Statistics from Altmetric.com
First, we allowed non-steroidal anti-inflammatory drugs (NSAIDs) in this trial, except for the 6 weeks (from 2 weeks prior to the initiation of trial drugs until 4 weeks) during which NSAIDs were allowed only for fever, in order to assess the status of body temperature appropriately.
Second, the per cent change in glucocorticoid dose were compared between the placebo and tocilizumab groups by means of analysis of covariance model with group as a factor and baseline as a covariate. Therefore, least squares means adjusted for baseline imbalance were presented in our manuscript, which are not consistent with simple summary statistics. We chose this model (not non-parametric one) because of its robustness in terms of the deviation from the normality assumption even with small sample size and prespecified this linear model in statistical analysis plan before the unblinding. We also confirmed that there were no outliers in our data to ensure the appropriateness of the analysis plan.
Third, the primary analysis population for efficacy evaluation in the double-blind phase (parts 1 and 2) was full analysis set, which consisted of patients who received at least one dose of the study medication. Therefore, all randomised patients, except for one patient in the placebo group whose underlying disease turned out to be lymphoma, not adult-onset Still’s disease, were included in efficacy analyses. This is consistent with the intention-to-treat principle. The criteria for escape during part 2 in our trial was non-achievement of American College of Rheumatology (ACR) 50 without fever under the initial dose of glucocorticoids while the endpoint at week 12 was ACR50 achievement.
Finally, as Bilgin et al pointed out, patients who received placebo had worse disease activity and took higher doses of glucocorticoids, and the efficacy outcomes in the placebo group were better than had been expected. This might be owing to careful patient enrollment or the effectiveness of slightly higher dose of predonisolone in the placebo group. However, this was the first trial using placebo in patients with adult-onset Still’s disease, and considering occasional fatal clinical courses of the disease, the investigators were compelled to be cautious.
We totally agree with Bilgin et al that establishing more sophisticated, validated tools or scoring systems is essential in trials and the management of adult-onset Still’s disease.
Handling editor Josef S Smolen
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent Not required.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.