Abstract
BACKGROUND: Among the biomarkers associated with immune-related adverse events (irAEs) induced by immune checkpoint inhibitors (ICIs) in cancer patients, interleukin-6 (IL-6) has emerged as a key predictive factor. However, it remains unclear whether high or low levels of IL-6 predispose patients to irAEs. Our objective was to evaluate the strength of the association between circulating IL-6 levels, measured in cancer patients before and after initiating ICIs, and the risk of irAEs. METHODS: In this multicenter, prospective pan-cancer cohort study, serum IL-6 levels were quantified immediately before the first (pre-ICI) and second (post-ICI) cycles of ICI therapy. To assess the association between IL-6 and irAEs, Fine and Gray competing risk regression models were fitted, considering irAEs as the main event and death as the competing event. The incremental predictive value of IL-6 levels was evaluated using Harrell's C-index. RESULTS: Overall, 224 patients were followed up for a median of 75.5 days after ICI initiation. The adjusted 1-year cumulative incidence of irAEs was 49.0% (95% confidence interval [95%CI], 41.9-55.6%). Multivariate regression models identified female sex (hazard ratio [HR], 1.81; 95%CI, 1.17-2.81; p=0.008), dual ICI therapy with nivolumab plus ipilimumab (HR, 1.86; 95%CI, 1.14-3.02; p=0.012) and post-ICI IL-6 levels (HR, 0.97; 95%CI, 0.94-1.00; p=0.049) as independent risk factors for irAEs. Using standardized post-ICI IL-6 levels, the effect was stronger, with an HR of 0.74 (95% CI, 0.55-1.00; p=0.049). Adding post-ICI IL-6 levels to a model containing established irAE risk factors improved the Harrell's C-index from 0.623 to 0.640. CONCLUSION: In cancer patients treated with ICIs, low rather than high post-ICI IL-6 levels, female sex and dual ICI therapy are independent risk factors for irAEs.