Indications for liver transplantation for hepatocellular carcinoma are evolving and so-called expanded criteria remain debated. Locoregional therapies are able to downstage hepatocellular carcinoma from beyond to within the Milan criteria. We aimed to investigate the efficacy of liver transplantation after successful hepatocellular carcinoma downstaging.We did an open-label, multicentre, randomised, controlled trial designed in two phases, 2b and 3, at nine Italian tertiary care and transplantation centres. Patients aged 18-65 years with hepatocellular carcinoma beyond the Milan criteria, absence of macrovascular invasion or extrahepatic spread, 5-year estimated post-transplantation survival of at least 50%, and good liver function (Child-Pugh A-B7) were recruited and underwent tumour downstaging with locoregional, surgical, or systemic therapies according to multidisciplinary decision. After an observation period of 3 months, during which sorafenib was allowed, patients with partial or complete responses according to modified Response Evaluation Criteria in Solid Tumors were randomly assigned (1:1) by an interactive web-response system to liver transplantation or non-transplantation therapies (control group). A block randomisation (block size of 2), stratified by centre and compliance to sorafenib treatment, was applied. Liver transplantation was done with whole or split organs procured from brain-dead donors. The control group received sequences of locoregional and systemic treatment at the time of demonstrated tumour progression. The primary outcomes were 5-year tumour event-free survival for phase 2b and overall survival for phase 3. Analyses were by intention to treat. Organ allocation policy changed during the course of the study and restricted patient accrual to 4 years. This trial is registered with ClinicalTrials.gov, NCT01387503.Between March 1, 2011, and March 31, 2015, 74 patients were enrolled. Median duration of downstaging was 6 months (IQR 4-11). 29 patients dropped out before randomisation and 45 were randomly assigned: 23 to the transplantation group versus 22 to the control group. At data cutoff on July 31, 2019, median follow-up was 71 months (IQR 60-85). 5-year tumour event-free survival was 76·8% (95% CI 60·8-96·9) in the transplantation group versus 18·3% (7·1-47·0) in the control group (hazard ratio [HR] 0·20, 95% CI 0·07-0·57; p=0·003). 5-year overall survival was 77·5% (95% CI 61·9-97·1) in the transplantation group versus 31·2% (16·6-58·5) in the control group (HR 0·32, 95% CI 0·11-0·92; p=0·035). The most common registered grade 3-4 serious adverse events were hepatitis C virus recurrence (three [13%] of 23 patients) and acute transplant rejection (two [9%]) in the transplantation group, and post-embolisation syndrome (two [9%] of 22 patients) in the control group. Treatment-related deaths occurred in four patients: two (8%) of 23 patients in the transplantation group (myocardial infarction and multi-organ failure) versus two (9%) of 22 patients in the control group (liver decompensation).Although results must be interpreted with caution owing to the early closing of the trial, after effective and sustained downstaging of eligible hepatocellular carcinomas beyond the Milan criteria, liver transplantation improved tumour event-free survival and overall survival compared with non-transplantation therapies Post-downstaging tumour response could contribute to the expansion of hepatocellular carcinoma transplantation criteria.Italian Ministry of Health.