Loading ...

TOPLINE:

An analysis of randomized controlled trials (RCTs) for giant cell arteritis (GCA) and Takayasu arteritis revealed that most trials with significant and nonsignificant outcomes were fragile. The fragility measures did not differ based on the risk for bias, drug type, disease subtype, or publication timing.

METHODOLOGY:

• Researchers conducted a systematic literature review through April 2025 and identified 32 RCTs of GCA and Takayasu arteritis to assess the fragility of trials, defined as the minimum number of outcome events needed to change to reverse the conclusions of trials, providing a numerical estimate to evaluate confidence of results with categorical outcomes.
• They analyzed 18 trials (14 GCA trials involving 1252 patients and 4 Takayasu arteritis trials involving 145 patients). Data extraction included primary and key secondary outcomes, event counts in the intervention and control groups, and risk for bias analysis using Cochrane Risk of Bias (RoB) 2.
• The fragility index (FI) was calculated for trials with significant outcomes (as the number of events needed to be nonevents in the intervention group for a particular outcome to become nonsignificant). Other analyses included the reverse FI for trials with nonsignificant outcomes and the fragility quotient as the ratio of FI or reverse FI to total trial participants. Trials were labeled fragile if the FI was ≤ 3, the reverse FI was ≤ 5, or the fragility quotient was ≤ 0.1.
• Eight RCTs involving patients with GCA and one with Takayasu arteritis had a significant primary outcome, from which the FI was calculated. Nine RCTs involving patients with GCA and three with Takayasu arteritis had a nonsignificant primary outcome, from which the reverse FI was calculated. Moreover, three trials had some significant outcomes, and others did not.
• Subgroup analyses examined differences based on the Cochrane RoB 2 assessment, type of drug (biologic/targeted synthetic disease-modifying antirheumatic drugs vs other agents), subtype of large vessel vasculitis, and publication timing (before/after 2015).

TAKEAWAY:

• The FI for trials with significant outcomes ranged from 1 to 12, and the fragility quotient from 0.019 to 0.150. Five of nine trials with significant outcomes had an FI ≤ 3, and eight of nine had a fragility quotient ≤ 0.1, marking them as fragile.
• For trials with nonsignificant outcomes, the reverse FI ranged from 1 to 9, and the fragility quotient from 0.009 to 0.330. About 8 of 12 trials with nonsignificant outcomes had a reverse FI ≤ 5, and 6 of 12 had a fragility quotient ≤ 0.1, indicating fragility.
• About 4 of 12 trials with nonsignificant outcomes had reverse FI values less than the number of participants with data missing for the assessed outcome.
• No significant differences were found in the FI, reverse FI, or fragility quotient when comparing trials on the basis of risk for bias, drug type, subtype of large vessel vasculitis, or publication timing.

IN PRACTICE:

“The development of robust outcome measures for use in clinical trials of GCA or [Takayasu arteritis] might reduce the fragility of their results and impart greater confidence in their results. Such measures might include a composite assessment of disease activity combining information from clinical assessment, blood biomarkers, and imaging,” the authors of the study wrote.

“We wish to highlight the need to consider the routine reporting of fragility indices while presenting the results of RCTs to facilitate a fragility indices while presenting the results of RCTs to facilitate a clearer understanding of the findings of the trial and the implications of changes in a few events or missing outcome data on the eventual result of a trial,” they added.

SOURCE:

The study was led by Durga Prasanna Misra of Sanjay Gandhi Postgraduate Institute of Medical Sciences in Lucknow, India. It was published online on August 25, 2025, in Autoimmunity Reviews.

LIMITATIONS:

The study only included RCTs with categorical outcomes. The estimation of the continuous FI was not possible due to the lack of access to primary data. Arbitrary cutoffs for fragility indices were used, owing to the unavailability of actual cutoff values.

DISCLOSURES:

No specific grants were received for this research. One author disclosed participating in unbranded educational activities and clinical trials with Novartis, Amgen, and AstraZeneca; clinical trials with AbbVie; and advisory boards with GSK.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.