Tumor necrosis factor inhibitor (TNFi) persistence and reasons for discontinuation in a predominantly male cohort with axial spondyloarthritis

Delamo I. Bekele, Elizabeth Cheng, Andreas Reimold, Christian Geier, Kavya Ganuthula, Jessica A. Walsh, Daniel O. Clegg, Maureen Dubreuil, Prashant Kaushik, Bernard Ng, Elizabeth Chang, Ryan Duong, Jina Park, Gail S. Kerr

Research output: Contribution to journalArticlepeer-review

1 Scopus citations


Although tumor necrosis factor inhibitors (TNFi) have favorably altered the treatment landscape for patients with axial spondyloarthritis (axSpA), there is limited data regarding TNFi persistence and reasons for discontinuation. This is an observational time-to-event study utilizing data collected for a prospective multiple-disease registry of US Veterans with axSpA treated with TNFi therapies and recruited over a 10 year period. Clinical, serological, and comorbid parameters were collected. Corporate Data Warehouse Pharmacy files provided courses of the 5 TNFi agents, and response to treatment was documented. Individual TNFi persistence was established utilizing univariate and multivariate Cox proportional models, and reasons for discontinuation were obtained by physician chart review. Two-hundred and fifty-five axSpA patients received 731 TNFi courses. A majority of patients (84.3%) had TNFi persistence at 12 months; 63.5% and 47.1% at 24 and 36 months, respectively. Compared to adalimumab, infliximab demonstrated greater persistence, certolizumab the least. Age, smoking status, BMI, comorbidity burden, inflammatory markers and HLA-B27 did not predict TNFi persistence or discontinuation. Stroke and peripheral arterial disease increased the probability of TNFi discontinuation. Secondary non-response (SNR) was the most common reason for discontinuation (46% of all courses); non-adherence (6%) and clinical remission (2%) were uncommon. Pain score at enrollment, myocardial infarction, African American race and inflammatory bowel disease (IBD) predicted TNFi response. While initial persistence of TNFi treatment was high, a large proportion of the patients discontinued initial TNFi therapy by 3 years, primarily due to loss of efficacy. While further research identifying potential predictors of TNFi discontinuation in axSpA is warranted, access to alternate disease-modifying therapies is needed.

Original languageEnglish (US)
JournalRheumatology International
StateAccepted/In press - 2021
Externally publishedYes


  • Axial spondyloarthritis
  • Medication persistence
  • Real-world
  • TNF inhibitors

ASJC Scopus subject areas

  • Immunology and Allergy
  • Rheumatology
  • Immunology


Dive into the research topics of 'Tumor necrosis factor inhibitor (TNFi) persistence and reasons for discontinuation in a predominantly male cohort with axial spondyloarthritis'. Together they form a unique fingerprint.

Cite this