Abstract

Background: Different methods exist to demonstrate anti-citrullinated protein/peptide antibodies (ACPA).

Aims: To evaluate discrepancy between four ACPA tests.

Patients and methods: Population 1 consisted of patients with a new diagnostic problem, including 86 patients with rheumatoid arthritis (RA) and 450 patients without RA. Population 2 consisted of 155 patients with RA who had long-standing disease. Population 3 consisted of 188 patients with psoriatic arthritis and in population 4 there were 192 patients with systemic lupus erythematosus. Populations 1 and 2 were tested with the anti-human fibrinogen antibody (AhfibA) test, anti-CCP2 from Eurodiagnostica (CCP2-euro), anti-CCP2 from Pharmacia (CCP2-phar) and anti-CCP3 test by Inova (CCP3). Samples were annotated as discrepant if positive in one and negative in at least one other test. Each discrepant sample was re-analysed in a different run. Populations 3 and 4 were analysed in the CCP2-euro and AhFibA test.

Results: In population 1, ACPA positivity was found in 17 of 450 (3.8%) patients without RA; 14 (82%) of these 17 samples were discrepant. In contrast, 61 of 86 (70.9%) patients with RA were ACPA positive of whom 18 of 61 (29.5%) were discrepant (70.9% vs. 29.5%, p<0.001). The discrepancies between tests could be partly attributed to borderline results, inter-assay discrepancy and inter-test variability. They were more prevalent in patients with systemic lupus erythematosus who were ACPA positive than in those with psoriatic arthritis who were ACPA positive.

Conclusions: Discrepancy between different ACPA tests was observed attributable to the occurrence of borderline results, inter-assay variability and mainly to inter-test variability. The lowest inter-test discrepancy is observed between tests that use the same substrate.