Background:The patient global assessment (PGA) is the most widely used patient-reported measure in rheumatoid arthritis (RA), being a component of disease activity scores and definitions of remission. Studies have however shown that, at least in established RA, PGA is mostly associated with pain, functional limitation, psychological distress and comorbidities, rather than with objective measures of inflammation. As such, the inclusion of PGA as a driver of intensification of immunosuppressive therapy is been questioned. Determinants of PGA may however differ in the earliest stages of RA, when pain processing mechanisms and damage accrual have not occurred yet. Whether the association of PGA with disease activity in RA changes over time is at present undetermined.Objectives:To analyze the associations between PGA and disease activity measures in patients with RA across different phases of the disease.Methods:1.002 RA patients from two independent cohorts were included: 1) a prospective longitudinal cohort of early RA (<12 months of symptoms) (n=601) with an observation period of 24 months upon initiation of therapy with methotrexate; 2) a cross-sectional cohort of established RA (duration >5 years) (n=401) with inadequate response to methotrexate. Determinants of PGA were assessed by Pearson’s correlation coefficients and multivariable linear regression.Results:In early RA, median (IQR) symptom duration at inclusion was 15 (9-27) weeks, 71.5% of the patients were female, 49.3% were autoantibody-positive, and 30% had radiographic evidence of ≥1 erosion. The mean (SD) DAS28 was 4.94 (1.19), and the mean (SD) PGA 57.6 (26.6). The proportion of patients at least in low disease activity (DAS28 <3.2) after 6, 12 and 24 months of treatment was 49.9%, 57.4% and 67% respectively. The associations of PGA with pain and functional limitation were moderate to good (r >0.40) at all time points, and were independent of other co-variates (Table 1). Swollen joints (SJC28) were independently and directly associated with PGA at all time points until 12 months, whilst the association become indirect at 24 months (Table 1). In ACPA-positive patients, PGA at 6 and 12 months was weakly but still significantly associated with SJC28 even in low disease activity states (adj r 0.15 and 0.13, p<0.05). Patients with established RA had a mean [SD] DAS28 of 5.3 [1.2] and were ACPA-positive in 69.6% of the cases. Independent determinants of PGA were pain, HAQ and tender joints, whilst no associations with SJC28 and acute phase reactants were found neither in the overall cohort nor in ACPA-positive patients.Conclusion:In established RA, PGA appears mostly related to factors outside the core domains of disease activity. In contrast, in the early phases of the disease, PGA may more strictly collect information on inflammatory-related symptoms. Better understanding of the relationship between patient reported outcomes and disease activity in the various phases of RA may thus be needed before introducing definitive changes in the current definitions of disease activity.