P09.20 Impact of Antibiotics During Immune Checkpoint Inhibitor (ICI) Therapy for Non-Small Cell Lung Cancer: A Real-World Analysis

Abstract

Conflicting evidence exists regarding the impact of antibiotic (ATB) use on effectiveness of ICIs among patients with cancer due to ATB effect on gut microbiota. In non-small cell lung cancer (NSCLC), these studies are mostly retrospective, single-institution series or unplanned subset analyses of clinical trials with no adjustments made for multiple potential confounders. There is significant heterogeneity across studies regarding ATB exposure, ICI line of therapy, types of ICI (PD/L-1 with/without CTLA-4 inhibitor), and outcomes reported (progression free [PFS], overall survival [OS]). A recent meta-analysis (Lurienne et al, 2020) reported that PFS was reduced by 1.2 months in patients exposed to ATB. Thus, this US claims-based analysis aimed to assess the real-world impact of ATB use on time to discontinuation of first-line (1L) ICIs, (surrogate of PFS) in patients with advanced NSCLC. Patients with ≥1 non-diagnostic hospital claim or 2 outpatient claims indicating NSCLC (plus 1 claim indicating metastasis) from January 1, 2015 to September 30, 2019, ≥1 claim for ICI therapy (limited to PD/L-1 agent in 1L), and ≥1 month pre and post earliest ICI claim were identified from Symphony Health longitudinal prescription and medical claims; patients with claims indicative of SCLC, secondary malignancies, and participation in a clinical trial were excluded. Patients were classified as having ATB use (anytime within 30 days pre and post ICI initiation) and no ATB (no claims for ATB in 30 days prior and during ICI therapy). Patient characteristics were summarized for each ATB group by frequencies for categorical variables and measures of centrality and spread for continuous variables; comparisons across groups were made by chi-square and Mann-Whitney tests. Median duration and 95% confidence intervals (CI) of ICI therapy was estimated by Kaplan-Meier estimator, adding 28 days to last claim for nivolumab or atezolizumab and 21 days for pembrolizumab. Among the 307 NSCLC patients included, 74 (24%) had ATB use, and 233 (76%) had no ATB use. Median age at diagnosis was similar across groups (65.5-66 years; P=0.79), while those with ATB use had a greater preponderance of females (54% vs 39%; P=0.02). Median follow-up from diagnosis was 13.7 months. Most common ATBs were β-lactams (45%), quinolones (39%), macrolides (23%), and tetracyclines (12%). ATBs were given a median of 7 days with 23% treated with >1 course for ATBs. Nivolumab (31% vs 35%), pembrolizumab (39% vs 31%), and atezolizumab (3% vs 8%) use were similar across those with and without ATB use, respectively (all P>0.05). Median durations of ICI therapy were similar among those with (6.9 months, 95% CI 5.4-9.9 months) and without ATB use (8.9 months, 95% CI 6.9-12.1 months; P=0.21). ATB use is common among NSCLC patients prior to and early during ICI therapy. ATB use did not have a statistically significant impact on time to ICI discontinuation, a measure of disease progression in a real-world US cohort of patients with NSCLC. However, the magnitude of decrement is similar to that reported in recent meta-analyses. Additional research is warranted to assess the true impact of ATB use with ICI in NSCLC.