Discontinuation of TNFi treatment among Rheumatoid Arthritis patients with a molecular signature of non-response to tumor necrosis factor-ɑ inhibitor therapies

Human Disease

Jeffrey Curtis, Joel Kremer, Dimitrios Pappas, Lixia Zhang, Erin Connolly-Strong, Johanna Withers, Viatcheslav Akmae and Alif Saleh


Background: A 23-feature blood-based molecular signature response classifier1 (MSRC) identifies rheumatoid arthritis (RA) patients who are unlikely to respond to tumor necrosis factor-ɑ inhibitor (TNFi) therapies. This study evaluated treatment outcomes over a 12-month period among patients prescribed TNFi therapies as a first targeted treatment and stratified according to MSRC results.

Methods: RA patients (N=175)1 from the CERTAIN study2 were followed 12-months following initiation of TNFi therapy. The MSRC evaluated patient gene expression data, anti-CCP and clinical metrics (sex, BMI, patient global assessment) prior to starting TNFi therapy to predict which patients were unlikely to respond to TNFi therapies. Prescription choices, response to therapies, and time on TNFi were assessed at 6, 9 and 12 months following TNFi treatment initiation. Response was measured using the ACR50 criteria.

Results: Relative to those without a molecular signature of non-response to TNFi therapies (n=93), a larger proportion of patients with high (n=26) and very high (n=47) signature of non-response discontinued their first TNFi therapy within the 12-month monitoring period (29.1% vs 42.3% and 59.6%, respectively). When a reason for discontinuation was provided, 63.6% reported issues with efficacy and 22.7% reported safety concerns. Predicted inadequate responders who did not reach ACR50 spent an average of 279 days (interquartile range 200-370) receiving TNFi therapy. Disease activity measures (DAS28-CRP, clinical disease activity index [CDAI], tender joint count and swollen joint counts) assessed at 6, 9 and 12 months following TNFi treatment initiation indicated less improvement from baseline among predicted inadequate responders compared to patients without a molecular signature of inadequate response. Among the 22 patients with more than one reported treatment cycle none (0/6) of the twelve PrismRA predicted inadequate responders achieved an ACR50 response to a second cycle TNFi therapy; however, 33% (2/6) achieved an ACR50 response to a non-TNFi treatment.

Conclusion: The molecular signature predictive of inadequate response to TNFi correlated with more discontinuation to TNFi therapy by 12 months, mostly due to lack of effectiveness; and despite this inactivity patients remained on TNFi treatment for 279 days on average.

View Full Paper