Abstract
BACKGROUND: Immune checkpoint inhibitor (ICI) therapy is increasingly used to treat non-small cell lung cancer (NSCLC). However, little attention has been given to the comparative analysis of adverse events (AEs) associated with different ICIs. METHODS: Disproportionality analysis and Bayesian confidence propagation neural network (BCPNN) were utilized to identify pharmacovigilance signals from the FDA Adverse Event Reporting System (FAERS). We compared the sex distribution of patients, risk of suffering more severe adverse reactions, and risk of hospitalization associated with different ICIs, using pairwise matrices that displayed odds ratio (OR) and their 95% confidence interval (CI). And we also compared the outcomes of reactions by using ordinal logistic regression. RESULTS: We analyzed 13,580 reports of AEs associated with five ICIs, namely, durvalumab, pembrolizumab, ipilimumab, atezolizumab, and nivolumab from January 2013 to October 2022. Significant differences were observed in sex distribution of patients, risk of suffering more severe adverse reactions, risk of hospitalization, and the outcomes of reactions. In terms of respiratory AEs, pembrolizumab exhibited a higher risk compared to durvalumab (OR = 2.48, 95% CI: 1.72-3.59), atezolizumab (OR = 1.84, 95% CI: 1.07-3.16), and nivolumab (OR = 4.21, 95% CI: 1.72-10.28), while ipilimumab exhibited a higher risk compared to durvalumab (OR = 2.76, 95% CI: 1.14-6.65) and nivolumab (OR = 4.67, 95% CI: 1.14-15.51). In terms of endocrine and metabolic AEs, durvalumab (OR = 7.80, 95% CI: 1.33-45.90) and nivolumab (OR = 5.20, 95% CI: 1.17-23.03) exhibited a higher risk compared to ipilimumab. CONCLUSION: Each ICI has distinctive features of pharmacovigilance signals. It is essential to acknowledge the AEs associated with the relevant system when clinicians administer ICIs.