Background Prognostic factors for overall survival (OS) in patients receiving second-line chemotherapy for advanced platinum-pretreated urothelial carcinoma (UC) include ECOG performance status (PS) >0, hemoglobin (Hb) <10g/dL and liver metastasis (LM). We hypothesized that time from prior chemotherapy (TFPC) has an independent impact on OS. Methods Of 12 available phase II trials evaluating second-line therapy for advanced UC (n = 748), 7 trials with available baseline TFPC, Hb, PS and LM were utilized (n = 559). These trials evaluated vinflunine (2 trials), docetaxel alone or with vandetanib, paclitaxel-gemcitabine, nanoparticle-albumin-bound paclitaxel, cetuximab alone or with paclitaxel, and volasertib. The Kaplan-Meier method was used to estimate OS from date of starting second-line therapy. Cox proportional hazards regression stratified for trial was used to evaluate the prognostic effect of factors on OS. External validation was conducted in a second-line phase III trial comparing best supportive care (BSC) vs. vinflunine plus BSC (n = 352). Results Median OS was 5.2, 7.1, 8.3, 7.6 and 10.6 months respectively for patients having TFPC <3 months, 3 to <6 months, 6 to <9 months, 9 to <12 months and >12 months. ECOG-PS >0 (HR = 1.66), LM (HR = 1.49), Hb <10 g/dL (HR = 1.67) and TFPC (HR = 0.79) were all significant prognostic factors on multivariate analysis. TFPC as a dichotomous (<3 months vs. 3+ months) or continuous (log-transform) factor were both significant. Type of prior chemotherapy (metastatic disease vs. perioperative) was not prognostic. External validation demonstrated TFPC remained significantly prognostic (p = 0.040) for PFS but not OS (p = 0.29) after adjusting for Hb, PS and LM. Median PFS for patients with 0, 1, 2 and 3-4 factors (PS > 0, Hb < 10g/dL, LM, TFPC <3 months) was 4.13, 3.84, 1.61 and 1.45 months respectively. Conclusions A shorter duration of TFPC exhibited a significant negative prognostic impact on OS independent of ECOG-PS >0, Hb <10 g/dL and LM in patients receiving second-line therapy for advanced UC. A prognostic grouping model consisting of these 4 factors may optimize risk stratification. Disclosure T.K. Choueiri: Research support to institution from Astrazeneca. R. Fougeray: Employee of Pierre Fabre. Y. Wong: Research support to institution from Imclone. S.S. Sridhar: Research support to institution from Celgene. J. Bellmunt: Research support to institution from Pierre Fabre. All other authors have declared no conflicts of interest.