Antibiotics and immune-related toxicity during pembrolizumab therapy in NSCLC: Real-world analysis.

Abstract

e20637 Background: Antibiotic exposure around immune checkpoint inhibitor (ICI) therapy has consistently been associated with inferior clinical outcomes, likely mediated by disruption of the gut microbiome and downstream immune dysregulation. While microbiome composition has been linked to ICI-related colitis, the impact of antibiotics on organ-specific immune-related toxicities remains incompletely characterized, particularly for renal and endocrine adverse events in non–small cell lung cancer (NSCLC). Methods: We conducted a retrospective cohort study using a large U.S. electronic health record network. Adults with NSCLC treated with pembrolizumab were identified. The primary exposure cohort included patients receiving any systemic antibiotic within ±2 weeks of a pembrolizumab dose; controls had no antibiotic exposure within 1 month of pembrolizumab and low ECOG performance status. Propensity score matching (1:1) adjusted for demographics, comorbidities, smoking-related diagnoses, chemotherapy exposure, concomitant medications (including proton pump inhibitors and corticosteroids), and available laboratory values. Outcomes were assessed after matching using Kaplan–Meier analyses and Cox proportional hazards models, reported as hazard ratios (HR) with 95% confidence intervals (CI). Results: Among 5,690 antibiotic-exposed and 7,371 control patients, 2,381 well-balanced matched pairs were analyzed. Median follow-up was shorter in the antibiotic cohort (329 vs 511 days). Despite shorter follow-up, increased hazards of early-onset toxicities were observed, including acute kidney injury (AKI; HR 1.45, 95% CI 1.20–1.75; log-rank p < 0.001) and adrenal insufficiency (HR 2.67, 95% CI 1.78–4.01; log-rank p < 0.001). Modest associations were observed for unspecified colitis (HR 1.47, 95% CI 1.09–1.97; log-rank p = 0.010), while toxic colitis showed a borderline increase (HR 1.37, 95% CI 1.00–1.89; log-rank p = 0.053). No significant differences were observed for thyroiditis, dermatitis, gastrointestinal bleeding, acute liver failure, or composite ICI-linked diagnoses. Findings were similar using a broader ±1-month antibiotic exposure window in sensitivity analyses. Conclusions: In matched NSCLC patients receiving pembrolizumab, antibiotic exposure near treatment was associated with increased hazards of renal and endocrine immune-related toxicity, with weaker associations for colitis. These results extend prior microbiome–ICI literature beyond efficacy and gastrointestinal toxicity, supporting a microbiome-mediated, organ-specific toxicity hypothesis and highlighting the importance of antibiotic stewardship and enhanced renal and endocrine monitoring during ICI therapy.