Background: The p53 tumor suppressor gene is frequently mutated in colorectal cancer, but reports on the effect of p53 mutations on response to adjuvant chemotherapy and survival are inconclusive. This study investigates whether p53 mutational status (wild-type, zinc or non-zinc binding mutations) impacts survival following adjuvant therapy containing fluorouracil/leucovorin with or without irinotecan (5FU/LV or IFL) in women and men with stage III colon cancer. Methods: As part of a retrospective analysis of prospectively accrued data, p53 mutational status was determined for 609 patients with stage III colon cancer who were randomized on CALGB 89803, a phase III adjuvant chemotherapy trial. p53 exons 5-8 were analyzed by direct sequencing or sequencing by hybridization. p53 mutations were identified in 276 tumors (45%), of which 134 were in the zinc binding and 142 were in the non-zinc binding regions of the core domain. Cox regression was used to study the impact of p53 mutational status, sex, and adjuvant chemotherapy on disease-free (DFS) and overall survival (OS). Results: p53 mutational status did not predict differential survival or response to adjuvant therapy among the 609 patients assessed. However, a significant sex by treatment interaction was observed for both DFS (P_interaction=0.008) and OS (P_interaction=0.002). Significant differences in DFS by p53 mutational status were observed among women (logrank P = 0.009). No such differences were observed among men (logrank P = 0.33). Similar results were observed for OS. There was marginal evidence of a treatment-related impact on the interaction between sex and p53 mutational status for both DFS and OS (DFS P_interaction = 0.07; OS P_interaction = 0.11). There was a trend toward improved OS when women with zinc binding mutations received IFL versus 5FU/LV (P = 0.08) and toward worse DFS when women with non-zinc binding mutations were treated with IFL versus 5FU/LV (P =0.08). Conclusions: This exploratory subset analysis suggests that p53 mutational status may be used to predict prognosis in a sex- and potentially chemotherapeutic regimen-specific manner.