Time-dependent patterns of immune-related adverse events by obesity status following immune checkpoint inhibitor therapy: A real-world cohort analysis.

11191 Background: Immune checkpoint inhibitors (ICIs) substantially improve cancer outcomes but are frequently complicated by immune-related adverse events (irAEs). Obesity alters immune regulation and has been paradoxically associated with enhanced ICI efficacy; however, its influence on the timing and persistence of irAEs remains poorly defined. Prior studies have largely evaluated irAEs as binary outcomes, potentially obscuring time-dependent risk patterns. Methods: We conducted a retrospective cohort study using the TriNetX U.S. Collaborative Network. Adults with solid tumors initiating nivolumab, pembrolizumab, ipilimumab, or durvalumab were identified. Patients with pre-existing autoimmune disease, hematologic malignancy, organ transplantation, or prior irAEs were excluded. Patients were stratified by baseline body mass index (BMI ≥30 kg/m² vs 20–24.9 kg/m²). irAEs were identified using diagnosis-based groupings mapped to organ-specific toxicity categories. Survival analyses were performed using landmarked Cox proportional hazards models evaluating early (1–90 days), delayed (91–365 days), and late (366–1825 days) irAE risk. Adjusted hazard ratios (AHR) were calculated after adjustment for age, sex, cancer type, immunotherapy agent, smoking status, diabetes, chronic kidney disease, chronic pulmonary disease, liver disease, and cardiovascular comorbidity. Results: Among 29,564 eligible patients (obese: 13,958; normal BMI: 15,606), obesity was not associated with overall irAE risk across 5 years of follow-up (AHR 0.98, p=0.19). In landmark analyses, obesity was associated with lower early irAE risk (AHR 0.94, p=0.003), no significant difference during the delayed period, and a persistent increase in late irAE risk (AHR 1.11, p=0.003).Time-stratified associations were most pronounced for gastrointestinal, renal, endocrine, and dermatologic irAEs, while pulmonary and neurologic irAEs were not significantly increased after multivariable adjustment. Conclusions: Obesity demonstrated a time-dependent association with immune-related toxicity following ICI therapy, characterized by attenuated early risk and persistent late risk emerging beyond one year. These findings highlight the importance of time-stratified analytic approaches and suggest that absence of early irAEs may not confer long-term safety in obese patients receiving ICIs. Outcome Time Window Adjusted HR (95% CI) p value Any incident irAE 1–90 d 0.94 (0.90–0.98) 0.003† Any incident irAE 366–1825 d 1.11 (1.04–1.19) 0.003† GI irAE 1–1825 d 0.82 (0.77–0.87) <0.001† GI irAE 1–90 d 0.76 (0.68–0.86) <0.001† GI irAE 91–365 d 0.82 (0.74–0.90) <0.001† GI irAE 366–1825 d 0.84 (0.75–0.93) 0.001† Renal irAE 366–1825 d 1.14 (1.01–1.29) 0.031† Endocrine irAE 91–365 d 1.04 (0.99–1.09) 0.089 Endocrine irAE 366–1825 d 1.05 (1.00–1.11) 0.069 †p<0.05.