• Arthritis Clinical Trials

Laure Gossec, MD, PhD, Paris Descartes University, Paris, France, presented results from the French Early Arthritis (ESPOIR) cohort study. Investigators analyzed patients who were in remission to determine if a patient's global status assessment and fatigue rating during the first year of early arthritis play a significant role in predicting structural progression over 3 years.

ESPOIR is an observational study of patients with early arthritis who had no prior disease-modifying antirheumatic drug (DMARD) therapy. Early arthritis was defined as having at least 2 swollen joints for <6 months. The outcome was change in total Sharp-van der Heijde score (SHS) from baseline to 3 years, adjusted for baseline radiographic score. Predictive variables included definitions of remission at 6 and 12 months, swollen and tender joints, C-reactive protein (CRP), global assessment, and fatigue rating (means of 6- and 12-month values). Remission definitions that were used were the ACR/EULAR Boolean remission (tender and swollen joint counts ≤1, CRP ≤1 mg/dL, and patient global ≤1/10), no-patient-reported outcome (PRO) near-remission (remission for all criteria except patient global), and fatigue remission (≤1/10 on a visual analog scale). Kappa agreement statistics were used to compare the definitions of remission. Multiple linear regression was used for prediction by remission definitions. Multiple linear regression and stepwise selection were used for prediction of radiographic score by remission components. Analyses were restricted to patients with all relevant data, with no imputation of missing data.

Remission data were available for 776 patients, and complete data were available for 520 patients. Among the patients with complete data, after 3 years, DAS28 decreased from 5.2±1.3 to 2.9±1.4, HAQ score decreased from 1.0±0.7 to 0.5±0.6, global assessment (0 to 10) decreased from 6.0±2.5 to 2.9±2.6, fatigue rating (0 to 10) decreased from 4.8±2.8 to 3.4±2.0, and SHS radiographic total score increased from 5.4±7.7 to 13.6±14.7. At 3 years, 57% of the patients were receiving methotrexate, and 16% were receiving biologic therapies.

Of the 776 patients, 7.4% achieved ACR/EULAR remission, 18.7% achieved no-PRO near remission, and 3.1% achieved fatigue remission (ie, with a fatigue score lower than 1/10). Of the 520 patients, 6.7% achieved ACR/EULAR remission, 18.7% achieved no-PRO near remission, and 3.1% achieved fatigue remission. Agreement between ACR/EULAR and the other remission definitions was moderate: ACR/EULAR versus no-PRO near remission – kappa, 0.48 (95% CI, 0.37 to 0.58); ACR/EULAR versus fatigue remission – kappa, 0.41 (95% CI, 0.23 to 0.58).

In the comparison of the remission models for the prediction of radiographic score, only swollen joint count and CRP were predictive of radiographic score. The PROs were not significant. Additional analysis of global cutoff in patients in no-PRO near remission (n=97) demonstrated no correlation between patient global and radiographic progression (Spearman correlation 0.025; p=0.575).

This analysis had several limitations. The comparison of models was not straightforward. There was a potential lack of power because of the low number of patients in remission. Fatigue remission is not a feasible outcome with a cutoff of 1/10.

No-PRO near remission was more frequent than ACR/EULAR Boolean remission in patients with early arthritis (18.7% vs 6.7%). Fatigue remission was rare (3.1%). Swollen joint count and acute-phase reactants were strong drivers of radiographic progression. Patients' global assessments had limited additional predictive value for radiographic progression. Further research is warranted.

• © 2012 MD Conference Express®