Canakinumab's Effect Against Subsequent Gout Flares and High‐Sensitivity C‐Reactive Protein Levels: A Causal Mediation Analysis

Objective The objective of this study was to quantify the value of early suppression of high‐sensitivity C‐reactive protein (hsCRP) levels as a biomarker of the protective role of canakinumab against future gout flares. Methods We conducted a post hoc causal mediation analysis of the Canakinumab Anti‐Inflammatory Thrombosis Outcome Study for gout flares. The 3‐month change in the log hsCRP level was the mediator of interest. We used linear regression for the hsCRP level mediator and Cox or Weibull regression for gout‐flare outcomes, combining them in causal mediation analysis. We examined the cohort overall, as well as stratified by prevalent gout at baseline. Results We analyzed 9,221 patients without prevalent gout and 747 with prevalent gout. The Cox regression hazard ratio (HR) for a gout flare was 0.50 (95% confidence interval [95% CI] 0.37–0.68) comparing canakinumab with placebo, of which 6% was explained by the mediated effect through hsCRP level reduction in the first 3 months. In the prevalent‐gout subgroup, the HR was 0.58 (95% CI 0.36–0.95), of which 31% was explained by the mediated effect through hsCRP level reduction. The Weibull analysis gave a proportion‐mediated estimate of 47%. The indirect effect via hsCRP level reductions was unclear in the subgroup without prevalent gout. Conclusion The first 3‐month reduction in hsCRP level was not a good biomarker for canakinumab's protective effect on future gout flares in the overall cohort. Among patients with prevalent gout, there may be a potential role for early hsCRP level reduction as a biomarker for interleukin‐1β inhibitors' future gout‐flare benefit.