RESUMO
OBJECTIVE: To investigate the associations between the Patient Global Assessment (PGA) and measures of disease activity in patients with rheumatoid arthritis (RA) in relation to disease duration and autoantibody status. METHODS: 1412 patients from three independent cohorts were studied: a prospective cohort of 810 patients with early RA followed up for 24 months; a cross-sectional cohort of 210 patients with established RA in low disease activity; a cross-sectional cohort of 401 patients with established RA in moderate-to-high disease activity. Correlations of the PGA were analysed by Pearson's coefficients and multivariable linear regression at baseline and at months 6, 12 and 24 in the overall populations and after stratification for autoantibody subgroup and remission status (Boolean remission, PGA near remission and non-remission). RESULTS: In patients with early RA in non-remission, swollen joints correlated independently with the PGA; the correlation became progressively weaker but persisted at all time points in autoantibody-positive patients (adjusted r=0.30-0.12) but lost significance after month 12 in autoantibody-negative patients. Swollen joints independently correlated with the PGA also in near remission until month 12 (adjusted r=0.18-0.16) in autoantibody-positive patients. No independent correlations of inflammatory variables were instead found in patients with established RA irrespective of disease activity and autoantibody status. CONCLUSIONS: In the early phases of RA, particularly in autoantibody-positive patients, inflammatory variables directly correlate with the PGA across different disease activity states. The optimal cut-off values of the PGA capable of identifying absence of disease should be better explored in relation to disease duration and autoantibody status.