Background:Fatigue is common among patients with rheumatoid arthritis (RA) and has major impact on the burden of disease. There is little knowledge regarding the factors predicting the longitudinal development of chronic fatigue.Objectives:To identify baseline predictors for the development of chronic fatigue in patients with RA who initiate biological DMARD (bDMARD) treatment, and to compare disease courses across categories of fatigue for 12 months follow-up.Methods:Different trajectories of fatigue were calculated from a cohort of 209 established RA patients initiating bDMARDs. Fatigue was assessed by use of the fatigue Numeric Rating Scale (0-10) from the Rheumatoid Arthritis Impact of Disease (RAID) questionnaire. The patients were assessed at 0, 1, 2, 3, 6 and 12 months. We defined three groups: no fatigue (≤3 at all visits), improved fatigue (>3 at baseline but ≤3 at follow-up) and chronic fatigue (≥ 4 at all visits). All patients had clinical/subjective assessments (28 tender/swollen joint count, assessor’s/patient’s global VAS, RAID score, widespread pain, pain catastrophizing, the Hospital Anxiety and Depression Scale and inflammatory markers (ESR, CRP and calprotectin (a major granulocyte protein sensitive for inflammation in RA patients)). All patients were assessed by ultrasound (grey scale (GS) and power Doppler (PD)) of 36 joints and 4 tendons with semi-quantitative scoring (0-3). Differences between groups at baseline was assessed by bivariate analyses, and logistic regression models adjusted for age and gender were used to explore baseline predictors of chronic vs improved fatigue. Trajectories of different groups were plotted as estimated marginal means in figures, and differences between groups assessed by mixed models with maximum likelihood random effects, adjusted for age and sex.Results:Table 1 describes demographics and clinical factors of the three groups with significant differences shown in bold. Logistic regression with multivariate assessments found anti-CCP and low inflammation (calprotectin) to be predictors of chronic versus improved fatigue. Sleep disturbance was highly predictive of chronic fatigue. Figure 1 illustrates the trajectories for the three groups at all visits, showing the chronic fatigue group to have significantly higher DAS28, level of widespread pain, depression and sleep disturbance in contrast to no higher level of inflammation assessed by CRP and ultrasound PD.Table 1.No fatigueImproved fatigueChronicfatigueNo fatigue vs Improved fatigueImproved fatigue vs chronic fatigueNo fatigue vs. chronic fatigue482943pppAge, mean (SD) years51 (2)48 (2)54 (2)0.280.090.28Female gender (%)35 (73)24 (83)38 (88)0.400.500.09Higher Education (%)31 (65)23 (79)20 (47)0.170.010.17Anti-CCP positive (%)29 (60)20 (69)36 (84)0.720.010.002RF positive (%)27 (56)17 (59)30 (70)0.760.110.15Disease duration, mean (SD) years7 (1)8 (1)11 (1)0.810.110.03RA disease activityDAS28CRP3.2 (0.1)3.9 (0.2)4.7 (0.2)0.0030.004<0.001Swollen joints (28)5.7 (0.7)5.6 (1.0)6.2 (0.7)0.900.600.63CRP mg/L mean (SD)9.4 (2.4)15.6 (4.1)11.0 (2.6)0.020.020.58Calprotectin mg/L mean (SD)1.6 (0.3)2.0 (0.4)1.5 (0.2)0.440.200.92Sum score PD mean (SD)14.3 (1.8)13.8 (2.5)12.3 (1.9)0.850.620.43Sum score GS mean (SD)31.6 (2.8)29.3 (3.4)28.2 (2.7)0.610.810.39Psychosocial factorsRAID sleep (VAS 0-10)1.2 (0.3)4.3 (0.6)6.7(0.4)<0.001<0.001<0.001RAID fatigue (VAS 0-10)1.4 (0.2)5.6 (0.3)7.1 (0.3)<0.0010.003<0.001Widespread pain (0-25)4.3 (0.4)7.0 (0.8)8.6 (0.7)0.0010.16<0.001HADS anxiety1.5 (0.3)1.4 (0.6)3.4 (0.7)0.260.580.10HADS depression0.8 (0.2)0.9 (0.4)3.0 (0.8)0.980.360.05Pain Catastrophizing (0-6)1.0 (0.2)2.5 (0.3)2.9 (0.3)<0.0010.31<0.001Conclusion:Sleep disturbance is a modifiable factor presently found to predict chronic versus improved fatigue. Thus, attention should be given to RA patients with sleep problems to seek to avoid development of chronic fatigue. This issue should be explored in further studies.Disclosure of Interests:Hilde Berner Hammer Consultant of: Has received fees as consultant from Roche, AbbVie and Novartis., Speakers bureau: Has received fees for speaking from AbbVie, BMS, Pfizer, UCB, Roche, MSD and Novartis, Brigitte Michelsen Grant/research support from: Research support from Novartis, Consultant of: Consulting fees Novartis, Joe Sexton: None declared, Till Uhlig Consultant of: Lilly, Pfizer, Speakers bureau: Grünenthal, Novartis, Sella Aarrestad Provan Consultant of: Novartis