Impact of diabetes on respiratory infection risk and hospitalization following immune checkpoint inhibitor initiation in non–small cell lung cancer: A propensity-matched retrospective cohort analysis.

11187 Background: Immune checkpoint inhibitors (ICIs) are a cornerstone of treatment for non-small cell lung cancer (NSCLC); however, patients with diabetes mellitus may have altered immune responses and higher susceptibility to infections. Data comparing infectious respiratory outcomes, healthcare utilization, and mortality between diabetic and non-diabetic NSCLC patients treated with ICIs is limited. We aimed to evaluate these outcomes in NSCLC patients receiving ICIs, stratified by diabetes status. Methods: We conducted a retrospective cohort study using the TriNetX Global Collaborative Network. Adults (≥18 years) diagnosed with NSCLC who initiated immune checkpoint inhibitor therapy (PD-1/PD-L1 or CTLA-4 inhibitors) within one year of diagnosis were included. Two cohorts were defined: patients with type 2 diabetes mellitus (DM) and patients without diabetes (non-DM). Propensity score matching (1:1) was performed based on demographics, comorbidities, cancer stage surrogates, smoking history, chemotherapy exposure, radiation therapy, and concurrent medications. Outcomes were assessed from 30 days to 3 years after ICI initiation and included bacterial pneumonia, viral pneumonia, severe sepsis with or without shock, ER visit/hospitalization, critical care needs, and overall survival. Hazard ratios (HRs) with 95% confidence intervals (CIs) were reported. Results: After propensity score matching, 905 patients were included in each cohort. Baseline characteristics were well balanced between groups. Compared with the non-DM cohort, patients with diabetes had a significantly higher risk of severe infectious complications, including bacterial pneumonia (HR 1.44, 95% CI 1.01–2.08), severe sepsis (HR 1.79, 95% CI 1.29–2.49). Rates of viral pneumonia did not significantly differ between groups (HR 0.79, 95% CI 0.45-1.38). Diabetic patients also demonstrated higher incidence of hospitalization (HR 1.17, 95% CI 1.05–1.30) and critical care utilization (HR 1.27, 95% CI 1.02–1.57). However, no statistically significant difference in overall survival between cohorts was detected (HR 1.12, 95% CI 0.98–1.29). Conclusions: Among NSCLC patients treated with immune checkpoint inhibitors, diabetes mellitus is associated with higher risk of respiratory infections and severe sepsis. Hospital visits and critical care utilization were more frequent in the diabetic cohort, without a corresponding difference in overall survival. These findings highlight the importance of heightened infection surveillance and proactive supportive care in diabetic patients undergoing immunotherapy. Prospective studies are warranted to better define underlying mechanisms and to optimize risk mitigation strategies in this vulnerable population.