Do the 2010 ACR/EULAR or ACR 1987 classification criteria predict erosive disease in early arthritis?

Background The new 2010 American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) classification criteria for rheumatoid arthritis (RA) aim at earlier diagnosis of RA compared to the 1987 ACR criteria. Objective To evaluate the ability of the 2010 ACR/EULAR and the 1987 ACR classification criteria to predict radiographic progression after 10 years of follow-up. Methods All early arthritis patients referred to Central Hospital in Jyvaskyl from 1997 to 1999 (cases with peripheral joint synovitis, other specific diseases excluded) were included in this 10-year follow-up study. Radiographs of hands and feet were analysed according to Larsen on a scale of 0-100. Results At 10 years, 58% of the patients had an erosive disease (defined as Larsen >= 2 in at least one joint). The discriminative power of the 2010 ACR/EULAR and the 1987 ACR criteria (erosive disease at 10 years) were comparable, with area under the curve 0.72 (95% CI 0.65 to 0.79) (2010 ACR/EULAR criteria) and 0.65 (95% CI 0.58 to 0.72) (1987 ACR criteria). The respective sensitivities and specificities were 0.87 and 0.70, and 0.44 and 0.47. At 10 years, median (IQR) Larsen score was 6 (0, 15) among patients who had fulfilled both sets of criteria, 2 (0, 8) in those who met the 2010 ACR/EULAR and did not meet the ACR 1987 criteria, 0 (0, 5) in those who met ACR 1987 criteria but did not meet 2010 ACR/EULAR criteria, and 0 (0, 2) among patients who did not fulfil either of the criteria. The percentage of patients with erosions was 69%, 64%, 32% and 26%, respectively. Conclusions The ability of the 2010 ACR/EULAR and 1987 ACR classification criteria to identify erosive disease in early arthritis is low. The discriminative power of the 2010 ACR/EULAR criteria of erosiveness in 10 years is slightly better than that of the 1987 ACR criteria.