Evaluating survival following severe immune-related adverse events requiring hospitalization.

Abstract

2666 Background: As immune checkpoint inhibitors (CPI) are increasingly approved for the treatment of multiple cancer types, hospitalizations related to severe immune-related adverse events (irAE) will increase in tandem. Here, we identify patients hospitalized due to irAEs and describe survival outcomes across irAE, CPI, and cancer type. Methods: We identified patients exposed to CPIs who were hospitalized from 1/2012 to 12/2020 at our tertiary care hospital by computationally extracting data from the electronic health record. We then performed manual chart review to include only confirmed irAE-related hospitalizations. Survival outcomes were analyzed using Kaplan-Meyer survival curves with log-rank tests. Results: Of 3137 patients treated with CPIs, 117 were hospitalized for irAEs (cumulative incidence 4%). Of these, 36% had melanoma, 15% had lung cancer, 9% had renal cell carcinoma (RCC), and the rest were distributed across other cancers. The average number of irAE-related hospitalizations per patient was 1.25 (ranging from 1-3). Among 153 irAEs requiring hospitalization, 39% were gastrointestinal (GI) (including hepatitis), 12% endocrine, 12% pulmonary, 12% neurologic, 12% musculoskeletal, 7% cardiovascular, 6% dermatologic, and the rest affected other organs. Across all patients hospitalized for irAEs, median survival following hospitalization was 980 days. Patients with GI and endocrine irAEs had longer median survival (1474 days and median not reached [NR], respectively), while patients with pulmonary irAEs had shorter median survival (64 days) [p=0.002]. Patients with melanoma and RCC had longer median survival (2796 days and NR, respectively), while patients with lung cancer had shorter median survival (165 days) [p=<0.001]. Survival was not significantly different across CPI type [p=0.20]. Conclusions: Although hospitalization for severe irAE was rare, these findings suggest that among these patients, survival differs by irAE and cancer type. This real world data can contribute to clinical models that predict the risk of hospitalization and the risk of death due to severe irAEs, which may inform patient counseling and treatment decision-making. [Table: see text]