Background Patients (pts) with rheumatoid arthritis (RA) often have inflammation-related anemia of chronic disease, partially due to IL-6 and its induction of hepcidin. High levels of the inflammatory marker C-reactive protein (CRP) are largely induced by IL-6. Treatment with JAK inhibitors has been associated with both rise and fall in hemoglobin (Hb) levels, possibly due to inflammation control and inhibition of JAK2 (which is involved in erythropoiesis), respectively 1,2.
Objectives To examine the association of changes in Hb upon treatment with ABT-494 with baseline (BL) CRP levels.
Methods This post hoc analysis used data from two phase 2b randomized controlled trials (RCT) of ABT-494: in RA pts with inadequate responses or intolerance to TNF inhibitors (TNF-IR, BALANCE-1)3, and in pts with inadequate responses to methotrexate (MTX-IR, BALANCE-2)4. The analyses included pts receiving ABT-494 at 6 or 12 mg. Pts were subgrouped by quartiles of BL CRP. For each RCT, and for the pooled data from both, changes from BL in Hb (g/dL) were determined after 12 weeks (wk) of treatment with ABT-494, for the overall population and for pts who achieved ACR20 or DAS28-CRP ≤3.2 responses at Wk 12. For ACR20 and DAS28-CRP ≤3.2, non-responder imputation was used. For changes in Hb, data are as observed. Significance was determined by ANCOVA.
Results In pooled data from 110 and 100 pts from BALANCE-1 and -2 respectively, higher BL CRP was associated with smaller mean decreases in Hb at Wk 12 (p=.074). In the MTX-IR trial, pts in the highest CRP quartile had a mean increase from BL in Hb (+0.28 g/dL) vs pts in the lower quartiles who had mean decreases in Hb (p<.01). In both RCTs, ACR20 or DAS28-CRP ≤3.2 responders had smaller decreases in Hb vs non-responders at Wk 12. Among DAS28-CRP ≤3.2 responders, pts in the higher CRP quartiles had significantly smaller mean decreases in Hb vs pts in the lower quartiles (p<.01 for MTX-IR and p<.05 for TNF-IR pts). Responders in the highest BL CRP quartile also had a mean increase in Hb (+0.31 and + 0.54 g/dL for MTX-IR ACR20 and DAS28-CRP ≤3.2 responders respectively, + 0.15 g/dL for TNF-IR DAS28-CRP ≤3.2 responders) (Fig. 1A–D). While similar trends between BL CRP levels and changes in Hb at Wk 12 were observed among ACR20 or DAS28-CRP ≤3.2 non-responders, the differences in Hb levels between the quartiles were not significant. Despite the small mean changes, the mean Hb values remained within the normal range for females (lower limit of normal 11.5 g/dL)
Conclusions Upon treatment with ABT-494, pts with the highest CRP (a potential surrogate for IL-6) at BL had a mean increase in Hb compared to decreases in those with lower CRP. This effect was more apparent in pts with clinical responses. This suggests that effective treatment of RA-associated inflammation with ABT-494 may counterbalance the small Hb reduction associated with JAK inhibition.
Schulze-Koops et al 2017, Rheumatology;56:46.
Genovese et al 2016, NEJM; 374:13.
Kremer et al 2016, Arthritis & Rheum;68:2867.
Genovese et al 2016, Arthritis & Rheum;68:2857.
Acknowledgements AbbVie: Study sponsor, involved in study design, data collection, analysis, and interpretation, and in publication writing, review, and final approval. Medical writing: Naina Barretto, of AbbVie.
Disclosure of Interest V. Strand Consultant for: AbbVie, Afferent, Amgen, Biogen Idec, Bioventus, BMS, Carbylan, Celgene, Celltrion, CORRONA, Crescendo, Genentech/Roche, GSK, Hospira, Iroko, Janssen, Lilly, Merck, Novartis, Pfizer, Regeneron, Sanofi, SKK, Takeda, UCB, Vertex, M. Genovese Grant/research support from: AbbVie, Lilly, Astellas, Vertex, Pfizer, Galapagos, Gilead, Consultant for: AbbVie, Lilly, Astellas, Vertex, Pfizer, Galapagos, Gilead, J. Kremer Grant/research support from: AbbVie, Lilly, Novartis, Pfizer, MedImmune, Sanofi, and Regeneron, Consultant for: AbbVie, Lilly, Novartis, Pfizer, MedImmune, Sanofi, and Regeneron, Employee of: CORONA, M. Schiff Consultant for: AbbVie, Speakers bureau: AbbVie, Y. Li Employee of: AbbVie, J. Sokolove Employee of: AbbVie