Background: Real-world evidence comparing long-term safety between PD-1 and PD-L1 inhibitors is limited, and the impact of competing mortality is unclear.
Methods: This retrospective cohort study utilized the TriNetX platform (2014-2024). Adults with advanced solid tumors initiating a PD-1 or PD-L1 inhibitor within 3 months of diagnosis were included. A 1:1 propensity score match created balanced cohorts. The primary outcome was the 1-year incidence of immune-related adverse events (irAEs). Standard Cox models and Fine-Gray competing risk models (accounting for death) were used to estimate hazard ratios (HRs) and subdistribution hazard ratios (sHRs), respectively.
Results: After matching, 22,672 patients (11,336 per group) were analyzed. In standard Cox analysis, PD-L1 inhibitors were associated with a lower risk of skin rash (HR 0.66, 95% CI 0.56-0.78), colitis (HR 0.79, 0.71-0.87), and nephritis (HR 0.71, 0.54-0.94) compared to PD-1 inhibitors. The competing risk analysis revealed that all-cause mortality was the overwhelmingly dominant outcome. Consequently, PD-L1 inhibitor use was associated with uniformly lower sHRs for every irAE analyzed (range: 0.85-0.88), with the sHR for mortality itself at 0.88 (0.84-0.92). This pattern, indicating the safety signal was heavily confounded by survival, was consistent across sex and age subgroups but more pronounced in males and older patients.
Conclusion: PD-L1 inhibitors are associated with a lower risk of specific irAEs. However, mortality is the dominant competing risk in advanced cancer, fundamentally shaping safety assessments. Real-world studies must account for competing mortality to avoid distorted conclusions.
Keywords: Competing risk analysis; Immune-related adverse events; PD-1 inhibitors; PD-L1 inhibitors; Real-World data; Safety.
© 2026. The Author(s).