Background: Use of Abx was associated with worse overall survival (OS) and progression-free survival (PFS) in our real-world cohort of mUC pts treated with ICI. (ASCO 2023). We now report the effect of type of Abx on ICI outcomes. Methods: In our cohort of 335 pts with mUC treated with ≥2 cycles of ICI with atezolizumab (A) or pembrolizumab (P) at CCF between 2015 and 2023, Abx class used in at least 20 pts was included. PFS and OS were calculated using Kaplan-Meier method, outcomes compared using log-rank testing and multivariate (MVA) Cox regression method. Results: Of 335 pts, 26.27% received A and 73.73% received P. Median follow-up was 26.8 mos. Details of type of Abx (cephalosporins (CPH), fluoroquinolones (FQ), penicillins (PCN) and trimethoprim-sulfamethoxazole (TMP-SMX), timing (30 or 60 (d) pre or post ICI start) and OS/PFS are shown in the Table. On MVA Cox regression analysis, Abx use within 60 (d) before ICI start was associated with worse OS (10.97 vs 16.03 mos.; (HR) 1.6 [1.2 -2.1], p=0.0004). Abx use within 30 (d) post ICI was associated with worse OS (6.9 vs 14.39 mos, p=0.008). Individual Abx and effect on OS and PFS are shown in Table. On MVA Cox regression analysis, Abx use within 60 (d) post ICI start was associated with worse PFS (3.12 vs 5.67 mos., HR 1.4; p=0.015). Conclusions: In our large cohort of pts with mUC treated with ICI, Abx PCN, CPH and FQ, within 30 and 60 (d) prior to starting ICIs demonstrated worse OS, Abx used after 30 and 60 (d) of ICI demonstrated worse OS and PFS, especially with CPH and PCN but not FQ. These findings have the potential to influence clinical practice, including using a higher threshold for prescribing antibiotics to pts with mUC when planned for or on ICI and the choice of Abx used.