Summary Background Patients with poor-risk rectal cancer defined by MRI can be at high risk of disease recurrence despite standard chemoradiotherapy and optimum surgery. We aimed to assess the safety and long-term efficacy of neoadjuvant chemotherapy with capecitabine and oxaliplatin before chemoradiotherapy and total mesorectal excision, a treatment strategy developed to enhance the outcome of this population. Methods Between November, 2001, and August, 2005, we enrolled eligible patients with poor-risk rectal cancer defined by high-resolution MRI and without metastatic disease. The protocol was amended in January, 2004, following clinically significant cardiotoxic events (nine events in eight of 77 patients), to exclude patients with a recent history of clinically significant cardiac problems. Patients received 12 weeks of neoadjuvant capecitabine and oxaliplatin (oxaliplatin 130 mg/m2 on day 1 with capecitabine 1000 mg/m2 twice daily for 14 days every 3 weeks) followed by chemoradiotherapy (54 Gy over 6 weeks) with capecitabine (825 mg/m2 twice daily), total mesorectal excision, and 12 weeks of postoperative adjuvant capecitabine (1250 mg/m2 twice daily for 14 days every 3 weeks). The primary endpoint was pathological complete response rate. We followed up patients for a median of 55 months (IQR 47–67). Efficacy analyses were undertaken for the intention-to-treat population, unless otherwise specified. This study is registered with ClinicalTrials.gov , number NCT00220051. Findings 105 eligible patients were enrolled. Radiological response rates after neoadjuvant chemotherapy and chemoradiotherapy were 74% (78/105) and 89% (93/105), respectively. 97 patients underwent surgery, of whom 95 underwent total mesorectal excision, of whom 93 had microscopically clear resection margins and 21 had pathological complete response (21/105 20%). 3-year progression-free and overall survival were 68% (95% CI 59–77) and 83% (76–91), respectively. 3-year relapse-free survival for patients who had complete resection was 74% (65–83). Following the protocol amendment for cardiovascular safety, only one further thromboembolic event was reported (fatal pulmonary embolism). Interpretation Intensification of systemic therapy with neoadjuvant combination chemotherapy before standard treatment is feasible in poor-risk potentially operable rectal cancer, with acceptable safety and promising long-term outcomes. Future development of this multidisciplinary treatment strategy in randomised trials is warranted. Funding UK National Health Service, Sanofi-Aventis.