ROME — The new ACR/EULAR criteria for diagnosing rheumatoid arthritis performed well, compared with two previously reported sets of criteria, in an early arthritis cohort of 736 people with early signs of synovitis or inflammatory tender joints, according to a report by Dr. Celina Alves.
The ACR/EULAR criteria “are the best choice, as they perform well in all patient subgroups, and they are easy to use in daily practice,” said Dr. Alves, a physician at Erasmus University Medical Center in Rotterdam. The new ACR/EULAR criteria “show particularly good performance in patients with persistent arthritis.”
Dr. Alves and her associates tested the performance of the criteria that a joint panel of the American College of Rheumatology and the European League Against Rheumatism released at a special session at the ACR's annual meeting last October.
The new criteria diagnose rheumatoid arthritis (RA) by assigning points in four clinical categories: number and size of affected joints, positive serology for rheumatoid factor or anti–cyclic citrullinated antibody, duration of symptoms lasting 6 weeks or longer, and an abnormal level of an acute-phase reactant (C-reactive protein or erythrocyte sedimentation rate). Patients who score 6 or more points out of a potential maximum of 10 are considered to have RA.
Publication of the ACR/EULAR criteria should occur later this year.
The researchers compared the performance of the ACR/EULAR criteria against two older sets of criteria: an algorithm published by Dr. Annette van der Helm-van Mil and associates (Arthritis Rheum. 2007;56:433-40), and an algorithm published by Dr. Henk Visser and associates (Arthritis Rheum. 2002;46:357-654).
Dr. Alves and her associates selected these two algorithms for comparison to the ACR/EULAR criteria after reviewing the rheumatology literature. They picked them because they “use routinely available variables that can be applied in daily practice by any rheumatologist,” she said in an interview.
To make the comparison, they used 736 people enrolled in the Rotterdam Early Arthritis Cohort (REACH). The group included 476 patients with inflammatory arthritis and 260 patients with two or more inflammatory tender joints but without apparent synovitis. The 476 patients with inflammatory arthritis included 282 with undifferentiated inflammatory arthritis and 194 with differentiated (classified) inflammatory arthritis.
In all 736 people studied, 21% were positive for rheumatoid factor, 20% were positive for anti–cyclic citrullinated protein, 5% had joint erosions, and 55% had persistent disease after 1 year.
To compare the three diagnostic algorithms, the researchers applied each of them to three separate groups from REACH. The first group included the 282 people with undifferentiated inflammatory arthritis. The second group included the 476 patients with either differentiated or undifferentiated arthritis. The third group included all 736 patients (the 476 from the second group, plus the remaining 260 people without inflammatory arthritis and at least two inflammatory tender joints).
The results showed that any of the models would be valid to use in the early stages of RA. “In patients with arthritis, all diagnostic models would perform well,” Dr. Alves said. “In patients with inflammatory complaints, including arthritis, the ACR/EULAR 2010 criteria and the van der Helm model performed best.”
Disclosures: Dr. Alves and her associates said they had no disclosures.
The ACR/EULAR criteria 'perform well in all patient subgroups, and they are easy to use in daily practice.'
Source DR. ALVES