Relative Impact of Pain and Disease Activity on Improvements in Fatigue: Results From 2 Baricitinib Phase 3 Clinical Trials

Abstract

Background/objective: Fatigue is common in patients with rheumatoid arthritis (RA). We assessed the relative impact of pain and disease activity on improvements in fatigue in 2 phase 3 baricitinib clinical trials.

Methods: RA-BEAM (NCT01710358) and RA-BEACON (NCT01721044) were randomized, double-blind, placebo-controlled studies in adults with moderate to severe RA. RA-BEAM assessed baricitinib + methotrexate (MTX) and adalimumab + MTX in patients with prior inadequate response/intolerance (IR) to MTX (MTX-IR). RA-BEACON assessed patients with IR to ≥1 biologic disease-modifying antirheumatic drug (bDMARD-IR). Measures included the Functional Assessment of Chronic Illness Therapy-Fatigue scale, Clinical Disease Activity Index (CDAI) for RA, and pain visual analog scale (VAS). Analyses were implemented separately for each study.

Results: Significant improvements were seen in disease activity and pain, which were greater with baricitinib versus adalimumab. A statistically significant improvement was seen in fatigue with both active treatments versus placebo. Moderate correlations were observed between improvements in disease activity and fatigue and between improvements in pain and fatigue in both MTX-IR and bDMARD-IR patients. Reductions in pain (≥50%) and remission or low disease activity (CDAI ≤10) had significant associations with fatigue improvement at week 24. In mediation analysis, improvements in fatigue attributable to CDAI and pain VAS in MTX-IR patients were 31% and 52%, respectively, for baricitinib, and 30% and 47%, respectively, for adalimumab. In bDMARD-IR patients, improvement in fatigue was attributed 48% to CDAI and 48% to pain VAS.

Conclusions: In both MTX-IR and bDMARD-IR patients, a large proportion of improvements in fatigue across treatment arms were accounted for by improvements in pain and disease activity.

FIGURE 1

Correlation between change from baseline in FACIT-F and (A) CDAI or (B) pain VAS in MTX-IR patients and (C) CDAI or (D) pain VAS in bDMARD-IR patients. R, Spearman correlation coefficient. The analysis was performed on pooled data from all treatment arms (including placebo) within each study.

FIGURE 2

Pairwise comparison between fatigue and categorical change at week 24 in MTX-IR and bDMARD-IR patients for (A) pain VAS and (B) CDAI. LS, least squares. The analysis was performed on pooled data from all treatment arms (including placebo) within each study. In A, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001 versus <30% pain improvement; + p ≤ 0.05, ++p ≤ 0.01, +++p ≤ 0.001 versus 30% to <50% pain improvement. In B, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001 versus CDAI >10; +p ≤ 0.05, ++p ≤ 0.01, +++p ≤ 0.001 versus 2.8 < CDAI ≤10.

FIGURE 3

Mediation analyses showing the proportion of direct effect and indirect effects mediated by pain and CDAI in (A) MTX-IR patients or (B) TNFi-IR patients. Direct effect = the treatment effect that cannot be accounted for by the indirect/mediation effect from improvement in CDAI and pain.