Comparative Risk of Cardiovascular Hospitalization With Immune Checkpoint Inhibitors Versus Chemotherapy.

PURPOSE: While immune checkpoint inhibitors (ICIs) are associated with rare, acute myocarditis, their comparative impact on chronic vascular events and heart failure versus conventional chemotherapy remains ill-defined. We compared the 2-year risk of hospitalization for major adverse cardiovascular events (MACEs) after ICI initiation versus chemotherapy and characterized risk heterogeneity across ICI regimens.

METHODS

We performed a global, propensity-matched cohort study using a federated electronic health record network (TriNetX). Patients initiating ICIs (2011-2024) for lung, skin, GI, or urinary malignancies were 1:1 matched to a historical cohort (1997-2011) treated with guideline-endorsed chemotherapy to benchmark toxicity against a standard-of-care control. The primary end point was hospitalization for MACE (heart failure, myocardial infarction, or ischemic stroke) within 24 months.

RESULTS

The final cohort included 30,720 matched patients (15,360 per group). At 2 years, ICI therapy was associated with a significantly higher risk of MACE hospitalization compared with chemotherapy (risk ratio [RR], 1.40 [95% CI, 1.26 to 1.56]; P < .001). This excess risk was consistent across heart failure (RR, 1.39), ischemic stroke (RR, 1.40), and myocardial infarction (RR, 1.33). Combination of ipilimumab plus nivolumab carried the highest early risk (0-6 months). However, durvalumab was associated with the highest 2-year cumulative incidence of MACE hospitalization (16.2%), significantly exceeding that of pembrolizumab and atezolizumab.

CONCLUSION

In this large real-world cohort, ICIs were associated with a substantial excess burden of severe cardiovascular events compared with traditional chemotherapy. The risk extends beyond acute myocarditis to include chronic vascular events, with distinct high-risk signals for combination therapy and durvalumab, underscoring the need for long-term cardiovascular surveillance.