Background: Administration of single-agent neoadjuvant immunotherapy (IO) proved to be effective and safe in the treatment of MIBC, and the identification of reliable predictors of treatment-failure would allow a more precise implementation in clinical practice. In advanced/metastatic urothelial carcinoma, ATB therapy has a negative impact on IO efficacy by modulating the intestinal microbiota towards a detrimental state of dysbiosis, eventually impairing the host anticancer immunity. However, evidences of such an effect in more confined disease, managed with an intention-to-cure attitude, are still lacking. Methods: A post hoc analysis was conducted in pts prospectively enrolled in PURE-01 study (NCT02736266), in which MIBC patients received 3 cycles of neoadjuvant pembrolizumab. ATB use was defined as any ATB administration between 30 days prior to the first pembrolizumab dose and the planned RC. Kruskal-Wallis and Chi-square tests for differences between patients treated or not with ATBs according to baseline characteristics were used. Endpoints of the study were pathologic complete response (ypT0N0) and 12- and 24-mo relapse-free survival (RFS). Multivariable logistic regression (MLR) tested the effect of ATB use on ypT0N0 rate. Secondary, we assessed RFS according to ATB use using Kaplan-Meier and multivariable Cox regression (MCR). Analyses were adjusted for baseline T stage of disease (stage II vs III), PD-L1 expression (CPS >10% vs <10%) and tumor mutational burden (TMB). Sub-analyses explored the effect of different ATB classes on the aforementioned outcomes. Results: The study cohort included 149 pts treated with neoadjuvant pembrolizumab, of which 140 (94%) underwent RC. Of all individuals, 48 pts (32%) received concomitant ATB treatment. Median TMB (9.3 Mut/Mb vs 11.4 Mut/Mb, p=0.005) and CPS (9.5% vs. 20%, p=0.04) were lower in the subgroup of patients treated with ATBs. At MLR analysis, ATB use was associated with significantly lower rate of ypT0N0 (OR 0.18, [95%CI] = 0.05-0.48, p=0.001). Patients receiving ATBs exhibited shorter 12-mo (80% [70-93] vs. 95% [91-99]) and 24-mo (63% [48-83] vs. 90% [83-97]) RFS rates than pts non receiving ATB. MCR analyses assessed that ATB treatment conferred higher risk of recurrence (HR =2.64 [1.08-6.50], p=0.03) compared to no ATB treatment, after adjusting for CPS, TMB and clinical stage at diagnosis. Fluoroquinolones were significantly associated with the worst outcomes (12-mo RFS 74% (55-99); 24-mo RFS: 61% (40-91); p=0.011; adjusted HR = 3.28 [1.12-9.60], p=0.03). Conclusions: ATB treatment was demonstrated as independently associated with lower rate of ypT0N0 and shorter RFS in MIBC treated with neoadjuvant pembrolizumab. More robust data testing the interactions between immunotherapy and gut and urinary microbiota are urgently needed. Clinical trial information: NCT02736266