Incidence and Comparative Risk of Hematological Adverse Events Across Different Immune Checkpoint Inhibitors: A Network Meta-Analysis of Randomized Controlled Trials

Cancer immunotherapy with immune checkpoint inhibitors (ICIs) has transformed oncologic treatment, yet hematological adverse events (HAEs) remain incompletely characterized across different agents. We conducted a systematic review and network meta-analysis to compare the incidence and risk of hematological toxicities (all grades) among various ICIs. A comprehensive literature search was performed across PubMed, EMBASE, Cochrane Central Register of Controlled Trials, and Web of Science through January 2026. Phase II and III randomized controlled trials (RCTs) evaluating ICIs and reporting HAEs were included. Frequentist network meta-analysis was performed to estimate odds ratios and rankings using surface under the cumulative ranking curve values for anemia, neutropenia, and thrombocytopenia. Thirty-seven RCTs encompassing diverse malignancies were included in the network meta-analysis. For anemia, ipilimumab demonstrated the most favorable safety profile, followed by toripalimab and pembrolizumab, while serplulimab and nivolumab ranked lowest. Regarding neutropenia, ipilimumab again showed the best safety ranking, followed by avelumab and pembrolizumab, whereas toripalimab and serplulimab exhibited less favorable profiles. For thrombocytopenia, ipilimumab ranked highest, followed by sintilimab and nivolumab, while toripalimab showed the least favorable ranking. No significant global inconsistency was detected across networks. Publication bias was identified for neutropenia but not for anemia or thrombocytopenia. This network meta-analysis reveals substantial heterogeneity in hematological safety profiles across ICIs, with ipilimumab consistently demonstrating favorable rankings across all HAEs. These findings suggest potential differences in hematologic safety profiles across treatments. However, given the star-shaped network structure, reliance on indirect comparisons, and observed heterogeneity, the results should be interpreted cautiously. While not definitive, the analysis may still offer preliminary insights to inform treatment selection, particularly for patients with baseline hematological vulnerabilities or those receiving concurrent myelosuppressive therapies, and may help guide the development of risk-stratified monitoring strategies in clinical practice.