BACKGROUND: Previous results of the EORTC intergroup trial 40983 showed that perioperative chemotherapy with FOLFOX4 (folinic acid, fluorouracil, and oxaliplatin) increases progression-free survival (PFS) compared with surgery alone for patients with initially resectable liver metastases from colorectal cancer. Here we present overall survival data after long-term follow-up. METHODS: This randomised, controlled, parallel-group, phase 3 study recruited patients from 78 hospitals across Europe, Australia, and Hong Kong. Eligible patients aged 18-80 years who had histologically proven colorectal cancer and up to four liver metastases were randomly assigned (1:1) to either perioperative FOLFOX4 or surgery alone. Perioperative FOLFOX4 consisted of six 14-day cycles of oxaliplatin 85mg/m(2), folinic acid 200 mg/m(2) (DL form) or 100 mg/m(2) (L form) on days 1-2 plus bolus, and fluorouracil 400 mg/m(2) (bolus) and 600 mg/m(2) (continuous 22 h infusion), before and after surgery. Patients were centrally randomised by minimisation, adjusting for centre and risk score and previous adjuvant chemotherapy to primary surgery for colorectal cancer, and the trial was open label. Analysis of overall survival was by intention to treat in all randomly assigned patients. FINDINGS: Between Oct 10, 2000, and July 5, 2004, 364 patients were randomly assigned to a treatment group (182 patients in each group, of which 171 per group were eligible and 152 per group underwent resection). At a median follow-up of 8·5 years (IQR 7·6-9·5), 107 (59%) patients in the perioperative chemotherapy group had died versus 114 (63%) in the surgery-only group (HR 0·88, 95% CI 0·68-1·14; p=0·34). In all randomly assigned patients, median overall survival was 61·3 months (95% CI 51·0-83·4) in the perioperative chemotherapy group and 54·3 months (41·9-79·4) in the surgery alone group. 5-year overall survival was 51·2% (95% CI 43·6-58·3) in the perioperative chemotherapy group versus 47·8% (40·3-55·0) in the surgery-only group. Two patients in the perioperative chemotherapy group and three in the surgery-only group died from complications of protocol surgery, and one patient in the perioperative chemotherapy group died possibly as a result of toxicity of protocol treatment. INTERPRETATION: We found no difference in overall survival with the addition of perioperative chemotherapy with FOLFOX4 compared with surgery alone for patients with resectable liver metastases from colorectal cancer. However, the previously observed benefit in PFS means that perioperative chemotherapy with FOLFOX4 should remain the reference treatment for this population of patients. FUNDING: Norwegian and Swedish Cancer Societies, Cancer Research UK, Ligue Nationale Contre Cancer, US National Cancer Institute, Sanofi-Aventis.