Diagnostic performance and predictive value of serum markers for the diagnosis of rheumatoid arthritis

Background

Rheumatoid factor (RF) is the classical serum marker for rheumatoid arthritis (RA). Anti-citrullinated peptide antibodies (ACPA) have a higher sensitivity and specificity (spec) for RA.

Objective

Our aim was to evaluate the diagnostic and predictive value for RA of RF and antibodies to pepA and pepB (two synthetic substrates for ACPA detection) in a setting relevant to routine clinical practice.

Methods

In this prospective multicentre study, samples were collected at academic and nonacademic centres from 1003 consecutive patients presenting for diagnostic work-up when the clinician included RA in the differential diagnosis. RF was detected by latex fixation. A research INNO-LIA™ RA (Innogenetics, Belgium) was used to detect anti-pepA and anti-pepB antibodies. Diagnoses were made by the clinician using ACR criteria after 1 year follow-up. ROC curve analysis was used to evaluate the diagnostic performance of the tests.

Results

The following diagnoses were made: definite RA (n = 144), non-RA (n = 629), undifferentiated (n = 156), and lost to follow up (n = 74). The first two groups were used to determine sensitivity, specificity, and positive predictive value (PPV). ROC curve analysis (Fig. 1) showed a higher area under the curve for RF than for anti-pepA and anti-pepB antibodies (0.839 versus 0.784 and 0.788, respectively), but in the high specificity region anti-pepA and anti-pepB antibodies performed better than RF (Table 1).

Figure 1

ROC curves.

Table 1 Diagnostic performance of serum markers using different cut offs

Conclusion

When high specificity is required, anti-pepA and anti-pepB antibodies have a markedly higher sensitivity than RF. The highest PPV are found when ACPA are very high.