Class-level differences and within-class heterogeneity in heart failure reporting with immune checkpoint inhibitors: a WHO vigiaccess pharmacovigilance study

Background/Introduction

Immune checkpoint inhibitors (ICIs) have revolutionised cancer care, but immune-related cardiotoxicity remains poorly catalogued beyond fulminant myocarditis. Emerging case series describe a clinically silent, acute left ventricular decompensation phenotype that lacks inflammatory biomarkers, implying that heart failure (HF) can arise de novo from immune checkpoint modulation rather than as a downstream scar of myocarditis. Whether this signal is uniform across PD-1, PD-L1, and CTLA-4 blockade, or differs among individual agents, has not been systematically explored in global pharmacovigilance data. Clarifying these patterns is essential for risk-stratified surveillance strategies in contemporary cardio-oncology.

Purpose

To evaluate whether the risk of acute cardiac failure differs across immune checkpoint inhibitor classes (PD-1, PD-L1, CTLA-4) and among individual agents using global pharmacovigilance data.

Methods

Spontaneous adverse-event reports for seven ICIs were extracted from the WHO VigiAccess platform. "Cardiac failure acute" (MedDRA preferred term) was the primary outcome. Disproportionality was quantified using reporting odds ratios (RORs) with 95% confidence intervals; class-level comparisons used the same methodology. Sensitivity analyses employed the composite "cardiac failure" term.

Results

A total of 136 reports of acute cardiac failure were identified among 4,791 ICI-related cardiac adverse events.

Agent-level analysis: Pembrolizumab showed the highest signal (ROR 1.60, 95% CI 1.14–2.26); nivolumab was neutral (ROR 0.96, 0.67–1.38); cemiplimab, atezolizumab, durvalumab, and ipilimumab all had RORs ≤ 1. Avelumab had zero reports.

Class-level analysis: PD-1 inhibitors demonstrated a higher signal compared with CTLA-4 blockade (ROR 1.57, 1.05–2.35), whereas PD-L1 inhibitors did not (ROR 0.75, 0.44–1.28). The PD-1 versus PD-L1 comparison yielded an ROR of 2.10 (1.18–3.74). Sensitivity analyses confirmed class-level findings.

Conclusion(s)

Pharmacovigilance data identify PD-1 inhibition, particularly pembrolizumab, as carrying a disproportionate signal for acute cardiac failure compared with CTLA-4 blockade, while PD-L1 inhibitors appear neutral. Agent-specific heterogeneity exists within classes, supporting heart failure as an under-recognised, differential immune-related cardiotoxicity rather than a mere sequela of myocarditis. These findings warrant systematic baseline and on-treatment cardiac surveillance in patients receiving PD-1–targeted therapies.