Background: To better communicate the results of randomized controlled trials (RCTs) of giant cell arteritis (GCA), we propose the use of the fragility index (FI), which is an intuitive measure defined as the minimum number of subjects whose status would have to change (e.g., from having the outcome to not) to render a statistically significant result nonsignificant, or vice-versa.
Methods: We conducted a systematic review and random-effects meta-analysis of RCTs of glucocorticoid (GC) sparing strategies for relapse-free maintenance in GCA, and used the FI to simplify the presentation of results.
Results: Ten RCTs (nine phase II and one phase III enrolling 645 subjects) were included. Tocilizumab, IV GC and methotrexate significantly improved the likelihood of being relapse free with relative risks and 95% confidence intervals of 3.54 (2.28, 5.51), 5.11 (1.39, 18.81) and 1.54 (1.02, 2.30); respectively. The median FI was 4.5 (range, 1-28), and was generally higher for negative RCTs (n = 6; median FI 4.5) than for positive RCTs (n = 4; median FI 3.5). The range of FI per treatment was (1-8) for methotrexate, (2-6) for anti-TNF agents, 4 for abatacept, 3 for IV GC pulses and (4-28) for tocilizumab.
Conclusion: Tocilizumab, IV GC and methotrexate improve the likelihood of being relapse-free in subjects with GCA. Assessment of GC sparing strategies in GCA has long depended on imprecise trials that would change significance if outcomes were reversed for a handful of subjects. FI may be used in rheumatology to simplify communication of statistical significance and overcome limitations of p-value.
Keywords: Fragility index; Giant cell arteritis; Meta-analysis; Randomized control trials; Tocilizumab.
Copyright © 2018 Elsevier Inc. All rights reserved.