Misdiagnosis of Rheumatoid Arthritis in a Long-Term Cohort of Early Arthritis Based on the ACR-1987 Classification Criteria

Objective: Correct diagnosis of early rheumatoid arthritis (RA) is essential for optimal treatment choices. No pathognomonic test is available, and diagnosis is based on classification criteria, which can result in misdiagnosis. Here, we examined the differences between actual and misdiagnosed RA cases in a long-term cohort of patients included based on the ACR-1987 classification criteria. Methods: Patients in the BARFOT (Better Anti-Rheumatic PharmacOTherapy) cohort (n=2543) with at least four follow-up visits during the initial 5 years from enrolment were assessed, and a change in diagnosis was reported by the treating rheumatologist. The groups were analysed with respect to the individual classification criteria, antibodies to citrullinated proteins (ACPA), disease activity (DAS28) and radiographic changes from inclusion up to 2 years. Results: Forty-five patients (1.8%) were misdiagnosed (RA-change group). When compared to those in the RA-change group, the patients who kept their diagnosis (RA-keep) were more often RF positive (64% vs 21%, p<0.001) or ACPA positive (59% vs 8%, p<0.001). They were also more likely to fulfil more than four ACR-1987 criteria (64% vs 33%, p<0.001) and to have radiographic changes at inclusion (RA-keep 27% vs RA-change 12%, p=0.04). The groups had a similar evolution of DAS28 and its components as well as of radiological joint destruction. Conclusion: Diagnosis of RA according to the ACR-1987 criteria had a high precision in this long-term cohort. A diagnosis of RA should be re-evaluated in patients who do not fulfil more than four ACR-1987 criteria especially in patients negative for RF.