Background/Purpose: Biologic disease modifying antirheumatic drugs (b-DMARDs) have revolutionized the management of Rheumatoid Arthritis (RA). Despite this success, a high percentage of the patients undergoing biologic treatment respond insufficiently. There is thus an urgent need to identify effective biomarkers to guide treatment selection. The aim of this study was to identify several specific reliable clinical and molecular predictors of the response of RA patients to TNF-α inhibitors (TNFi) and to B-cell depletion by anti-CD20 antibody (Rituximab, RTX).

Methods: In a prospective RA cohort study, we collected serum from RA patients with moderate or high disease activity prior to treatment with TNFi or RTX and analyzed baseline levels of 27 proteins that constitute a multibiomarker test of inflammatory profile of these samples using a multiplex immunoassay. The patients’ response was determined 24 weeks after starting bDMARDs treatment, according to the EULAR response criteria (good vs. moderate/no). We compared the inflammatory molecules between the response and non-response patient groups and analyzed their discriminative ability. Logistic prediction models were created to assess the added value of potential inflammatory predictors.

Results: Among 91 total RA patients, 45 of 66 patients in the TNFi group and 20 of 25 patients in the RTX group responded to the biologic treatment. As previously reported both, high DAS28 and CRP levels at baseline were predictive of response to TNFi. High rheumatoid factor titers and DAS28 at baseline were predictive of response to RTX. Instead, smoking habit and hyperlipidemia at baseline were predictors of a worse response to any of these b-DMARDs. Of the molecules analyzed by the multiplex assay, we identified seven molecules as potential predictors of TNFi response [fibroblast growth factor basic (FGFb), interleukin (IL)-10, IL-13, IL-15, IL-2, IL-8, and monocyte chemotactic protein -1 (MCP-1)] mainly involving molecules related to leukocyte’s activation and recruitment to inflammatory tissues.  On the other hand, three highly expressed cytokines/chemokines in RA serum were identified as predictors of RTX response [IL-9, interferon-inducible protein 10 (IP10) and macrophage inflammatory protein 1 beta (MIP1b)]. Receiver operating characteristic analyses for those multiple biomarkers revealed areas under the curve (AUC) ranging from 0.78 to 0.94, with a sensitivity of 100% and specificity varying from 80 to 90% for TNFi, and AUC from 0,76 to 0,97 with a sensitivity of 96% and specificity varying from 79 to 88 % for RTX. Combinations of these 7 and 3 biomarkers, respectively, significantly increased the accuracy of prediction.

Conclusion: By analysis of serum inflammatory profile, we identified specific and distinctive serum biomarkers that, in coordination with known clinical and serological profiles, have a high ability to predict the response of RA patients to TNFi or RTX treatments. Funded by PI-0285-2017, ISCIII, PI18/00837 and RIER RD16/0012/0015 co-funded with FEDER

« Back to 2019 ACR/ARP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/biomarkers-identified-by-serum-inflammatory-profile-analysis-to-predict-biologic-treatment-response-in-rheumatoid-arthritis-patients/