The toxicity profile and temporal dynamics of dual PD-1/CTLA-4 immune checkpoint blockade: a real-world...

Combination therapy with PD-1 and CTLA-4 inhibitors improves survival in advanced cancers but is associated with heightened toxicity. Whether this represents additive toxicity from each drug or a distinct synergistic profile remains unclear. This study utilized the FDA Adverse Event Reporting System (FAERS) to systematically characterize the toxicity landscape of dual PD-1/CTLA-4 blockade.

We analyzed FAERS database (2004-2024) for nivolumab, ipilimumab, and their combination. Disproportionality analysis was performed using Reporting Odds Ratios (RORs) and the Ω shrinkage measure model to detect drug-drug interactions signals. To explore potential synergistic signals, we used two exploratory criteria: 1) Emergent signals: adverse events (AEs) with a significant reporting association only for the combination; 2) Supra-additive reporting signals: combination ROR at least 50% higher than the higher monotherapy ROR.

We identified 15, 252 combination therapy reports. Disproportionality analysis revealed 23 emergent preferred terms (PTs) unique to the combination, including immune-related AEs like cardiotoxicity (ROR = 1.89). Strong supra-additive reporting signals were observed for immune-mediated hepatitis (ROR = 205.80) and endocrine toxicity (ROR = 557.95). The Ω shrinkage measure model detected statistical interaction signals for cytokine release syndrome (Ω=2.35) and immune-mediated dermatitis (Ω=2.48). Among cases with evaluable onset dates, the majority of severe AEs were reported within the first 90 days after treatment initiation, with descriptive differences observed across cancer types. Integrated analysis revealed hepatobiliary disorders were the most frequently reported AEs (28.1% of 675 cases), followed by endocrine disorders (14.2%).

Our findings are consistent with the hypothesis that dual PD-1/CTLA-4 blockade may be associated with a toxicity profile involving potential supra-additive reporting associations. These signals include emergent safety events not detected with monotherapies, reporting associations in specific organ systems, and observed temporal patterns. These results are hypothesis-generating and await confirmation in independent studies.