OBJECTIVES: To compare the 1973 and 2004 World Health Organization (WHO) systems for the interval to tumor recurrence (TR), tumor progression (TP), and overall survival (OS) using either the superficial/muscle invasive or strict TMN pathologic staging in patients with urothelial carcinoma with ≥10 years of follow-up. METHODS: A total of 269 tumors from an institutional review board-approved bladder tumor registry were graded using the 1973 and 2004 WHO systems. Kaplan-Meier plots, the log-rank test, the chi-square test, and the Cox proportional hazard model were used to relate the clinical and histologic variables. RESULTS: The Cox model analyses, which were multivariate and included tumor stage (coded as pT1 or less versus pT2 or greater) as a significant covariate to grade, were performed and in all tumor stages were significant. The 2004 WHO grading system was more closely associated with TR (P = .025) and TP (P = .012) than was the 1973 WHO grading system (P = .47, and P = .046, respectively). OS was similar and significant for both. The OS plots for the 1973 WHO system showed a significant overlap between Stage pT1 or less, grade 2 and 3 tumors. For those with high-grade Stage pTa and high-grade Stage pT1 disease, TR and TP were similar; however, OS was significantly longer (P = .05, log-rank test) for those with Stage pTa. The OS was similar for those with high-grade Stage pT1 disease and those with Stage pT2 or greater (P = .069, log-rank test). For those with pTa, the 2004 system predicted TR and TP, but the 1973 system only predicted TP. Neither predicted OS. CONCLUSIONS: The results of our analysis have shown that the 2004 WHO system is superior to the 1973 system for predicting clinical outcomes in patients with urothelial carcinoma, independent of pathologic stage. Its primary usefulness is in those with Stage pTa.