Improvements in Fatigue in 1536 Patients with Rheumatoid Arthritis and Correlation with Other Treatment Outcomes: A Post Hoc Analysis of Three Randomized Controlled Trials of Abatacept

INTRODUCTION: A post hoc analysis of three randomized controlled trials of abatacept in rheumatoid arthritis (RA) was conducted to explore the effect of abatacept on fatigue in RA and its correlation with other outcomes. METHODS: In this analysis of AGREE (early RA) and AIM and ATTAIN (established RA), changes in baseline fatigue (0–100 mm scale), pain, sleep (AIM and ATTAIN only) and Disease Activity Score (DAS) 28 (C-reactive protein; CRP) were calculated at days 29, 85, and 169. Agreement between improvements ≥minimum clinically important differences (MCID) in fatigue and other outcomes were evaluated using agreement statistics (kappa) in each study and at each time point. RESULTS: Of 1536 patients (mean disease duration: 6.2 months [AGREE], 8.5 years [AIM], 12.2 years [ATTAIN]), mean (SE) decreases in fatigue from baseline to day 169 with abatacept were 28.9 (1.7), 25.3 (1.2), and 21.9 (1.6) in AGREE, AIM, and ATTAIN, respectively, with corresponding decreases of 16.0, 13.7, and 13.4 at day 29. Most patients (67.8%; 624/920) reported improvements ≥MCID in fatigue with abatacept at day 169; 79.2% (671/847) and 57.8% (388/671) reported improvements ≥MCID in pain and sleep, respectively; 18.9% (158/836) were in DAS28 (CRP) remission. Agreement between improvement in fatigue and other outcomes was low (kappa range 0.30–0.51 [pain], 0.14–0.26 [sleep], and 0.02–0.12 [DAS28 (CRP) remission]). CONCLUSIONS: Abatacept resulted in rapid improvements in fatigue and pain in patients with RA. However, low agreement between improvements in these outcomes indicates that fatigue and other outcomes including pain and sleep may represent different domains of response. FUNDING: Bristol-Myers Squibb. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1007/s40744-017-0054-6) contains supplementary material, which is available to authorized users.