the RA samples whith OD below the OD of the calibrator D, but with calculated results above the defined cut-off at 50U. Therefore, we calculated results by the equtation of 3rd degree polynomal curve which fitted perfectly the measured OD values to the defined units.
Out of 97 RA patients, 56 were anti-CCP positive (58%). From 69 patients simultaneously tested for RF and anti-CCP, both tests were positive in 24/69 (35%) and both negative in 23/69 (33%). Anti-CCP were positive in 15/38 (40%) patients with negative RF. Despite lower anti-CCP positivity rate, 16% of samples in RF neg. group exhibited high anti-CCP values:
The clinical sensitivity of the anti-CCP test in our RA patients was 58%, which is lower than previously reported (68%). The discrepancy might partially be due to different calculation of the results. We believe that the introduction of polynomal standard curve could contribute to a more consistent validation of samples exhibiting OD bellow the lowest calibrator.
RF pos (n = 31)
RF neg (n = 38)