The performance of anti-cyclic citrullinated peptide assays in diagnosing rheumatoid arthritis: a systematic review and meta-analysis.

OBJECTIVES We assessed the ability of anti-cyclic citrullinated peptide (CCP) tests to diagnose rheumatoid arthritis (RA), comparing the effect of manufacturer assay type, study design (single- and two-gated) and duration of disease (early vs. established). METHODS We searched seven databases for relevant diagnostic studies containing data on CCP tests in known or suspected RA patients. We used a bivariate model to produce summary estimates for test sensitivity, specificity, and positive and negative likelihood ratios. Summary Receiver Operating Characteristic (sROC) curves were derived to compare early versus established RA. RESULTS 83 studies were identified and included. For individual manufacturer tests there was considerable variation in both pooled sensitivity (range 67-83%) and specificity (range 90-96%) estimates. This heterogeneity was also observed when grouping studies into two-gated and single-gated designs. Study design and disease duration impacted on sensitivity, with single-gated study designs and early RA patients resulting in lower estimates than two-gated and established disease, respectively. CONCLUSIONS This review highlights the large number of CCP tests that are now commercially available and the considerable variation in their diagnostic performance. This variation, although partly influenced in this analysis by the study design (single-gated vs. two-gated), seems to have different levels of impact depending on the manufacturers. The Thermo Fisher Scientific EliA and Inova Diagnostics Quanta Lite (CCP2) tests showed the least between-study variation in sensitivity and specificity suggesting they have the most consistent diagnostic performance overall.