Cardiovascular safety profile of CD20 × CD3 bispecific antibodies compared to immune-checkpoint inhibitors: A faers analysis, 2014–2025

Abstract

Abstract Background: Bispecific T-cell engagers (BTEs) have evolved from CD19 × CD3 constructs such as blinatumomab to newer CD20 × CD3 agents, including mosunetuzumab, glofitamab, and epcoritamab. While cytokine-release syndrome (CRS) is a well-recognized toxicity, the broader cardiovascular (CV) safety profile of BTEs, especially relative to immune-checkpoint inhibitors (ICIs), which carry a known myocarditis risk, remains poorly defined. We analyzed the U.S. FDA Adverse Event Reporting System (FAERS) to characterize CV adverse event signals associated with BTEs and benchmark them against ICIs from 2014 to 2025. Methods: FAERS data (Q1 2014 through Q1 2025) were queried using validated regular-expression algorithms to classify primary-suspect drugs as BTE-CD19, BTE-CD20, or ICIs. CV adverse events were categorized into six MedDRA-defined groups: myocarditis/pericarditis, arrhythmia, heart failure/cardiomyopathy, acute ischemia/arterial thrombosis, venous thromboembolism, and blood pressure disorders. Disproportionality was assessed using reporting odds ratios (RORs) versus the full FAERS background. Temporal trends were evaluated using Poisson regression. Time-to-onset (TTO) for myocarditis/pericarditis was compared using Kaplan–Meier analysis. A sensitivity analysis excluded “cytokine” prefererred terms to account for potential CRS miscoding. Results: Among ~14 million reports, we identified 8,863 BTE-CD19, 3,208 BTE-CD20, and 234,114 ICI primary suspect cases. Arrhythmia was the leading CV signal for CD20 BTEs (2.81 events/100 reports; ROR 2.21, 95% CI 1.78–2.73), with a risk ratio of 2.20 (1.79–2.72) versus ICIs. CD19 BTEs showed a weaker arrhythmia signal (ROR 1.36). Myocarditis/pericarditis was predominantly reported with ICIs (4,503 events; 1.92/100; ROR 15.8). Year-over-year Poisson modeling revealed declining arrhythmia reporting, most pronounced for CD20 BTEs (IRR 0.74, p = 0.004). TTO analysis showed median onset of myocarditis at 0 days (IQR 0–93) for CD20 BTEs, 11 days for CD19 BTEs, and 27 days for ICIs (log-rank p = 4×10⁻⁷). Sensitivity analysis excluding cytokine terms changed key RORs by <2%. Conclusions: CD20 × CD3 bispecifics are associated with a distinct arrythmia signal, more frequent and earlier than that observed with ICIs. However, overall arrhythmia reporting has declined over time, likely reflecting improved mitigation strategies such as step-up dosing and early monitoring. Nonetheless, the absolute incidence (2.8%) exceeds that seen with ICIs, supporting baseline ECG and rhythm surveillance during CD20 BTE initiation. In contrast, myocarditis remains primarily linked to ICIs, with later onset and a different pathophysiologic profile, underscoring the importance of troponin monitoring and early cardiology consultation. Prospective cardio-oncology registries are warranted to validate real-world risks, clarify mechanisms, and inform tailored monitoring as BTE use expands in practice.