Predictors of Response to Anti-TNF Therapy in Axial Spondyloarthritis

spondyloarthritis, axial spondyloarthritis, SpA, axSpA
Credit: Medical Images
Researchers sought to determine the modifiable and nonmodifiable risk factors affecting anti-TNF response in axSpA.

Cohort study data published in Rheumatology identified several predictors of nonresponse to anti-tumor necrosis factor-α (TNFα) in patients with axial spondyloarthritis (axSpA), including lower socioeconomic status (SES), fewer years in education, the presence of clinical comorbidities, and poorer mental health. 

Analyses were conducted as part of the British Society for Rheumatology Biologics Registers in Ankylosing Spondylitis (BSRBR-AS) study, a prospective cohort of patients with axSpA in Great Britain. Between 2012 and 2017, patients with axSpA were recruited from 83 secondary care centers across the country. Enrollees were naive to biologic therapy at baseline. Patients assigned to the biologic subcohort began anti-TNFα treatment at recruitment and were followed up at 3 months, 6 months, and 1 year, and annually thereafter. Patients in the nonbiologic cohort could later switch to the biologic cohort and initiate anti-TNFα treatment. 

At each study visit, patients underwent a clinical assessment and completed self-report questionnaires. The primary outcome was treatment response, defined as meeting any of the following criteria: achieving 20% and 40% improvement per Assessment of Spondyloarthritis International Society criteria (ASAS20 and ASAS40, respectively), displaying a reduction of ≥1.1 in the Ankylosing Spondylitis Disease Activity Score (ASDAS), and moving from high or very high disease activity (ASDAS, ≥2.1) to moderate or inactive disease (ASDAS, <2.1). Forward stepwise logistic regression models were used to determine predictors of treatment response. 

The current analysis used data from 335 patients, among whom 69% were men. Median age at enrollment was 47 years (interquartile range [IQR], 36-56 years). Median time from TNFα initiation to first clinical assessment was 14 weeks (IQR, 12-17 weeks). ASAS20 and ASAS40 response criteria were achieved by 175 (52%) and 111 (33%) patients at first follow-up visit, respectively. 

In addition, 122 (47%) participants had a ≥1.1-point reduction in ASDAS, and 87 (35%) achieved moderate or inactive disease states. Lack of ASAS20 response was predicted by the absence of full-time employment, greater body mass index, and higher baseline anxiety. Failure to meet ASAS40 criteria was similarly predicted by employment, although not by body mass index or anxiety. 

Lower educational attainment was associated with lack of response per ASAS40 and ASDAS disease state criteria. Poorer mental health and greater number of clinical comorbidities were also associated with nonresponse to ASDAS criteria. In receiver operator characteristic curves, models that used ASDAS criteria demonstrated the highest discriminatory capacity for nonresponse. 

Related Articles

Independent of the axSpA criteria used, lower SES and educational attainment predicted poorer response to initial treatment with anti-TNFα. Although mental health and comorbidities may be appropriate targets for intervention, certain demographic factors are more difficult to tackle. Still, screening for these predictors of nonresponse may be useful to tailoring axSpA patient care.


Macfarlane GJ, Pathan E, Jones GT, Dean LE. Predicting response to anti-TNFα therapy among patients with axial spondyloarthritis (axSpA): results from BSRBR-AS [published online January 28, 2020]. Rheumatology (Oxford). doi:10.1093/rheumatology/kez657