Lead Investigator: Michael Ward NIH
Title of Research Proposal: Predicting Treatment Response to Tumor Necrosis Factor Inhibitors in Patients with Ankylosing Spondylitis
Vivli Data Request: 3369
Funding Source: None
Potential Conflicts of Interest: I have no financial or commercial conflicts of interest in the proposed work.
Summary of the Proposed Research:
Axial spondyloarthritis (axial SpA) is a group of inflammatory spine conditions that affects 0.9-1.4% of general population, and ankylosing spondylitis (AS) is the prototypic disease. Tumor Necrosis Factor inhibitors (TNFi) have been widely used as the second line treatment for patients with active AS when patients have inadequate response to non-steroidal anti-inflammatory drugs (NSAIDs) or cannot tolerate NSAIDs. The treatment response to TNFi, however, is heterogenous. In our previous systematic review of randomized control trials of TNFi in patients with AS, about one-half of the participants (39.% to 58.9%) achieved the Assessment in SpondyloArthritis international Society 40% response (ASAS40) at week 24, an indicator of major response. In clinical practice, both patients and clinicians are interested to know, how likely an individual patient will achieve a major response or remission after initiating TNFi therapy. Although some studies have suggested baseline C-reactive protein (CRP) may predict therapeutic response to TNFi, other potential factors, such as disease duration, gender, HLA-B27 status, and the interactions between these features and with CRP, have not been examined. Therefore, limited information is available to assist clinicians and patients make treatment decisions based on expected efficacy. The goal of our study is to investigate the predictive value of baseline clinical features for achieving a major response, in order to identify subgroups of patients more likely to have a meaningful clinical response to TNFi. We plan to request and pool participant-level data from the active arms of randomized controlled trials of TNFi (adalimumab, etanercept, golimumab, infliximab) in active AS, and perform classification tree analysis to predict a major response to TNFi at week 12 and week 24. Classification tree analysis would allow us to assess several predictors and their interactions at the same time, and the result will be in the form of probability score that can be used directly in clinical practice. The results will help patients and physicians understand the likelihood of response prior to the start of TNFi and may help in deciding who should not be treated with TNFi, or how soon to declare primary non-response. In the future, a similar analysis of anti-IL17 medications may allow more personalized treatment.
Statistical Analysis Plan:
Inclusion criteria of RCTs: We plan to include RCTs on patients with active AS, with at least one intervention being TNFi. Specifically:
1) RCTs that evaluated the efficacy of TNFi in adult patients with active AS;
2) AS was defined as fulfilling modified New York criteria;
3) TNFi include original adalimumab, certolizumab, etanercept, golimumab, and infliximab; we will include FDA approved the dosing regimen.
4) For RCTs that has the option of early escape, we will only use the outcomes before the time of early escape.
Exclusion criteria: To ensure the homogeneity of the study population, we will exclude studies on patients with axial spondyloarthritis, and other related spondyloarthritis; we will also exclude open label studies.
Data request from other sources:
We will submit a data request to Vivli for RCTs that involves adalimumab and etanercept; and in addition, we have made a request to the Yale University Open Data Access (YODA) Project for RCTs that involves Golimumab and Infliximab. Certolizumab does not have any eligible studies. The full list of the requested studies from YODA is:
Study NCT ID Study Drug Data Request
NCT00265083 Golimumab YODA
NCT01248793 Golimumab YODA
NCT02186873 Golimumab YODA
NCT00202865 Infliximab YODA
NCT00207701 Infliximab YODA
We will use descriptive analysis to summarize the data. We will group the participants into major response vs. clinically important response vs. no response based on the change of ASDAS at week 24, and perform classification tree analysis using the variables listed in the Main Predictor/independent variable section to identify predictors for a major response at week 24. A similar analysis will be performed for response at week 12. We plan to perform a sensitivity analysis using ASAS40, ASAS20 as the outcome measures of major response and clinically important response, and for partial remission.
We will first examine whether missing data were missing completely at random. For missing outcomes data, we will use intent-to-treat analysis, and carry the last observation forward. We will use multiple imputation to address missing values, and compare these results with a complete case analysis.
Heterogeneity of the studies:
To reduce the heterogeneity of the RCT participants, we limited the study population to patients with active AS, and excluded similar conditions, such as axial spondyloarthritis. The inclusion criteria of the selected RCTs are similar. Because we are not comparing medications, heterogeneity of treatment effects is not an issue for this study. The classification tree method intrinsically identifies groups that differ with respect to the probability of the outcome. Therefore, we will be able to learn if structural factors in the studies (i.e. country of origin, specific TNFi) impact the results.
Study ID: NCT00195819
Sponsor ID: M03-606
Study ID: NCT00085644
Sponsor ID: M03-607
Study ID: NCT01114880
Sponsor ID: M11-991
Study Evaluating Etanercept Treatment of PatientsWith Ankylosing Spondylitis
Sponsor: Pfizer Inc.
Study ID: NCT00421915
Study Comparing Etanercept 50 mg Once Weekly to 25 mg Twice Weekly in Patients With Ankylosing Spondylitis
Sponsor: Pfizer Inc.
Study ID: NCT00418548
Study Evaluating Etanercept and Sulphasalazine in Ankylosing Spondylitis
Sponsor: Pfizer Inc.
Study ID: NCT00247962
- Wang R, Dasgupta A, Ward M. Predicting Major Treatment Response to Tumor Necrosis Factor Inhibitors in Patients with Ankylosing Spondylitis [abstract]. Arthritis Rheumatol. 2020; 72 (suppl 10).
- Wang R, Dasgupta A, Ward MM. Predicting Probability of Response to Tumor Necrosis Factor Inhibitors for Individual Patients With Ankylosing Spondylitis. JAMA Netw Open. 2022;5(3):e222312. doi: 10.1001/jamanetworkopen.2022.2312