Association between antibiotic exposure and the risk of checkpoint inhibitor pneumonitis in lung cancer patients.

[INTRODUCTION] Checkpoint inhibitor pneumonitis (CIP) is a potentially fatal immune-related adverse event in patients receiving immune checkpoint inhibitors (ICIs). Antibiotics (ATBs) usage has risen among such patients, yet its relationship with CIP remains uncertain. We investigated the association between ATB exposure and CIP development in lung cancer patients treated with programmed death-1/programmed death-ligand 1 (PD-1/PD-L1) inhibitors.

[METHODS] We retrospectively collected clinical data of lung cancer patients who received PD-1/PD-L1 inhibitors for the first time, according to predefined eligibility criteria. ATB exposure was defined as receipt of ATBs within 3 months before ICIs initiation, or within 3 months after ICI initiation but before CIP onset. Patients with ATB exposure were further stratified by timing relative to ICIs initiation and CIP onset (pre-ICIs only, post-ICIs only, or both). CIP was adjudicated by two oncologists and three thoracic radiologists on the basis of clinical symptoms, imaging findings, and laboratory/microbiological testing to exclude alternative etiologies, including infection. Associations between ATB exposure and CIP risk were evaluated using Fine-Gray competing-risk regression models.

[RESULTS] Among 998 patients, 233 (23.3%) had ATB exposure and 142 (14.2%) developed CIP, with a median onset of 82 days (range, 9 to 552). After multivariable adjustment, ATB exposure was associated with a higher risk of CIP (subdistribution hazard ratio [sHR] = 9.05, 95% CI 6.32 to 12.94, p < 0.001). The magnitude of association increased with CIP severity (mild to severe), with sHRs of 5.81, 11.40, and 12.46, respectively. Broad-spectrum ATBs (sHR = 7.27, 95% CI 5.19 to 10.17, p < 0.001) and oral administration (sHR = 7.06, 95% CI 5.07 to 9.84, p < 0.001) were associated with greater CIP risk than narrow-spectrum agents and intravenous administration, respectively. Patients exposed to ATBs exclusively after ICI initiation had the highest CIP risk (sHR = 11.19, 95% CI 6.92 to 18.09, p < 0.001). Prophylactic ATB use was associated with worse outcomes among patients with CIP (adjusted p = 0.014).

[CONCLUSION] In lung cancer patients treated with anti-PD-1/PD-L1 therapy, ATB exposure was associated with a substantially increased risk of CIP. ATB use, particularly broad-spectrum and orally administered regimens, should be carefully evaluated, and ICI-treated patients receiving ATBs warrant close monitoring. In patients with established CIP, routine prophylactic ATB therapy should be used judiciously given its association with worse outcomes.