BACKGROUND: National Surgical Adjuvant Breast and Bowel Project R-04 was designed to determine whether the oral fluoropyrimidine capecitabine could be substituted for continuous infusion 5-FU in the curative setting of stage II/III rectal cancer during neoadjuvant radiation therapy and whether the addition of oxaliplatin could further enhance the activity of fluoropyrimidine-sensitized radiation. METHODS: Patients with clinical stage II or III rectal cancer undergoing preoperative radiation were randomly assigned to one of four chemotherapy regimens in a 2x2 design: CVI 5-FU or oral capecitabine with or without oxaliplatin. The primary endpoint was local-regional tumor control. Time-to-event endpoint distributions were estimated using the Kaplan-Meier method. Hazard ratios were estimated from Cox proportional hazard models. All statistical tests were two-sided. RESULTS: Among 1608 randomized patients there were no statistically significant differences between regimens using 5-FU vs capecitabine in three-year local-regional tumor event rates (11.2% vs 11.8%), 5-year DFS (66.4% vs 67.7%), or 5-year OS (79.9% vs 80.8%); or for oxaliplatin vs no oxaliplatin for the three endpoints of local-regional events, DFS, and OS (11.2% vs 12.1%, 69.2% vs 64.2%, and 81.3% vs 79.0%). The addition of oxaliplatin was associated with statistically significantly more overall and grade 3-4 diarrhea (P < .0001). Three-year rates of local-regional recurrence among patients who underwent R0 resection ranged from 3.1 to 5.1% depending on the study arm. CONCLUSIONS: Continuous infusion 5-FU produced outcomes for local-regional control, DFS, and OS similar to those obtained with oral capecitabine combined with radiation. This study establishes capecitabine as a standard of care in the pre-operative rectal setting. Oxaliplatin did not improve the local-regional failure rate, DFS, or OS for any patient risk group but did add considerable toxicity.