Corticosteroid exposure, tapering, and survival in immune-related colitis: a 10-year real-world melanoma cohort

Background Immune-related diarrhoea and colitis (irC) is a common toxicity of immune checkpoint inhibitors (CPIs) in melanoma and frequently requires prolonged corticosteroid (CS) therapy. The impact of CS exposure on CPI efficacy remains uncertain, and evidence guiding tapering strategies and biological therapy is limited. The aim of this study was to evaluate the association between CS exposure (peak and cumulative) and survival in patients with melanoma having irC to identify predictors of prednisolone tapering failure and to describe treatment patterns in irC. Materials and methods This 10-year retrospective cohort study included patients with melanoma from the Danish Metastatic Melanoma Database who developed irC during CPI therapy. Clinical characteristics, CS exposure, and treatment patterns were obtained through systematic medical record review. Associations between CS exposure and survival were assessed using multivariable Cox proportional hazards regression. Predictors of tapering failure were evaluated using multivariable logistic regression. Results A total of 122 patients with irC requiring medical treatment were included. Higher cumulative CS exposure (>4000 mg) was associated with reduced overall survival (OS), whereas no significant association was observed at lower exposure levels. Peak CS dose was not independently associated with survival. Tapering failure occurred in 50% of patients. Anti-programmed cell death protein 1 monotherapy and early biological therapy were associated with a lower risk of tapering failure. Conclusions Higher cumulative CS exposure was associated with reduced OS in patients with melanoma having irC, although this may reflect underlying disease severity and treatment responsiveness rather than a direct causal effect. Associations among biological therapy, treatment regimen, and tapering outcomes require confirmation. Prospective studies are warranted to clarify causal relationships and optimise treatment strategies.