Objective. To examine the relationship between disease activity and fatigue over time in early rheumatoid arthritis (RA). Methods. Data were from patients with early RA (duration of symptoms ≤12 months) enrolled in the Canadian Early Arthritis Cohort (CATCH). Patients rated the level of their fatigue over the past week using an 11-point numerical rating scale for up to 5 years of follow-up. Fatigue severity was classified as low (≤2), moderate (>2 but <5), or high (≥5). Fatigue severity ratings in patients who achieved a low disease activity state (Disease Activity Score in 28 joints [DAS28] ≤3.2) were compared to those in patients who did not achieve a low disease activity state within 3 months of cohort entry. Results. Of 1,864 patients included, 88% met RA criteria, and 72% were women. The mean ± SD baseline DAS28 was 4.9 ± 1.5. Nineteen percent of the patients reported moderate baseline fatigue, and 59% reported severe baseline fatigue. Fatigue was correlated with pain and patient global disease activity ratings (r = 0.56-0.67, P < 0.001), and was weakly correlated with the DAS28, tender joint count, swollen joint count, physician global assessment of disease activity, erythrocyte sedimentation rate, and C-reactive protein level. Patients who reported low fatigue by 3 months had significantly lower fatigue throughout follow-up compared to those who had moderate or high fatigue at 3 months (P < 0.001). Patients who achieved a DAS28 ≤3.2 within 3 months had significantly lower fatigue ratings than those with a DAS28 >3.2 (mean ± SD fatigue severity score 2.7 ± 2.6 versus 4.6 ± 3.0; P < 0.001), with improvements in fatigue that persisted through 5 years of follow-up. Maximal improvements in fatigue lagged behind remission by 6 months. Conclusion. Fatigue is common in early RA, and improvements may occur after remission. Early treatment response within 3 months after treatment initiation was associated with short-term and long-term improvements in fatigue over time.